Home

XML Prague 2011 Conference Proceedings

1. XML XML document Q i document with defaults grammar filling in semantic data types defaults constraints RELAX NG DSRL Schematron schema schema schema Figure 2 Outline of the validation procedure Itis important to add default contents before Schematron validation because YANG requires that the data tree against which XPath expressions are evaluated already have all defaults filled in http xmlsoft org xmllint html http www thaiopensource com relaxng jing html http www schematron com http code google com p pyang 113 Configuring Network Devices with NETCONF and YANG 4 6 Implementation An open source implementation of the YANG to DSDL mapping is available as a part of the pyang tool A tutorial describing practical validation of instance docu ments can be found at YANG Central While the first step of the mapping from input YANG modules to the hybrid schema is written in Python the transformation of the hybrid schema to the final DSDL schemas is implemented in XSLT Interestingly XSLT 1 0 would suffice for the latter purpose if it were not for the instance identifier type in YANG A value of this datatype is a simplified XPath expression that points to a node in the data tree In order to be able to check for the presence of this node Schematron must be able to dynamically evaluate the XPath expression This is not possible in standard XSLT 1 0 but
2. xml lt toName e if e namespaceDeclarations length gt 0 for var i 0 i lt e namespaceDeclarations length i var name e namespaceDeclarations i if name prefix length gt 0 xml xmlns name prefix name namespaceURI else xml xmlns name namespaceURI if e attributeNames length gt 0 for var i 0 i lt e attributeNames length i var name e attributeNames i var value e getAttributeValue name localName name namespaceURI xml We xml toName name xml xml value replace amp amp quot xml xml gt break case XMLItemEvent END_ ELEMENT xml lt toName e gt break case XMLItemEvent CHARACTERS xml e value break case XMLItemEvent COMMENT xml lt e value gt break 141 Efficient XML Processing in Browsers case XMLItemEvent PROCESSING INSTRUCTION xml lt e localName e value gt break y 4 Implementation Performance Characteristics An implementation of these interfaces was built into the WebKit 7 browser platform and tested in the Safari browser A set of test documents with a maximum tree depth of three with numerous repeating elements were created in sizes of 1MB 10MB and 100MB These documents were loaded within the implementation for three different configurations e xhr an
3. is non empty array and amp x300 gt lt xf trigger appearance minimal gt lt xf label gt lt xf output value gt lt xf label gt lt xf action ev event DOMActivate gt lt xf setvalue ref instance isearch search value current gt gt lt xf toggle case hide autocompletion gt lt xf action gt lt xf trigger gt lt xf repeat gt lt xf case gt lt xf case id hide autocompletion gt lt xf switch gt lt body gt lt html gt 6 Conclusion There is no technical problem for XForms in supporting JSON Mapping the JSON possibilities into an XML document is the key for a full integration allowing intuitive XPath use Other notations can be supported as easily Implementations just have to allow developers to provide their own functions to convert to and from XML for each Client side XForms implementations written in Javascript such as XSLTForms can access JSON data on different domains using JSONP with which there is no cross domain limitation XForms is JSON compatible and architectures used to JSON data exchanges can benefit from XForms implementations 24 A JSON Facade on MarkLogic Server Jason Hunter MarkLogic Corporation lt jhunter marklogic com gt Ryan Grimm lt grimm xqdev com gt Abstract What would happen if you put a facade around MarkLogic Server to have it act as a JSON store This paper explores our experience doing just that MLJSO
4. list ipv description At least one IPv6 address is required min elements 1 key address leaf address type inet ipv6 address leaf subnet mask length type uint8 range 0 128 default 64 container state config false container supply levels typedef percent type uint8 range 0 100 default 0 106 Configuring Network Devices with NETCONF and YANG leaf water type percent leaf milk type percent leaf coffee type percent leaf temperature type uint8 rpc put the kettle on input leaf recipe type enumeration enum espresso enum turkish enum Cappuccino default espresso notification low level warning leaf ingredient type enumeration enum water enum milk enum coffee 4 Mapping YANG Data Models to DSDL The network management community realized the risks associated with creating and supporting an entirely new data modelling language Therefore the NETMOD WG charter contained along with the main YANG deliverable a standardized mapping of YANG data models to Document Schema Definition Languages 2 This mapping work has been published recently as a Standard Track RFC 13 Resulting DSDL schemas are expected to serve the following purposes 107 Configuring Network Devices with NETCONF and YANG e leverage existing XML tools for validating various instance do
5. amp amp apos TEET EJAN instance result replace instance gt The same methods can be applied using query parameters Example 9 XForms Submission Using Query Parameters in XSLT verbose way lt xforms submission id doSearch method get resource http localhost 8080 orbeon exist rest db app users mail encode for uri lower case instance search amp amp query encode for uri declare namespace request http exist db org gt xguery reguest user mail reguest get parameter mail 0 instance result replace instance gt This is working but we can do much simpler relying on XForms to do the encoding all by itself The complete XForms model would then be Example 10 XForms Submission Using Query Parameters in XSLT XForms friendly lt xforms model gt lt xforms instance id search gt lt search xmlns gt lt mail gt lt guery gt declare namespace request http exist db org gt xquery request user mail request get parameter mail 0 lt guery gt lt search gt lt xforms instance gt 186 XQuery Injection lt xforms instance id result gt lt empty xmlns gt lt xforms instance gt lt xforms submission id doSearch method get b ref instance search resource http localhost 8080 orbeon exist rest db app users instance result replace instance gt lt xforms model gt 3 5 Related Attacks We have explor
6. 120 Configuring Network Devices with NETCONF and YANG lt optional gt lt optional gt lt element name cm milk gt lt ref name coffee machine state supply levels percent gt lt element gt lt optional gt lt optional gt lt element name cm coffee gt lt ref name coffee machine state supply levels percent gt lt element gt lt optional gt lt interleave gt lt element gt lt optional gt lt optional gt lt element name cm temperature gt lt data type unsignedByte gt lt element gt lt optional gt lt interleave gt lt element gt lt optional gt lt interleave gt lt start gt lt define name coffee machine state supply levels percent gt lt data type unsignedByte gt lt param name minInclusive gt 0 lt param gt lt param name maxInclusive gt 100 lt param gt lt data gt lt define gt lt grammar gt lt interleave gt lt element gt lt element gt lt start gt lt grammar gt A 3 RELAX NG Schema Global Definitions lt xml version 1 0 encoding utf 8 gt lt grammar xmlns http relaxng org ns structure 1 0 xmlns cm http example com coffee xmins nma urn ietf params xml ns netmod dsdl annotations 1 datatypeLibrary http www w3 org 2001 XMLSchema datatypes gt lt define name ietf inet types ipv6 address gt lt data type string gt 121 Configuring Network Devices with NETCONF and
7. Prague Prague Spring Prague Conservatory Prague Ruzyn u011lb Airport Prague University Prague Castle Prague Metro 21 JSON for XForms Prague East District Prague West District Prague Offensive 5 2 The XForms Page An instance is required for building the request lt xf instance id isearch mediatype application json gt action opensearch format json search lt xf instance gt Another instance is reguired for storing the responses lt xf instance id iresults mediatype application json gt lt xf instance gt A constraint is added to check whether the typed value matches an entry lt xf bind nodeset search constraint instance iresults 2 upper case gt upper case current gt A submission is defined without mentioning the callback function name which will be automatically added and without even indicating that JSONP has to be used because in XSLTForms this is set to be the default mode for cross domain requests Submission is performed even if the input control is not validated lt xf submission method get replace instance instance iresults b separator amp validate false action http en wikipedia org w api php gt Controls are used to allow input and output lt xf input id search ref search incremental true delay 500 gt lt xf label gt Subject lt xf label gt lt xf sen
8. A good example of an optimizing rewrite is one which eliminates intermediate results from an expression by unnesting a nested FLWOR expression UnnestInnerFor for a in for b in e return g return f where not rw uses variable f xs QName b gt for b in e for a in g return f The matched expresion in this example first creates a sequence by executing g for each of the items returned by e then iterates that intermediate sequence executing f Rather than create this intermediate sequence this rewrite rule unnests the inner FLWOR into a single FLWOR thus saving memory and execution time In this rewrite rule the variable names are place holders and will match any variable name their actual names are carried forward and used in the replacement expression where the place holder names are written The where clause in this rule ensures that f is not affected by bringing a new variable b into scope for it This could probably be ensured by the rewrite rule framework itself if correct analysis and variable renaming was applied 218 Declarative XQuery Rewrites for Profit or Pleasure 3 3 Rewrite Rules in XQuery Modules Rewrite rules can be included in XQuery modules by using an XQuery prolog option whose value is the rule Such rules are also imported along with the functions and global variables in a module Using this mechanism a library writer can not only write new XQuery functions but can also extend th
9. json as xs string as element json The xmlToJSON function does the reverse it accepts an XML element and returns a JSON string It does not support the passing of arbitrary XML documents It accepts only lt json gt elements whose contents follow the informal schema used by MLJSON internally to represent JSON structures declare function json xml1ToJSON Selement as element json as xs string 4 Sample JSON to XML Conversions To understand these functions let s look at their input and output The following XQuery script takes a typical JSON data structure in this case one used by Google Charts and converts it from a JSON string to XML import module namespace json http marklogic com json at mljson lib json xqy 26 A JSON Facade on MarkLogic Server json jsonToXML s iconkeySettings stateVersion 3 time notime xAxisOption NOTHING playDuration 15 iconType BUBBLE sizeOption NOTHING xZoomedDataMin null xZoomedIn false duration multiplier 1 timeUnit none yZoomedDataMin null xLambda 1 colorOption NOTHING nonSelectedAlpha 0 4 dimensions iconDimensions vZoomediIn false VAxisOption NOTHING vLambda 1 vZoomedDataMax null showTrails true xZoomedDataMax nul J It returns lt json type object gt lt iconKeySettings type array gt lt stateVersion type numb
10. 2 description The Sname strong Scontent best price This was easily implemented as a short Javascript program where the binding was automatically assessed by an XML event handler Using streaming events avoided an intermediate DOM as well as allowed stated based interpretation of the document without knowledge of the overall document structure An example invocation of this application is as follows var reader new XSONReader reader binder elementHints order attributes count isInteger gt true p Jj reader binder elementHints price isFloat true reader binder elementHints guantity isInteger true reader binder elementHints description isMixed true reader ondone function data alert JSON stringify data null reader parse xml 6 Conclusions and Future Work Exploring and implementing an event based parsing framework for web browsers has certainly been an interesting adventure The space saving outcomes by using event based parsing for large XML documents should prove to be useful but the parsing time penalty is a direct trade off that must be considered In some applica tions the ability to load large data streams and subset or filter the data will prove to be useful despite the parsing time penalty The parse time trade off may be avoided in the future by providing closer integ ration between the consumer of XML events an
11. item and rw subtype b item gt a eq b 217 Declarative XQuery Rewrites for Profit or Pleasure This turns an expression using the existentially quantified operator into one using the eq operator if its arguments are singletons The result of this normaliza tion rewrite can then be considered for possible rewriting using the aforementioned optimizing rewrite Along similar lines there are certain expressions that cannot be constant folded as they do not meet the criteria of having a constant sub expression However they may still have anumber of constant operands which could be folded if the expression were rewritten using mathematical transitivity rules rw AddAddTransFold A B C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double gt A C B where rw is constant A gt A C B where rw is constant B This rewrite rule finds nested addition operators where either B and C or A and C are constant and uses mathematical transitivity rules to move the constant ex pressions together so that they will be constant folded 3 2 Optimizing Rewrites Optimizing rewrites will usually operate on the result of normalizing rewrites and aim to replace slower expressions with faster alternatives Sometimes this judgement can be based on heuristics whilst at other times some cost analysis might be per formed
12. www w3 org TR XMLHttpRequest2 3 Crockford D The application json Media Type for JavaScript Object Notation JSON July 2006 http www ietf org rfc rfc4627 txt 4 Le Hors A et al Document Object Model DOM Level 2 Core Specification November 2000 http www w3 org TR DOM Level 2 Core 5 The Unicode Standard Version 6 0 0 2011 ISBN 978 1 936213 01 6 http www unicode org versions Unicode6 0 0 6 Simple API for XML http www saxproject org 7 WebKit http www webkit org 147 148 EPUB Chapter and Verse Electronic origination of the Bible Tony Graham Mentea lt tgraham mentea net gt Mark Howe Jalfrezi Software lt mark cyberporte com gt Abstract The link between the Bible and publishing technology is at least as old as Gutenberg s press Four hundred years after the publication of the King James Bible we were asked to convert five modern French Bible translations from SFM a widely used ad hoc TROFF like markup scheme used to produce printed Bibles to EPUB We used XSLT 2 0 and Ant to perform all stages of the conversion process Along the way we discovered previously unimagined creativity in the original markup even within a single translation We cursed the medieval scholars and the modern editors who have colluded to produce several mutually incompatible document hierarchies We struggled to map various typesetting features to EPUB E Reader compatibility made us nostal gic for brow
13. 1 2 Related Work The XQuery Formal Semantics document itself 2 defines normalization mapping rules which use a graphical notation to describe query rewrites This is thorough but impractical from an authoring and execution perspective Many academic papers on the subject of XQuery optimization use an ad hoc notation for expressing both the query algebra and the rewrites performed on it ie 7 meaning that every paper presents not only its original concepts but also an entirely new notation to become familiar with Michael Kay has previously proposed a solution using XSLT directly to perform rewrites 6 by manipulating an XML representation of the program s expression tree Whilst effective this approach is not as natural or readable as an approach that uses the expression syntax directly in its patterns and transformations The Glasgow Haskell Compiler allows rewrite rules in pragmas 5 Their rule syntax has a simplicity that derives from the regular nature of the Haskell syntax but is not powerful enough to take into account further information available through query analysis 2 Rewrite Rules 2 1 Examining the Problem After parsing and static analysis an XQuery program is turned into an expression tree representing the operations in the program and how they should be executed The process of applying rewrites to this tree is one of identifying relevant subtrees that satisfy the conditions for the rewrite and replacing that
14. Cheap DBMS POETDB u poet u ep u Ezra Pound u 45 Usura Place Hailey ID u co u Christopher Okigbo u 7 Heaven s Gate Idoto Anambra Nigeria u work u cantos u The Cantos u poet ep u mauberley u Hugh Selwyn Mauberley u poet ep u thunder u Paths of Thunder u poet co hy def not found baseuri ruri reguest uri reguest environ response code 404 Not Found return u lt p gt Unable to find s try the lt a href s gt index of works lt a gt lt gt p gt S ruri baseuri simple service GET POETDB SERVICE ID poetdb text html def poetdb baseuri request environ SCRIPT NAME def get_work wid uri join baseuri work wid name poet POETDB u work wid puri join baseuri poet poet o return lt p gt Poetic work lt a href s gt s lt a gt by lt a href s gt linked gt 75 Akara Spicy Bean Fritters and XML Data Services poet lt a gt lt p gt uri name puri def get poet pid uri join baseuri poet pid name address POETDB u poet pid return lt p gt Poet lt a href s gt s lt a gt lt p gt uri name getters u work get work u poet get poet entitytype shift path info reguest environ if not entitytype entitytype u work if entitytype not in POETDB return not_found baseuri eid shift path info reguest environ if not eid Return a
15. E o E 8 F N Jj M o xhr events no handler Figure 2 Time Performance in Comparison to XMLHttpReguest It should be noted that while the overall parsing time is slower due to the delivery of events to Javascript based event listeners the overall application performance may be faster Any application that deconstructs an XML document into some combination of service invocations or mappings to local data structures will not have to traverse the XML DOM at the end of the parsing As such the overall time may improve for any application action where XML parsing of web resources is involved In fact in the simple example of finding links shown in the previous sec tion the overall processing time was twice as fast for the XMLReader implementation over the XMLHttpReguest implementation 144 Efficient XML Processing in Browsers 5 An Example Application One example application amongst many for event oriented parsing is the unmar shalling of XML into local data structures specifically JSON An event handler for XML that has algorithms for building JSON and so Javascript objects from XML automatically can easily be built using XMLReader that avoids the construction and memory consumption of a whole document DOM This kind of binding process can be used to access XML services while providing JSON directly to the caller One interesting aspect of such a tool is that it can be configured to filter subset and use s
16. Eliot Crozier Ken Using the NETCONF Protocol over the Blocks Extensible Exchange Protocol BEEP RFC 4744 December 2006 IETF http tools ietf org html rfc4744 8 Chisholm Sharon Trevino Hector NETCONF Event Notifications RFC 5277 July 2008 IETF http tools ietf org html rfc5277 9 Badra Mohamad NETCONF over Transport Layer Security TLS RFC 5539 May 2009 IETF http tools ietf org html rfc5539 10 Bjorklund Martin Ed YANG A Data Modeling Language for the Network Configuration Protocol NETCONF RFC 6020 October 2010 IETF http tools ietf org html rfc6020 11 Schonwalder J rgen Ed Common YANG Data Types RFC 6021 October 2010 IETF http tools ietf org html rfc6021 12 Bierman Andy Guidelines for Authors and Reviewers of YANG Data Model Document RFC 6087 January 2011 IETF http tools ietf org html rfc6087 13 Lhotka Ladislav Ed Mapping YANG to Document Schema Definition Languages and Validating NETCONF Content RFC 6110 February 2011 IETF http tools ietf org html rfc6110 115 Configuring Network Devices with NETCONF and YANG 14 Clark James Murata Makoto Ed RELAX NG DTD Compatibility OASIS Committee Specification 3 December 2001 http relaxng org compatibility 20011203 html 15 Schonwalder J rgen Pras Aiko Martin Flatin Jean Philippe On the future of Internet management technologies IEEE Communications Magazine October 2003
17. Psalm titles lists paragraphs tables etc e Add inline markup e Highlights glossary terms e Verse end milestones e Cleanup The structural stages are consecutive since they all operate on lt t line gt elements and as successive stages run there are fewer and fewer lt t line gt elements remaining and one of the important functions of the cleanup phase is producing error messages for any that remain Footnotes are moved to the point of their footnote reference or just into the preceding lt t line gt in translations that don t use footnote references before adding paragraphs etc since the footnotes are ignored when merging lines Adding inline markup was interesting since e Highlighted bold italic etc text can contain notes which are XML at that point and glossary terms e Both notes and glossary terms can contain highlighted text e Highlights can nest but all use the same x SFM markup to end any sort of highlighted region The stage for adding lt hi gt elements therefore finds the first highlight start code li b etc in a text node child of an element and then processes the remainder of that text node and any following sibling nodes to find the end code When found the text node fragments and any intervening elements are wrapped in a lt hi gt ele ment with the appropriate type attribute value Along the way the intervening elements were themselves recursively processed to turn highlight code
18. Snumber gt s select mode current gt lt xsl apply templ lt xsl copy gt lt xsl text gt 4fFxA lt xsl lt xsl template gt text gt 161 EPUB Chapter and Verse 3 1 3 Auxiliary files The processing to produce the OPF and NCX files depends on several auxiliary files booklist xml For every book in the OSIS the book s identifier title and file name and for each chapter in the book the chapter s identifier number and whether or not it contains footnotes Generated from the OSIS files themselves using collection lt All the XML files in osis dir gt lt xsl variable name 0sis docs select collection concat file osis dir select xml gt notelist xml For every external note the note s identifier title filename and location in the Bible plus the filename of each figure in the note Generated from the OSIS for the notes EPUB hierarchy xml Expanded hierarchy of the specific books for the EPUB with the correct titles for the divisions within the hierarchy EPUB notelist xml Information from notelist xml for just the notes in the particular EPUB EPUB figurelist txt EPUB files txt EPUB notelist txt Lists of filenames one file per line for Ant to insert into the particular EPUB 3 2 The millstone of milestones As mentioned previously ancient and modern hierarchies within Bible documents create particular challenges for XM
19. and in any case SFM for printed Bibles only has to work for one origination process Tools to convert from SFM to XML are therefore likely to be useful for the foreseeable future Definitions of the SEM codes 151 EPUB Chapter and Verse 2 From SFM to OSIS 2 1 The OSIS schema Open Scriptural Information Standard 14 is an XML format for describing Scripture and related text The OSIS effort has good theological linguistic and XML creden tials Its authors include Steve DeRose XML XPath 1 0 etc and Patrick Durasau OpenDocument format Topic Maps 3 The OSIS 2 1 1 documentation 15 however is still labelled draft despite being published in 2006 and it shows signs of multiple editors and last minute changes as well as having references to elements that don t exist and some hyperlinks that don t work The commonsense advice from the experts about converting SFM to OSIS 5 is The first task in preparing to convert SFM files to OSIS is to clean the text The more regular your source files are the more likely the conversion process will operate correctly Maybe we could have picked one SFM format and modified currently four other translations to match But our SFM is sacred text in more ways than one since the client wants to continue to use the SFM as SFM for some purposes We have worked around the inconsistencies as much as we can and have kept the SFM changes to a minimum pending our own day of rec
20. hello xg hello xs index html xproject 46 CXAN a case study for Servlex an XML web framework expath web xml project xml 4 The CXAN website Now that we have met the basics of the framework let us have a look at a real world example the CXAN website The purpose of this website has stated in the introduc tion is double first to be an online catalog of packages to be browsed by human with a graphical environment and secondly to be a central server for CXAN clients to commnicate to through HTTP and XML in order to maintain a local install of CXAN packages by allowing searching and installing packages Both aspects are actually very similar The website has to respond to HTTP re quests sent to structured URIs In both cases the responses carry the same semantics The difference is that in one case the format of the information is a HTML page aimed at human beings and in the other case the format is an XML message aimed at a computer program This section will first focus on the HTML part of the website then will show how both parts are actually implemented in common So the main business of the website is to maintain a set of packages and to provide a way to navigate and to search them as well as to display their details and down loading them There must also be a way to upload a new package So the overall architecture is pretty straight forward a plain directory on the filesystem to save the package files an XM
21. less code as an incentive e declarative transparency structuring usage for easy translation from one system of declarations to others and to reuse standard declarations systems such as XSLT patterns where appropriate The result is an XML processing library that s truly different from anything else out there Whether it suits one s tastes or not is a matter of taste The Web server system Akara proper is designed to be layered upon Amara 2 x in a RESTful context as a lightweight vehicle for deploying data transforms as services 2 The basic tree APIs Amara 2 0 comes with several tree APIs and makes it fairly easy to design custom tree APIs by extension 55 Akara Spicy Bean Fritters and XML Data Services 2 1 Parsing XML into simple trees The most fundamental tree API is just called amara tree It s very simple and highly optimized but it lacks some of the features of the Bindery API which is recommen ded unless you really need to wring out every ounce of performance import amara from amara import tree MONTY XML lt monty gt lt python spam eggs gt What do you mean bleh lt python gt lt python ministry abuse gt But I was looking for argument lt python gt lt monty gt doc amara parse MONTY XML doc is an amara tree entity node the root of nodes representing the elements at tributes text etc in the document assert doc xml type tree entity xml type doc xml chil
22. lt a gt lt b gt lt a gt doc xml write lt xml version 1 0 encoding UTF 8 gt n lt a gt lt b gt lt a gt 2 3 Using XPath XPath is also available for navigation amara tree as well as Bindery and other derived node systems fully supports XPath which means all the other implement ations do as well Use the xml select method for nodes from amara import bindery MONTY XML lt monty gt lt python spam eggs gt What do you mean bleh lt python gt lt python ministry abuse gt But I was looking for argument lt python gt lt monty gt doc bindery parse MONTY XML m doc monty pl doc monty python print pl xml select u string spam 59 Akara Spicy Bean Fritters and XML Data Services for p in doc xml select u python p xml write 2 4 Parsing HTML Amara integrates html5lib for building a bindery from non well formed HTML and even non well formed XML though the latter is always an abomination from amara bindery import html H lt html gt lt head gt lt title gt Amara lt title gt lt body gt lt p class DESC gt XML processing toolkit lt p gt Python meets lt br gt XML lt html gt doc html parse H Use bindery operations print unicode doc html head title Use XPath print doc xml select u string html head title Re serialize to well formed output doc xml write The last line in effect tidies up the messy
23. p 90 97 16 Biron Paul V Malhotra Ashok XML Schema Part 2 Datatypes Second Edition W3C Recommendation 28 October 2004 http www w3 org TR 2004 REC xmlschema 2 20041028 17 Van der Vlist Eric RELAX NG O Reilly amp Associates 2004 xviii 486 p ISBN 978 0 596 00421 7 http books xmlschemata org relaxng A Schemas for the Coffee Machine Data Model This appendix contains the schemas mapped from the coffee machine YANG module The hybrid schema is listed in Section A 1 and the subseguent sections then contain the validating schemas RELAX NG Schematron and DSRL The schemas are complete except for the long regular expressions constraining IPv4 and IPv6 addresses which have been trimmed A 1 Hybrid Schema lt xml version 1 0 encoding UTF 8 gt lt grammar xmlns http relaxng org ns structure 1 0 xmlns nma urn ietf params xml ns netmod dsdl annotations 1 xmlns cm http example com coffee xmlns a http relaxng org ns compatibility annotations 1 0 xmlns dc http purl org dc terms datatypeLibrary http www w3 org 2001 XMLSchema datatypes gt lt dc creator gt Pyang 1 0 DSDL plugin lt dc creator gt lt de date gt 2011 01 27 lt dc date gt lt start gt lt grammar nma module coffee machine ns http example com coffee gt lt dc source gt YANG module coffee machine revision 2010 03 26 lt dc source gt lt start gt lt nma data gt lt interleave gt lt element name cm conf
24. quence of characters and loaded into an XML DOM that is stored in responseXML If the response is not XML but is in some representation that encodes a sequence of Unicode characters the responseText is still usable This is the case for JSON where the responseText can be passed to a JSON parser or evaluated within the Javascript scripting environment There are several deficiencies in this model First if the response is not XML and not a sequence of characters there is little support for handling the response data e g raw image data Fortunately there is anew Blob interface that may address this need by some web applications 2 135 Efficient XML Processing in Browsers Second in the case where the response is not XML the use of an intermediary representation as a sequence of characters may be wasteful For example for formats like JSON the syntax could be parsed efficiently within internals of the implement ation to return a JSON object directly to the calling application The character rep resentation should be reserved for unrecognized or non specialized character based media types Finally in the case of XML documents the whole document treatment of the response isn t what is always wanted by the calling application The response may be intended to be bound to specific data structures As such the intermediary DOM data structure is wasteful in both processing time and memory consumption Even further a web applicat
25. return result return item This code uses the static type of the argument expression to perform early evaluation in some cases and to remove the unneeded effective boolean value operation in other cases Whilst with a little context it is possible to decode what is going on in this function it is far from succinct or easily understandable Using the rewrite rule notation this method can be replaced with the following single rule fn EBVFold boolean e gt false where rw subtype e empty sequence gt exists e where rw subtype e node gt e where rw subtype se xs boolean This rule is far easier to understand and easily shows the optimization behaviour when the argument expression has been inferred to be one of three different types In short when the argument is empty the effective boolean value is false When the argument is a single xs boolean value the effective boolean value returns it and is therefore unnecessary When the argument consists only of nodes then the effect ive boolean value gives the same result as the exists function which itself has rewrite rules which will be applied to it In total the lines of code attributed to these optimizations was reduced from 676 lines to 165 a 75 reduction As a crude measurement of understandability and maintainability this figure shows some of the benefits of using rewrite rule notation In addition this should ma
26. u tag u poet Each triple is current resource id relationship string result xpath expression Notice the U convenience function which takes an object and figures out a way to get you back a Unicode object Python s iterator goodness makes it easy to organize this data in any convenient way for example from itertools import groupby from operator import itemgetter from amara lib import U for rid triples in groupby generate metadata doc itemgetter 0 print Resource rid for row in triples print t row 0 row 1 U row 2 The output is Resource co u co u name u Christopher Okigbo u co u place u Idoto Anambra u co u opus u r2e0ele5 Resource r2eQeled u r2e0ele5 u title u Heaven s Gate Resource co u co u tag u biafra u co u tag u poet 68 Akara Spicy Bean Fritters and XML Data Services 5 Incremental parsing Imagine a 10MB XML file with a very long sequence of small records If one tries to use a convenient tree API you will end up trying to load into memory several times the full XML document but very often when processing such files all that matters in processing is one record at a time One could switch to SAX but then lose the convenience of the tree API Amara provides a system for incremental parsing which yields subtrees according to a declared pattern provided as the function amara pushtree which requi
27. world lt p gt lt body gt lt html gt lt web body gt lt web response gt 40 CXAN a case study for Servlex an XML web framework The response includes a status code and the status message of the HTTP response It can also and usually do contain an entity content the body of the response In this case this is an HTML page with the content type text html Optionally the response document can set some headers on the HTTP response Besides components in XSLT XQuery and XProc a webapp can contain resources They are also identified using a regex but then the path is resolved in the webapp s directory and the webapp container returns them as is from the filesystem The map can set their MIME content type and can also use a regex rewrite pattern to rewrite a resource URI path to a path in the filesystem This is useful to have the webapp container serving directly paths like style css and images png without ac tually calling any component and without having to generate the request document A web application is thus a set of components along with a map mapping re quest URIs to components It is packaged using the Packaging System format with the following structure expath pkg xml expath web xml the webapp component one xsl two xproc tres xam any dir more components xproc images logo png style layout css Because this is a standard package with the addition of the webapp descriptor ak
28. 1 2 Iterators AI functions XPath selectors and other operators in MXOuery are implemented through iterators An iterator takes as input zero an iterator can return a constant value or more iterators and outputs a token stream Iterators are then combined into a tree DAG to represent the guery 4 1 1 3 Stores While MXOuery tries to stream tokens as much as possible in many cases additional functionality is needed such updates full text or stream storage A store provides an interface for getting an iterator for an XDM instance and for additional operations like indexed path access applying updates or fulltext retrieval XDM tokens in MXOuery carry a link to their store in their node IDs see also Section 4 2 3 4 1 1 4 Customizability and Platform Abstractions In order to easily adapt MXOuery for environments with restricted processing re sources several components can be detached Most of the functions and extension modules are loaded dynamically Stores for particular functionality streams fulltext updates are instantiated on demand and can be omitted when this functionality is not needed Furthermore even the parser compiler frontend can be removed if not interactive compilation is needed Within the scope of this prototype we have not exercised these options If an additional code functionality reduction is reguired some of these options will become useful Since the Java language space has become fragemented over various
29. 1 4 times The Figure 3 illustrates the FSM of the provided example The figure shows that street follows the town After the street we can have another street which is limited by maxOccurs flat room or exit edge i e no element room can follow the street or flat The following W3C XML document confirms to the schema lt xml version 1 0 gt lt address gt lt town gt Edinburgh lt town gt lt street gt Parkside Terrace lt street gt lt room gt 1 lt room gt lt address gt If we select the street element Insert After element after is room and the candidates are street and flat street equals to the selected element so we can insert it after flat edges contain element after room so we can insert it after Insert Before element before is town and the only candidate is street since street equals the selected element Client side XML Schema validation fat town street room S maxOccurs O gt ee Figure 3 FSM of address element we can insert it before Insert Into 1 element street is of simple type and does not admit elements inside Insert Into 2 element street does not admit attributes inside Delete the minOccurs field is equal to 1 and the real occurrence rate is equal to 1 since we cannot delete the selected element Edit does not apply to elements If we select the address element Insert After Insert Before Insert Into 2 and Edit does not appl
30. 111 Configuring Network Devices with NETCONF and YANG Whenever this pattern is applied the instantiating pattern must provide the values for prefix and start as its parameters for example lt sch pattern id id2573371 is a example sorted leaf list gt lt sch param name pref value ex gt lt sch param name start value nc rpc reply nc data gt lt sch pattern gt 4 4 Mapping Default Contents The YANG to DSDL mapping uses a subset of DSRL for specifying default contents as indicated in the hybrid schema For example the hybrid schema for the coffee machine data model Section A 1 contains the following annotated element pattern lt element name cm recipe nma default espresso gt lt element gt The nma default annotation is mapped to the DSRL schema for the put the kettle on request as the following element lt dsrl element map gt lt dsrl parent gt nc rpc cm put the kettle on lt dsrl parent gt lt dsrl name gt cm recipe lt dsrl name gt lt dsrl default content gt espresso lt dsrl default content gt lt dsrl element map gt The mapping procedure for DSRL is slightly complicated by the following two factors 1 A default value may be specified for an element appearing inside a named pattern definition As DSRL offers nothing similar to Schematron abstract patterns the only possible solution is to use the corresponding element map repeatedly at all places where the named pattern is refer
31. 8880 poetdb work cantos in a browser you will get a page saying Work The Cantos click for poet info and if you click the link it will take you to a page with the representation of the poet resource http localhost 8880 poetdb poet ep based on the relative link set up in the POETRYDB data structure Now you re really getting into the Web application space and rubbing up a bit against REST in that resources such as poet and work are clearly identified by URL and clearly referenced within the content via hypermedia i e good old Web links 9 Error handling and making things more robust Try out the following URL on the above service 74 Akara Spicy Bean Fritters and XML Data Services http localhost 8880 poetdb poet noep You get the dreaded 500 error The Web is a wild place and you never know what input or conditions you re going to be dealing with so anticipating and gracefully handling errors is important Let s set it up so that the server returns a 404 Not Found error in case the URL path doesn t match anything in the database Let s also set up some basic link index pages to help the user In general the following is amuch more complete and functional example from wsgiref util import shift path info request uri rom amara lib iri import join from akara services import simple service ci from akara import request respons POETDB SERVICE ID http example org my services poetdb
32. An example for a translation rule is shown in Table 4 by means of an SQL SELECT statement Table 4 Translation of SELECT query from SQL to XQuery tableSQL nodesxXQu SELECT exprListgo for sourcel sourcelyoy FROM sourcelsoz SourceNgg ae WHERE boolExprso for sourceN sourceNyoy GROUP BY groupByso group by groupByyo y ORDER BY orderListgo where boolExpryp order by orderListyo return nodesyoy 5 Implementation and Evaluation 5 1 Implementation We have implemented our formal translation framework as a cross compiler from SPARQL and SQL to XQuery which is available as a command line tool and as a web service 11 The code is written in C and follows the classical compiler archi tecture Lexer Flex Parser Bison Semantic Checks and Code Generation towards textual XQuery which can be consumed by any XQuery engine The current focus is on complete and correct translation Thus only a limited amount of optimizations are present mostly pushdown of constants and filter predicates within basic pattern matches As the results show these optimizations are already useful but for more complex queries more effort to minimize interme diate results is necessary 90 Translating SPARQL and SQL to XQuery 5 2 Evaluation We evaluated our translation with regard to correctness completeness and perform ance Although at this stage we have no formal verifica
33. Answer you generate HTML that contains Javascript And the Javascript is the part that does all the interesting stuff So you re writing in two languages instead of one and they have an awkward relationship writing XSLT that generates Javascript code that responds to user events by invoking another XSLT transforma tion that in turn generates more Javascript can t realistically be considered an elegant architecture The success of CSS hasn t helped XSLT s cause either CSS and XSLT aren t in direct competition no one will deny that XSLT s capabilities go far beyond what CSS can do But many of the original aspirations of XSLT the separation of content from presentation can be achieved to a reasonable extent using CSS alone So enhancements to CSS reduce the incentive to learn XSLT just as enhancements to HTML forms diminish the attractions of XForms At this stage one could write off XSLT on the browser as a failed technology But I wouldn t be speaking on the topic if my aim was to deliver an obituary Because the fact is the state of software development on the browser is far from rosy and I m convinced that far better tools are both needed and feasible and I m also con 126 XSLT in the Browser vinced that developers would be much better off making more use of declarative languages if those languages delivered what developers need The browser has become a monoculture for software developers HTML5 plus Javascript is what we
34. CXAN website application defines an abstract representation of a page an element page with a title some para image italic text code snippets list and some table elements Each pipeline has to focus only on building such a simple description of the page to display to the user In the web descriptor the application sets a transformer for all pages which is an XSLT stylesheet This stylesheet generates the web response element Servlex is expecting including the HTML version of the page transformed from its abstract respresentation Of course when the browser receives the HTML page it displays it When doing so it finds that the page refers to some images CSS stylesheets and Javascript files So the same cycle starts again it sends as much HTTP request as resources to retrieve But because those resources are static files they are handled differently When Servlex receives the request it looks in the web descriptor for a component matching the URI It then founds a specific kind of component resource components The resource components specify also a URL pattern and set the content type of the resource they can also contain a rewriting rule based on the third parameter of fn replace The resource is then read from within the package based on its name and returned directly to the user lt resource pattern style css media type text css gt lt resource pattern images gif media type image gif gt lt resource
35. DITA NG A Relax NG implementation of DITA 3 3 3 Simpler schema construction The order in which entities are declare in DTDs is very important as some need to redefine others and thus each module is split between an ent and an mod file the shell DTD needs to carefully include each module for a specific order The Relax NG schema only needs to include a single mod file for a domain 4 Processing Relax NG based DITA documents 4 1 Getting deliverables through DITA OT What is needed to get the DITA OT to process DITA documents based on Relax NG is to set a system property to specify the parser configuration that it should use for Xerces That needs to be set to a parser configuration that adds the Relax NG default values For example on the command line that starts the DITA processing one can just add Dorg apache xerces xni parser XMLParserConfiguration com oxygenxml relaxng defaults RelaxDefaultsParserConfiguration In addition to this you need also to set the validate ant property to false it defaults to true as otherwise DITA OT will complain about a missing DTD or XML Schema this is something that DITA OT should fix all they need is to set also the dynamic validation feature for Xerces that means it will force validation only if a DTD or schema is specified 4 2 Editing in oXygen I experimentally integrated the Jing compiled with support for default attribute values and switched all the parsers to use the Xerces comp
36. Ps 3 9 gt 3 9 lt reference gt lt hi type italic gt lt hi type bold gt au S lt hi type small caps gt EIGNEUR lt hi gt lt hi gt lt hi gt lt hi type italic gt lt hi type bold gt gu appartient le salut lt hi gt lt hi gt 37 39 Jon 2 10 Pr 21 31 Voir lt hi type italic gt lt hi type bold gt lt w gloss benediction gt b n diction lt w gt lt p hi gt lt hi gt lt note gt C est au S lt hi type small caps gt EIGNEUR lt hi gt qu appartient le salut lt 1 gt lt l level 1 gt Que ta b n diction soit sur ton peuple lt 1 gt lt l type selah gt Pause lt verse eID Ps 3 9 gt lt 1 gt lt 1g gt lt div gt lt chapter eID Ps 3 gt lt chapter osisID Ps 4 sID Ps 4 n 4 gt lt div type section gt lt title gt Exauc au milieu de l hostilit lt title gt 156 EPUB Chapter and Verse lt title type psalm canonical true gt lt verse osisID Ps 4 1 sID Ps 4 1 n 1 gt Du chef de che ur Avec instruments a cordes Psaume De David lt verse eID Ps 4 1 gt lt title gt a e2 88 98 GS 88 99 Key Each q makes a separate lt l gt Line groups are implicit in the SFM but are grouped using xsl for each group into lt lg gt in the OSIS For each verse start milestone osisID identifies the verse or range of verses sID matches eID in the corresponding end milestone and tn preser
37. Tomcat handles this safely and only exposes a yes no answer to user entries and sends error messages to the server log but on other forms this could also be an option leading to a third kind of attacks If we know the database structure for any reason this could be because we ve successfully leaked information in error messages because the application s code is open sourced or because you ve managed to introspect the database using func tions such as xmldb get child collections we can also update user information with forged authentication data let Su user role orbeon admin 1 return update value u mail with eric example com update value Su password with foobar This can be done by pasting the URL encoded value of the following string 3 Protection Now that we ve seen the harm that these attacks can do what can we do to prevent them A first set of recommendations is to limit the consequences of these attacks 1 Do not store non encrypted passwords 2 Use a user with read only permissions to perform read only queries 3 Do not enable extensions modules unless you really need them If the authentication realm of our example had followed these basic recommenda tions our attacks would have had limited consequences and they are always worth to follow but they do not block the possibility to perform attacks To block the attacks themselves we need a way to avoid that the values that are copied int
38. XML this interface has been widely used to allow browser based applications to interact with a variety of web services and content many of which are now in other formats which includes JSON 3 At the time of origin of this interface the design pattern of building and unmarshalling whole XML documents into a DOM 4 probably made sense but there are now many use cases where processing a document to build a whole DOM representation is undesired if not infeasible Experience with XMLHttpRequest in web applications has brought the efficiency and ease of use of using XML to communicate data responses or other content to web applications into question This is especially true when there are simpler formats like JSON that are available and readily usable within a browser s scripting envir onment But even for JSON the design of the XMLHttpRequest interface leaves a lot of room for improvement due to the both conflating the original intent of loading XML with alternate formats and the whole document design pattern Once a request has been made to a web resource using XMLHttpRequest and that request has successfully returned the entity body of the response is loaded into two simultaneous objects readonly attribute DOMString responseText readonly attribute Document responseXML The decoded sequence of Unicode 5 characters representing the resource is loaded into the responseText and if possible the XML document is parsed from that se
39. YANG lt param name pattern gt trimmed lt param gt lt data gt lt define gt lt define name ietf inet types ipv4 address gt lt data type string gt lt param name pattern gt trimmed lt param gt lt data gt lt define gt lt grammar gt A 4 Schematron Schema lt xml version 1 0 encoding utf 8 gt lt sch schema xmlns sch http purl oclc org dsdl schematron gueryBinding exslt gt lt sch ns uri http exslt org dynamic prefix dyn gt lt sch ns uri http example com coffee prefix cm gt lt sch ns uri urn ietf params xml ns netconf base 1 0 prefix no gt lt sch pattern id coffee machine gt lt sch rule context nce rpc reply nc data cm configuration cm ipv6 gt lt sch report test preceding sibling cm ipv6 cm address current cm address gt Duplicate key cm address lt sch report gt lt sch rule gt lt sch pattern gt lt sch schema gt A 5 DSRL Schema lt xml version 1 0 encoding utf 8 gt lt dsrl maps xmlns dsrl http purl oclc org dsdl dsrl xmlns cm http example com coffee xmlns nc urn ietf params xml ns netconf base 1 0 gt lt dsrl element map gt lt dsrl parent gt nc rpc reply nc data cm configuration cm ipv6 lt dsrl parent gt lt dsrl name gt cm subnet mask length lt dsrl name gt lt dsrl default content gt 64 lt dsrl default content gt lt dsrl element map gt lt dsrl element map gt lt dsrl
40. You can do this in several ways but I ll focus on the modeling by ex ample approach Examplotron see Introducing Examplotron is an XML schema language where an example document is basically your schema The following listing is a regular XML document and is also an Examplotron schema LABEL MODEL lt xml version 1 0 encoding utf 8 gt lt labels gt lt label gt lt name gt Addressee name lt name gt lt address gt lt street gt Address street info lt street gt lt city gt City lt city gt lt state gt State abbreviation lt state gt lt address gt lt label gt lt labels gt It establishes a model that there is a labels element at the top containing a label element child and so on In this case the intention is that there are multiple label element children and Examplotron allows you to clarify this point using an inline annotation LABEL MODEL lt xml version 1 0 encoding utf 8 gt lt labels xmlns eg http examplotron org 0 gt lt label eg occurs gt lt name gt Addressee name lt name gt lt address gt lt street gt Address street info lt street gt lt city gt City lt city gt lt state gt State abbreviation lt state gt lt address gt lt label gt lt labels gt Specifically eg occurs indicates 0 or more occurrences The following is an XML document that conforms to the schema VALID LABEL XML lt xml version 1 0 encoding
41. address gt 2001 db8 1 c0ffee lt address gt lt subnet mask length gt 64 lt subnet mask length gt lt ipv6 gt lt configuration gt lt state xmlns http example com coffee gt lt supply levels gt lt water gt 91 lt water gt lt coffee gt 73 lt coffee gt lt milk gt 19 lt milk gt lt supply levels gt lt temperature gt 67 lt temperature gt lt state gt lt data gt lt rpc reply gt The client may also guery specific parts of the data tree One way for doing this are XPath expressions which however need not be implemented by all servers The mandatory mechanism for narrowing a guery which every server has to support are subtree filters see 4 sec 6 For example the client can ask for just the temperature parameter by using this reguest lt rpc message id 124 xmins urn ietf params xml ns netconf base 1 0 gt lt get gt lt filter type subtree gt lt state xmlns http example com coffee gt lt temperature gt lt state gt lt filter gt lt get gt lt rpe gt Another important method is edit config which enables the client to modify server s configuration datastores Apart from the essential running configuration datastore which governs the device operation other configuration datastores may also be available A typical example is the candidate datastore where a new config uration may be prepared and then committed to become the new running configur ation 101 Configu
42. all nodes in a HTML document should be in the XHTML namespace http www w3 org 1999 xhtm1 While node names in XHTML are defined to be lowercase the standard implementation in HTML for node nodeName has always 200 XQuery in the Browser reloaded Runtime Dispatcher XML XDM Store DOM5 HTML Store DOM Event Figure 1 XOIB Architecture Overview been to return an uppercase value To avoid breaking existing code this is how the attributes are defined in HTML5S Table 2 DOM5 HTML attributes and their use for traversal Function Semantics Used for String node namespaceURI hardcoded to http retrieving the namespace of www w3 org 1999 xhtml lan attribute element String node localName node s name without retrieving the name of an namespace in lowercase __ attribute element String node nodeName returns node localName in not used Stringelement tagName uppercase NamedNodeMap retrieving the attributes of node attributes a Node Node node firstChildNode axis navigation node nextSiblingNode node parentNode Therefore we can just use node localName and node namespaceURI to get the same behavior without difference between an XHTML and an HTML document An important decision in the design of XQIB was the linking between the DOM as a materialized updateable tree and and the stream of immutable tokens Instead of building a shadow structure of tokens that are generated when
43. and added to the on disk repository so it will be available even after you restart Tomcat The manager lists the installed applications and allows you to remove them or to install new ones e o E Categories ae Manager gt C localhost 8090 serviex manager Manager lt gt 88 p m projects me DEVICES a Name Date Modified Installed webapps EL neptune 7 V G expath Today 17 21 4 Macintosh HD gt aj crypto 22 06 10 i o x xqts browser gt Ol expath Today 16 47 e x dev tools sachet B files 16 02 10 e x webapp tools E Uranus gt z http client 17 05 10 e x expath PLACES E pkg 19 12 10 e x xnippets FA Desktop v 3 servlex 10 09 10 e x oauth example N fgeorges gt z build 20 11 10 e x h2o A Applications README 10 09 10 e oie Documents v 2 samples 12 09 10 a v g hello zal t GE expath v 3 hello world Today 17 12 X hello xsitforms E sever v Gdist Today 17 19 e x cxan mm projects hello world 0 4 0 xaw Today 17 15 gt a src 23 11 10 Deploy a webapp Ch File No file ch Java ploy pp oose File o file chosen a i Ga xproject Today 17 15 s SEARCH FOR f Clear the application cache Reset T Toda pene k o V A gt Gzip 17 05 10 X Yesterday v gt h20 09 12 10 v A Caneel Open After you installed the webapp you can directly access it at the address http localhost 8080 servlex hello world This is a very simple application The home page contain
44. around fifty although some are used only for a single translation The build properties file for a translation which is just a text file specifies the stages and their order We cannot pretend to have handled every possible SFM variant but it is possible to go a long way by mixing and matching the stages that we have so far defined Example 1 Sample build properties file Project code project NBS Source suffix suffix NS2 153 EPUB Chapter and Verse Major language lang fr Identifier identifier Bible fr NBS 2002 Glossary file glossary extras indx xml OSIS stages to process stages accents lang fr structure section title chapter chapter single verse start notes no fm note xref verse join line group table psalm title other line line break paragraph inline glossary after verse end split at verse note into verse cleanup EPUB volumes to produce epubs NBSNoDeutNoNotes NBSNoDeutNotes epub include notes yes Contents XML file contents basedir converted prot contents xml build properties also defines any parameters to be passed to the stage stylesheets We ve yet to find a convenient way to dynamically pass arbitrary parameters to An s lt xs1t gt task so to avoid hardcoding or limiting the parameters that can be passed to a stage when you modify build properties you then run ant to remake the build xml file that controls Ant so that it uses the cu
45. as 16 JSON for XForms lt exml anonymous xmlns exml http www agencexml com exml xmlns gt lt a gt stringA lt a gt lt b gt lt ce gt stringC lt c gt lt d gt stringD lt d gt lt b gt lt exml anonymous gt and for default context Tatt e a equals stringA e b c equals stringC e b d equals stringD 3 2 3 JSON Names JSON names which cannot be used as XML names are replaced by _______ in the empty namespace An extra attribute is used to store the JSON name when necessary and the new XPath functions fullname and local fullname are created to return the same value as name and local name or the value of this attribute when present Example a 6 pb s A B d j already n underscores are represented as lt exml anonymous xmlns exml http www agencexml com exml xmlns xsi http gt www w3 org 1999 XMLSchema instance xmlns exsi http www agencexml com exi xmlns gt lt exml fullname a amp amp b gt A B lt gt lt exml fullname d ja gt already lt gt lt exml fullname gt underscores lt gt lt exml anonymous gt and for default context fullnamef a amp b equals A B e local fullname 1 equals a amp b e fullnamef d ja equals already e local fullname 2 equals d ja fullname JJ equals underscores 17 JSON for XForms
46. as the input for the transformation and it will take the output as the new page This variant only runs the transformation once when the page is loaded and therefore offers no means for changing the page after it has loaded 2 4 2 XOuery in the Browser plugin Last year we presented the latest release of our XOIB browser plugin 8 It offers the functionality of XOuery inside the browser Unlike the plugins presented before it is not just a box inside a browser but integrates seamlessly into the HTML DOM It is possible to manipulate the website using all updating XQuery operations insert replace remove rename It also allows subscribing event handlers to DOM events While XOIB combines XML technologies declarative programming and browser DOM interaction it suffers from being a plugin it needs to be manually installed and will only have limited availability 3 XOuery in the Browser JavaScript Edition API and Functions This section presents the API we suggest for programming browser applications In general most functions of XOuery 3 0 XOuery Update Facility and XOuery Scripting Extension are supported 195 XQuery in the Browser reloaded 3 1 An example Here is an example of a page containing XQuery code Example 6 Simple XQIB example lt html xmlns http www w3 org 1999 xhtml gt lt head gt lt script type text javascript src mxqueryjs mxqueryjs nocache js gt lt script gt lt script type applica
47. each variable must be non empty Therefore we add an explicit check fn exists to a where clause in XQuery The intermediate result represented by GRAPH_0 is limited by a FILTER expression as shown in Section 3 3 and the reduced set of possible bindings is assigned to SGRAPH 1 In the next step the result is sorted by the attribute col Finally the 88 Translating SPARQL and SQL to XQuery output is generated by the function formatSparq1XML which renders the result according to the SPAROL Query Results XML Format which is defined in 19 To make the code generated by the cross compiler more concise a number of custom XQuery functions are used As an example the source code of the function getSubj from the xqllib package is shown in Example 4 This function returns the identifier of a given node which can be obtained by reading the rdf 1ID attribute The source code of the remaining functions can be found in 10 Example 4 Source Code of the Custom Function getSubj declare function xqllib getSubj Ssubj as node as xs string return subj rdf ID Example 5 Translation Result of the SPARQL Code 2 Example to XQuery declare namespace chemistry http www xql2xquery org chemistry let doc chemistry fn collection chemistry let Sresult let GRAPH 0 for element in doc_chemistry rdf ID let value element xgllib getSubj Selement for value_col in xgllib getData chemistry Selem
48. future work 2 Background 2 1 Related Works We have found two similar programs BXE Browser Based XML Editor 4 and Xopus 6 BXE is a browser based XML WYSIWYG Editor which works on almost all Mozilla based editors on any platform However its development stopped in 2007 and in addition we have been unable to get it to run Client side XML Schema validation Xopus is a client side WYSIWYG XML editor It allows working with structured and complex content without need for technical knowledge through a graphical interface The author cannot break the XML structure or write content that does not conform to the W3C XML Schema Xopus is proprietary software and it has limited flexibility for embedding within other applications It does not process wildcards correctly and works only in IE and Firefox In addition its resource usage can be substantial 2 2 Schema Validation Algorithms H S Thompson and R Tobin 8 proposed a technique to convert W3C XML Schema content models to Finite State Automata FSA including handling of nu meric exponents and wildcards The authors presented three algorithms e Conversion to FSA Converts regular expressions to FSAs It has two stages converting particles and converting terms e Unique Particle Attribution Supplements the first algorithm to check the Unique Particle Attribution constraint e Subsumption Checks two FSAs to confirm that one accepts only a subset of what the oth
49. gt e where rw subtype e xs decimal or rw subtype this xs float or rw subtype this xs double declare option rw rule rw SubZeroFold e 0 gt e where rw subtype e xs decimal or rw subtype this xs float or rw subtype this xs double declare option rw rule rw ZeroSubFold 0 e gt e where rw subtype e xs decimal or rw subtype this xs float or rw subtype this xs double declare option rw rule rw AddEmptyFold e gt declare option rw rule rw SubEmptyFoldl e gt 223 Declarative XQuery Rewrites for Profit or Pleasure declare option rw rule rw SubEmptyFold2 e gt declare option rw rule rw MulEmptyFold e gt declare option rw rule rw DivEmptyFoldl e div gt declare option rw rule rw DivEmptyFold2 div e gt declare option rw rule rw IDivEmptyFoldl e idiv gt declare option rw rule rw IDivEmptyFold2 idiv e gt declare option rw rule rw ModEmptyFoldl e mod gt a declare option rw rule rw ModEmptyFold2 mod e gt declare option rw rule rw UnaryMinusEmptyFold gt declare option rw rule rw IfTrueFold if true then then else else gt gt then declare option rw rule rw IfFalseFold if false then then else else gt gt else declare option rw rule rw Boole
50. http xmlprague cz Institute for Theoretical Computer Science http iti mff cuni cz Ubigway s r o http www ubigway com ix Institute for Theoretical Computer Science i DA bac E Center of research in Computer Science and Discrete Mathematics funded by the Ministry of Education of the Czech Republic Established in 2000 current project approved for 2010 2011 Staff of 60 researchers include both experienced and young scientists ITI is a joint project of the following institutions Faculty of Mathematics and Physics Charles University Prague Faculty of Applied Sciences University of West Bohemia Pilsen Faculty of Informatics Masaryk University Brno Mathematical Institute Academy of Sciences of the Czech Repub lic Institute of Computer Science Academy of Sciences of the Czech Republic For more information see http iti mff cuni cz m Publication preprints are available in ITI Series http iti mff cuni cz series Sponsors Gold Sponsors Mark Logic Corporation http www marklogic com The FLWOR Foundation http www flworfound org Silver Sponsors oXygen http www oxygenxml com Syntea software group a s http syntea cz Bronze Sponsors Mercator IT Solutions Ltd http www mercatorit com Mentea http www mentea net lt oXygen gt Xml editor m MarkLogic MENTEA xi xii Preface This publication contains papers presented at XML Prague 2011 XML
51. into presentation oriented formats such as XSL FO And if you re converting XML into HTML for rendition in a browser it obviously makes sense to download the XML to the client and do the conversion there The language specification of course never uses the words browser or client but there are many features of the language that were designed with this environment in mind for example the fact that most run time errors are recoverable which mirrors the way that browsers handle HTML syntax errors So what went wrong Why didn t everyone immediately use the new capability The answer is that few web sites are interested in delivering content in a format that can t be displayed by every browser That meant that the first few years web developers were reluctant to serve XML because XSLT wasn t implemented on browsers such as IE3 which were still widely deployed Later they were reluctant because XSLT wasn t implemented or was poorly implemented on Netscape and 125 XSLT in the Browser Firefox Today nearly all modern desktop browsers support XSLT 1 0 and the level of compatibility is pretty good despite one or two glitches like the failure of Firefox to support the namespace axis and the very tardy support for the document function in Opera But there are still users using ancient browsers and why shouldn t they My TV is twenty years old and still meets my needs Meanwhile the younger generation who would be horrified by a
52. it during the editing Ext JS and other frameworks provide functionality for checking simple types e g Ext isNumber Ext isDate etc as of now we do not implement the W3C XML Schema simple types in detail 4 3 Updating the Document When an user selects an action the application should physically update the docu ment The real updates deserve thorough study and was out of scope for our work to date which only updated the DOM In this subsection we provide basic ideas and complications that could rise during the development In the case of attributes the naive approach is simple we add or delete the ap propriate attribute within the element and then re style the element with the new set of attributes This assumes that the presentational impact of the attribute change is limited to the sub tree dominated by the host element which need not be the case Element insertion and deletion actions are more likely to have wider impact Therefore after the insertion or deletion the parent element should be re styled 4 4 Limitations and Assumptions The algorithms are straightforward and provide necessary editing functionality However the proposed method has some shortcomings It s not possible to edit several elements or attributes at once or to rename elements or attributes or to specify a sequence of actions It s not possible to edit invalid documents or even to pass through an invalid state for example in order to move a required el
53. lib util import coroutine coroutine def receive nodes text list while True node yield text list append node xml encode return XML lt doc gt lt one gt lt a gt 0 lt a gt lt a gt 1 lt a gt lt one gt lt two gt lt a gt 10 lt a gt lt a gt 11 lt a gt lt two gt lt doc gt text list target receive_nodes text_list pushtree XML u a target send target close print text List 6 And much more Amara provides many facilities beyond those covered above such as XSLT 7 Akara Web Framework A system for writing REST friendly data services Functions can be written using the core library features perhaps as unit trans forms Apply simple wrappers to turn functions into RESTful services Akara runs as a repository of services and allows you to discover these using a simple GET Service classes have IDs independent from locations of individual service end points Built in facilities for Web triggers AKA Web hooks AKA Web hooks like DBMS triggers declaration that one event actuates another in this case HTTP re quests Modern multi process Web server dispatch engine for services 70 Akara Spicy Bean Fritters and XML Data Services 7 1 A simple complete Akara module For the basic set up of an Akara module one can start with echo py and then cus tomize accordingly The following is a complete module for which you can indicate a URL of an XML document and get from
54. library Furthermore this function is only available in HTML and therefore GWT and our library can currently not be used on XHTML pages 4 2 Selected Implementation Aspects When translating MXQuery using we initially encountered various compiler errors see Section 4 2 6 due to the library differences to either J2ME or J2SE We solved them by removing functionality and gradually reintroducing it until we could run some basic queries Once we had a version that compiled we added the integration with the browser rewrite added code to re enable the remaining missing features 4 2 1 Solution architecture Similar to the original XQIB plugin we use the Store abstraction as a means to en capsulate browser specific aspects as shown in Figure 1 From an MXQuery point of view the DOM is yet another store that supports full path navigation and updates This store is created when the b dom function is present in the code If other XDM instances are needed normal MXQuery stores are generated 4 2 2 Mapping the DOM to MXQuery tokens The DOM Document Object Model is the fastest way of accessing the data in a browser A website is mapped into a DOM tree when it is loaded which can then be accessed through JavaScript The XHTML DOM and the HTML DOM have some small differences XHTML is naturally namespace aware because it is derived from XML whereas HTML is not To make the DOM implementations compatible again the W3C has specified that
55. like below defines a lang attribute with the default value sen Example 1 Relax NG schema specifying a default attribute lt xml version 1 0 encoding UTF 8 gt lt grammar xmlns http relaxng org ns structure 1 0 xmlns a http relaxng org ns compatibility annotations 1 0 datatypeLibrary http www w3 org 2001 XMLSchema datatypes gt lt start gt lt element name test gt lt attribute name lang a defaultValue en gt lt empty gt lt element gt lt start gt lt grammar gt Adding this support had 3 parts 1 Add support in Jing to store default values in the Attribute patterns 2 Create a class that given a Relax NG schema provides easy access to the attribute default values 3 Create a parser component that integrates the default values support in Xerces so that all the processing is done automatically on any parsing These steps are detailed in the following sections below 2 1 Support for storing default values There was an alternate approach here Trang already creates an internal model that contains all the annotations so one possibility was to work on that level However that complicates the next step of providing access to the default values and the change in Jing is minimal 169 DITA NG A Relax NG implementation of DITA Jing does not store any annotation when it parses a schema so that needed to be changed to retain the default value and set it on the attribute pattern 2 2 Acc
56. lt xsl copy of select t opening chapter start gt lt xsl copy gt lt xsl apply templates select node mode current gt lt xsl copy gt lt xsl template gt lt xsl template match o chapter t is opening chapter start mode move up gt e XSLT was not a good fit for processing SFM highlight markup when the text also contained XML elements 3 From OSIS to EPUB 3 1 Anatomy of an EPUB To a first approximation an EPUB 7 is a Zip encoded archive containing some XHTML and CSS and a bit of metadata EPUBs typically have a epub suffix An EPUB has a fairly contrived signature achieved by requiring that the first file in the Zip archive is named mimetype contains only the text application epub zip and is included in the archive uncompressed The operating system can then check that a file is an EPUB by examining up to the first 58 bytes of the file In practice all EPUB Reading Systems that we have used can cope if the signature isn t exactly right EPUB text is usually a profile of XHTML 1 1 but it can be DTBook DAISY Talking Book Preferred Vocabulary and may include generic XML as XML islands Text may be styled using a defined subset of CSS2 The metadata is also XML The two most significant and variable metadata files are the Open Packaging Format OPF file 13 which contains Dublin Core metadata the manifest of files in the EPUB and the spine that lists their reading
57. mapping This is why I only had to declare the Atom namespace on the top feed element All the children picked up the default namespace until it got to the div element which redefined the default as XHTML which was then passed on to its p child You can create namespace declarations manually using the special NS prefix ns construct Just make sure it comes beyond any other type of child specified for that element This is useful when you have QNames in content e g gen erating XSLT or schema or SOAP or some other horror This courtesy does not apply to attributes If you don t declare an namespace attribute for an attribute it will have none e Structwriter also tries to be smart with strings versus unicode I still recommend using Unicode smartly when working with XML but if you get lazy and just specify something as a string Structwriter will just convert it for you e Notice the use of a generator expression line 25 to generate the multiple category elements 3 1 Generating XML and HTML gradually The above works well if you have are generating an XML document all at a go but that s not always the case Perhaps you are generating a huge document little by little Perhaps you are generating a document in bits based on processing of asyn 62 Akara Spicy Bean Fritters and XML Data Services chronous events In such cases you might find useful the coroutine or pseudo coroutine if you insist form of the structwri
58. never mentioned For limiting refresh duration dependencies between nodes and XPath expressions should be maintained and updated if necessary when the value of a node is changed As a consequence any XForms implementation needs an enhanced XPath en gine 2 3 Server Exchanges XForms can directly be used with native XML databases but can also be used with any other server architecture Every AJAX based application is sending data to its client software according to requests with parameters Only XML serialization and flat GET parameters list submission are specified in XForms 1 1 2 4 JSON vs XML for browsers There are technical facts to consider JSON is lighter no namespace no distinction between elements and attributes JSON is shorter minimal closing tags 14 JSON for XForms JSON objects are evaluated and stored by Javascript without any extra API being required JSON values are not just text strings numbers booleans and dates might be present too JSON natively supports any character in names supports named and anonymous arrays JSON doesn t natively support queries and external query libraries are not as rich as XPath just child or descendant axes XML processing is not standardized in browsers and require specific Javascript instructions IE8 is still using MSXML3 Even if XSLT is also well supported by browsers it cannot be used for dynamic treatments There are also human facts to consi
59. not aware of any client side schema validator when the project was started Also missing is the function fn normalize unicode because neither GWT nor any third party libraries provide support for it with reasonable code footprint At the time where this paper is written the newer main library module API is not yet implemented but we are working on it Samples are available at 5 4 Implementation 4 1 Background 4 1 1 MXQuery MXOuery is an XQuery engine written in Java with the goal of good portability and a small footprint targeting also mobile and embedded devices It supports a broad range of XQuery related standards including XQuery 1 0 the Update Facility Fulltext and parts of XQuery 3 0 and Scripting It is implemented in Java and available as open source with an Apache 2 0 License The design of MXQuery is inspired by query processing database architectures using an iterator style query plan generated by a parser optimizer frontend Therefore it is well suited for streaming execution keeping the runtime memory needs low unless blocking or nested operations require materialization 198 XQuery in the Browser reloaded 4 1 1 1 Tokenstream based Data Model MXOuery uses a Token model for its XDM representation similar to the BEA XORL XOuery engine Tokens are conceptually very similar to parse events SAX StAX but are objects which directly carry all relevant information instead of a parse type identifier 4 1
60. of the other module The location is determined by the argument of the augment statement As a matter of fact it is expected that augment will be one of the main methods for building more specific YANG modules from generic components For instance 1 shows the following example in which a module for general network interfaces is augmented with new nodes specific for Ethernet interfaces import interfaces prefix if augment if interfaces if interface when if type ethernet container ethernet leaf duplex As an example of a complete YANG module which however demonstrates only a small subset of YANG features below is a data model for the imaginary coffee machine that was used in the examples in Section 2 Note that in order to keep the example short we do not follow all the recommendations for structure and contents of YANG modules stated in 12 module coffee machine namespace http example com coffee prefix cm import ietf inet types 105 Configuring Network Devices with NETCONF and YANG prefix inet description This is an example YANG module for an Internet enabled coffee machine organization Example Ltd contact R A Bean lt bean example com gt revision 2010 03 26 description Initial and final revision container configuration container ipv4 leaf address mandatory true type inet ipv4 address leaf subnet mask length type uint8 range 0 32
61. order if you page through the document and the Navigation Control File NCX file 11 which functions as the table of contents for the EPUB 159 EPUB Chapter and Verse Converting OSIS to EPUB is dazzlingly simple by comparison to producing OSIS there are only three stylesheets for converting OSIS to XHTML plus a handful of other stylesheets for generating the OPF and NCX files whereas the other metadata files are static files We have provided links in osisbyxsl to use epubcheck 8 an open source EPUB validator Current translations result in one to four EPUBs each typically Protestant Cath olic canon with without notes The resulting zipped EPUB weighs in at between 3 5Mb and 6Mb depending on the variant Because of the length of some biblical books we opted for one XHTML file per chapter and one for each chapter s notes as well as a cover copyright page contents pages and a glossary In total there are 831 XHTML files in the Catholic version of the Bible en Francais Courant Skeleton NCX and OPF files for an EPUB are populated based on the specific hierarchy of books for that edition and the set of notes and graphics that are to be included Translations may also include introductions and back matter as well as diagrams photos maps and explanatory notes that are interspersed in the text of the Bible We are calling all of these external notes since we have them marked up in OSIS with graphic files for diag
62. package google apis an XSLT and XQuery library to access Google APIs over HTTP B l Qr 2 4 Je Web cxan org beta pkg google apis v 3 vw Search with Google Home News Packages Authors Categories Tags google apis W ID google apis Name http fgeorges org lib google apis Title Google APIs for XSLT and XQuery Author Florent Georges Categories Web API Tags web api google library 0 3 0 E File google apis google apis 0 3 0 xar Dependency http expath org lib http client The client is invoked from the command line although a graphical or web front end could be written It understands a few commands in order to find a package install it or remove it in the local repository The following screenshot shows how to look for packages with the tag google There is one the package with the ID google apis We then display the details for that package We also search for an HTTP Client implementation then install it before installing the Google APIs All informations and packages are retrieved directly from the CXAN website 37 CXAN a case study for Servlex an XML web framework gt cxan tag google tags google subtags library web api google apis Google APIs for XSLT and XQuery gt cxan show google apis ID google apis Package http fgeorges org lib google apis Description Google APIs for XSLT and XOuery Author fgeorges Categories web api Tags web api google library 3
63. pattern images jpg media type image jpg gt lt resource pattern images png media type image png gt So far so good But what happens when the component encounters en error For instance if the user sended wrong parameters or the database was not reachable In that case the component just throw an error using the standard XPath function fn error or the XProc step p error This error is caught by Servlex which passes it to an error handler configured in the web descriptor which formats a nice error page for the user as well as the corresponding web response in particular the HTTP status code of the response for instance 404 or 500 Thanks to the standard XPath error mechanism the error itself can carry all the information needed the CXAN 49 CXAN a case study for Servlex an XML web framework application allows the error item to contain the HTTP status code to use as well as a user friendly message All the logic is thus implemented by XProc components They talk to eXist using its REST like API and p http request The following is an abstract view of the most complex pipeline in the website the Package component HTTP Request C GET PUT save the body in a temporary dir get package in DB format to a page create empty entry in DB consistent No move the file in its final destination insert new version in DB descriptors 000 HTTP Response A flowcha
64. processors Most XML databases provide such a feature usually trying to provide an API at a slightely higher level sacrificing the power of a full HTTP support Its place in the XML eco system is similar to the place of the Servlet technology in the Java eco system quite low level but providing the ability to build more complex systems on top of it entirely in XML Servlex is an open source implementation of the Webapp Module based on Saxon and Calabash as its XSLT XQuery and XProc processors and on the Java Servlet technology for its networking needs It can be installed in any servlet con 38 CXAN a case study for Servlex an XML web framework tainer like Tomcat Glassfish or Jetty It is available on Google Code at http code google com p servlex 2 The Webapp Module The overall treatment of an in bound HTTP request is as follows in the Webapp Module S a mel l That is the client sends a request It is received by the webapp container It is translated to an XML representation by the HTTP Bridge This XML representation is a simple XML vocabulary giving information about the HTTP verb the request URI the URI parameters the HTTP headers and the entity content e g in case of a PUT or a POST Based on the request URI and on a set of maps the Dispatcher finds the component to call in order to handle the request Once the correct component is found it is called with the request as parameter For
65. quality of the XQuery execution does not change much MySQL s query processing and storage engines are well tuned for such simple retrieval queries but sees a strong competition from the XML databases with Sedna catching up at scale factor 5000 100000 u 9 E 10000 g t ARQ 1000 Saxon o 5 100 Zorba 5 QizX z 10 lt BaseX 5 i Sedna 10 50 100 500 1000 5000 scaling factor Figure 1 BERLIN SPARQL Query 1 10000 a E 1000 N v MySQL 100 Saxon o 5 10 Zorba x 0izX 1 BaseX gt Sedna 0 1 T T T T T 1 10 50 100 500 1000 5000 scaling factor Figure 2 BERLIN SOL Ouery 1 For a query with an OPTIONAL clause Figure 3 the XQuery translation from SPAROL requires a join and duplicate elimination which is currently implemented using nested loop We measured a significant difference in the quality of the XQuery op 92 Translating SPARQL and SQL to XQuery timizers Saxon and BaseX seem to exploit more optimization possibilites than the other XQuery implementation and scale in the same way as ARQ but maintain a lead of more than one order of magnitude The other XQuery systems scale somewhat worse but still maintain a measureable lead at scale 5000 For the translation from SQL Figure 4 the simpler query structure gives all optimizers the chance to detect the relevant information leading to almost the same pe
66. re being offered and we have very little choice but to accept it There s very little scope for those outside the charmed circle of browser vendors to offer an alternative and as a result there s very little questioning of whether what s on offer is really the best we can do or not But I think we can do better and that s what I hope to demonstrate in the next two sections of this paper 2 The Present During XML Prague 2010 anumber of speakers made complaints about the limita tions of browsers similar to the complaints in the previous section not just in relation to XSLT but also XQuery XForms XProc and other technologies A presentation on XQuery in the browser 2 attracted a lot of interest though sadly there is only a skeleton paper in the proceedings That implementation used the plug in mechanism offered by many browsers which in practice has many limitations it doesn t offer seamless ability to manipulate the HTML of the page as well as other components of the browser space such as the address bar status bar history and cookies A common observation made in discussion at last year s conference was that if you implemented a processor for your favourite language in Javascript all these problems would go away Javascript as a language has matured immensely in the last few years and browser vendors fall over each other trying to excel in the per formance of the Javascript engines There appeared to be a consensus forming th
67. several extensions offer this functionality We use the EXSLT function dyn evaluate for this purpose 5 Conclusions In this paper we gave an overview of the recent efforts of the IETF working groups NETCONF and NETMOD concentrating especially on the issues of data modelling using the YANG language The discussion of the mapping from YANG to DSDL allows to compare YANG with the existing XML schema languages The NETCONF protocol is now supported by numerous vendors of networking devices including a few big players The NETCONF WG web page lists about a dozen independent implementations Also the YANG data modelling language has already been used in several data modelling projects despite the fact that the standard 10 was published only few months ago For instance YANG has been used by the IPFIX WG of the IETF for modelling configuration and state data of various devices producing and processing IP traffic flow data 3 The success of YANG and to some extent the NETCONF protocol will be ulti mately determined by the willingness of the IETF community to adopt these tech nologies developing new as well as re implementing old configuration and state data models To aid this adoption the NETMOD working group was rechartered in 2010 with the main aim of creating several basic YANG data models that other IETF working groups can use to build their specialized models upon So far the first draft of a model for network interfaces was publis
68. some data formats where XML is more natural than JSON Probably the key advantage of MLJSON is that it works well as a cloud service An enterprise system can use MLJSON to store retrieve and query JSON and not need to know anything about the internal details The MLJSON system would just appear like a decoupled scalable high performant text indexed JSON store 34 CXAN a case study for Servlex an XML web framework Florent Georges H20 Consulting lt fgeorges fgeorges org gt Abstract This article describes the EXPath Webapp Module a standard framework to write web applications entirely with XML technologies namely XQuery XSLT and XProc It introduces one implementation of this module Servlex It uses the CXAN website the Comprehensive XML Archive Network as a case study Keywords EXPath webapp XProc XSLT XQuery 1 Introduction The EXPath project defines standard extensions for various XPath languages and tools Most of them are extension function libraries defining sets of extension functions you can call from within an XPath expression e g in XLST XProc or XQuery like the File Module the Geo Module the HTTP Client and the ZIP Module resp functions to read write the filesystem functions for geo localisation a function providing HTTP client features and functiond to read write ZIP files EXPath also defines two modules of a different nature the Packaging System and the Webapp Module The Packaging Sys
69. strong gt lt quote gt lt name ak rel name gt Thomas Eliot lt name gt lt address ak rel place ak value concat city province gt lt street gt 3 Prufrock Lane lt street gt lt city gt Stamford lt city gt lt province gt CT lt province gt lt address gt lt opus year 1932 ak rel name ak resource gt lt title ak rel name gt The Wasteland lt title gt lt opus gt lt tag eg occurs ak rel name gt old possum lt tag gt lt label gt lt labels gt labelmodel examplotron_model MODEL A INSTANCE A 1 lt labels gt lt label id co added 2004 11 15 gt lt name gt Christopher Okigbo lt name gt lt address gt lt street gt 7 Heaven s Gate lt street gt lt city gt Idoto lt city gt lt province gt Anambra lt province gt lt address gt lt opus gt lt title gt Heaven s Gate lt title gt lt opus gt lt tag gt biafra lt tag gt 67 Akara Spicy Bean Fritters and XML Data Services lt tag gt poet lt tag gt lt label gt lt labels gt doc bindery parse INSTANCE A 1 model labelmodel for triple in generate metadata doc to be print triple 0 triple 1 U triple 2 Triples but only RDF if you want it b The output is u co u name u Christopher Okigbo u co u place u Idoto Anambra u co u opus u r2e0ele5 u r2e0ele5 u title u Heaven s Gate u co u tag u biafra u co
70. subtree rather than the entire document This validation is also available after mutation with the Amara API Validation can be a bit expensive though not noticeably unless you re dealing with huge docs so it should be used judiciously The penalty is only paid upon actual validation Mutation document access and other operations proceed at regular speed 65 Akara Spicy Bean Fritters and XML Data Services With a somewhat irregular XML document it can be tricky to use bindery object traversal e g doc labels label without risking AttributeError A model used in parsing a document makes the binding smarter setting a default value to be returned in cases where a known element happens to be missing somewhere in the instance document LABEL MODEL lt xml version 1 0 encoding utf 8 gt lt labels gt lt label gt lt quote gt What thou lovest well remains the rest is dross lt quote gt lt name gt Ezra Pound lt name gt lt address gt lt street gt 45 Usura Place lt street gt lt city gt Hailey lt city gt lt state gt ID lt state gt lt address gt lt label gt lt labels gt TEST LABEL XML lt xml version 1 0 encoding utf 8 gt lt labels gt lt label gt lt name gt Thomas Eliot lt name gt lt address gt lt street gt 3 Prufrock Lane lt street gt lt city gt Stamford lt city gt lt state gt CT lt state gt lt address gt lt label gt lt labels gt f
71. tell us whether they are referencing the same node no matter how we re trieved the references To compare by document order the DOM nodes can also be used The DOM Level 3 Core offers the function compareDocument Position which allows to compare two nodes directly Where this interface is not available we can fall back to an alternative method based on the Lowest Common Ancestor LCA The LCA can be found by walking through the tree up to the root while memoizing intermediate nodes From the root node we then compare the intermediate node lists of the two nodes The LCA is the last node which we can find in both lists If one of the nodes is the LCA we know the document order Otherwise we can look at the children of the LCA and determine which path comes first Using the nodes themselves as IDs we can reduce runtime overhead and when compareDocument Position is available we also have a very efficient way of ordering nodes 202 XQuery in the Browser reloaded 4 2 4 Applying Updates When we have an updating query we get as result of the query a PUL Pending Update List containing all updates that need to be applied Performing these up dates on DOM5 HTML in a correct and efficient way needs some special consider ations Deletion is the most straightforward primitive since we can just call the node s removeFromParent function For insertion the token stream needs to be converted into DOM nodes It is very important not to inse
72. that everything matching the wildcard in pkg must be given the name id so it can be accessed in the request by the XPath web request web path web match name eq id The URI query parameter and the HTTP request headers are also easily accessible by name The entity content aka the request body if any is also passed to the component The bodies though are passed a bit differently Instead of being part of the request document the bodies are passed in a separate sequence I say bodies because in case of a multi part request we can have several of them They are parsed depending on their content type so a textual body is passed as a string item an XML content is parsed as a document node an HTML content is tidied up and parsed in a docu ment node and everything else is passed as a base 64 binary item A description of each body is inserted in the web request though describing its content type and a few other infos The component is called with the request document and must provide as result a response document The same way the request document represents the HTTP request the client sent the response document represents the HTTP response to send back to the client It looks like lt web response status 200 message 0k gt lt web header name Extra Header value gt lt web body content type text html gt lt html gt lt head gt lt title gt Hello lt title gt lt head gt lt body gt lt p gt Hello
73. then its access endpoint will be http localhost 8880 The user can HTTP POST some data to this URL and the decorated function will be invoked akara echo body body ctype ecounter is the decorated function Simple service implementation functions wrapped as HTTP POST methods receive the HTTP POST body and the HTTP Content Type header as parameters The latter is a convenience All the other HTTP headers are also available using WSGI more on this later The following version demonstrates some basic security features import amara from akara services import simple service respons http github com zepheira akara tree master lib demo echo py 71 Akara Spicy Bean Fritters and XML Data Services ECOUNTER SERVICE ID http purl org akara services demo element counter Config info is pulled in at global scope as AKARA MODULE CONFIG Security demo create a URI jail outside of which XML operations won t leak URI JAIL AKARA MODULE CONFIG get uri jail Create the assertion rule for the URI jail ALLOWED lambda uri base baseuri uri startswith URI JAIL True Create a URI resolver instance that enforces the jail restricted resolver irihelpers resolver authorizations ALLOWED simple service GET ECOUNTER SERVICE ID ecounter text plain def ecounter uri e g curl http localhost 8880 ecounter uri http hg akara info b testd
74. to learn a new syntax for querying JSON within XForms New XPath extension functions should be enough to guaranty a full support Whilst XSLTForms has its own XPath engine other XForms implementations using an external library as XPath engine should just add functions Whenever extra elements are required for internal storage their number should be minimal and their names should not be in conflict with ordinary JSON names XPath functions name and local name should return the effective JSON names when possible Javascript notation to retrieve an item in a array pos where pos starts from 0 should be almost preserved pos where pos starts from 1 3 1 3 XML Schema conformance XML Schema Recommendation already defines the xsi type and the xsi nil attributes and they are supported in the XForms Recommendation The xsd maxOccurs attrib ute enables to specify how many occurences there might be for an element 3 2 Proposed Representation of JSON Objects With XML 1 0 3 2 1 Elements Attributes and Namespaces Elements are used for JSON properties within the empty namespace Meta data are stored in attributes within a specific namespace Extra elements are only used in anonymous situations and are always in a spe cific namespace 3 2 2 Extra Document Element XML 1 0 requires a unique document element so such an element is always added Example a stringA bz c stringC d stringD would be serialized
75. utf 8 gt lt labels gt lt label gt lt name gt Thomas Eliot lt name gt lt address gt http www ibm com developerworks xml library x xmptron 64 Akara Spicy Bean Fritters and XML Data Services lt street gt 3 Prufrock Lane lt street gt lt city gt Stamford lt city gt lt state gt CT lt state gt lt address gt lt label gt lt labels gt The following is an XML document that does not conform to the schema INVALID LABEL XML lt xml version 1 0 encoding utf 8 gt lt labels gt lt label gt lt quote gt What thou lovest well remains the rest is dross lt quote gt lt name gt Ezra Pound lt name gt lt address gt lt street gt 45 Usura Place lt street gt lt city gt Hailey lt city gt lt state gt ID lt state gt lt address gt lt label gt lt labels gt The quote element is not in the model One specifies the XML model to use when parsing to Bindery from amara bindery model import label model examplotron_model LABEL MODEL doc bindery parse VALID LABEL XML model label model doc xml validate doc bindery parse INVALID LABEL XML model label model try doc xml validate except bindery BinderyError e print e doc xml write Parse INVALID LABEL XML succeeds but the xml validate method fails and raises an exception because of the unexpected quote element Note it s no problem to validate just an element s
76. version of the Micro Edition J2ME the Standard Edition and newer Google proposals such as Android Dalvik Google App Engine and GWT MXQuery aims to use a subset of language expressions and library classes that is broadly available At the beginning of the XOIB JS project MXOuery had an abstraction layer that hid the differences between J2ME J2SE 1 4 and J2SE 1 5 1 6 As it turned out Google has chosen to make GWT a different Java subset than J2ME so several of the abstractions were not sufficient 199 XQuery in the Browser reloaded 4 1 2 Google Web Toolkit GWT Fundamentals GWT provides a Java API as a subset of J2SE which exposes several browser fea tures most prominently DOM access to the browser The application and the re quired libraries need to be available as Java source code This Java code is then translated into Javascript creating by default a monolithic code file Since the available functionality varies from browser to browser several different versions of the code are generated At runtime a small Javascript file projectname nocache js has to be included in the Web site which will detect the browser type and download the actual application Javascript code file encapsulated in an HTML file This file is loaded into an IFrame using an XMLHttpRequest and inserted into the current page using the JavaScript function document write Therefore the same origin policy applies preventing central hosting of an GWT based
77. we utilize the the runtime function call mechanism of MXQuery treating the pre compiled query like a module 4 2 6 Compatibility issues 4 2 6 1 Missing functionality in GWT For many browser functions GWT only supports the functionality available on all browsers This reduces many functions for DOM access to the level of Internet Ex plorer 6 This forced us to rewrite or extend DOM classes like Node or Element with functionality such as element setAttributeNS or element localName 4 2 6 1 1 Missing Java Functions In GWT we do not have a Java VM Virtual Machine thus the class loading ability is missing The mainline MXQuery relies heavily on on demand loading of classes for functions and operators in particular for extensibility We needed to hardcode all functions as a big Java file transformed from the XML metadata files used in MXOuery Another missing area are I O classes There is no implementation of OutputStream PrintStream etc available so we had to include these classses from the Apache Harmony project Furthermore noCalendar class is included in GWT This class backs most date and time operations in MXQuery The Calendar classes from the Apache Harmony project depend on IBM s internationalization library ICU which is nearly impossible to port to GWT Therefore we first used a third party Calendar emulation gwt calendar class which uses the GWT internationalization features to provide the full functionality pro
78. you can use something like the following command line to add a poet curl request POST data binary http localhost 8880 newpoet lt lt END lt newpoet id co gt lt name gt Christopher Okigbo lt name gt lt address gt Christopher Okigbo lt address gt lt newpoet gt END You can verify the result easily enough by querying the low level database file gt gt gt import shelve gt gt gt d shelve open tmp poet gt gt gt print d keys co 78 Akara Spicy Bean Fritters and XML Data Services gt gt gt print d co u Christopher Okigbo u Christopher Okigbo Note This tutorial uses shelve for simplicity but for real world applications you almost certainly want to use another persistence back end such as sqlite Also these examples are not safe for concurrent access from multiple module instances which is just about guaranteed for a real world application 11 Conclusion Akara s design makes it easy to integrate with other persistence facilities from rela tional to state of the art DBMS and certainly modern cloud style storage services It has seen a wide variety of use with mixed and matched components whether incorporating Web based transform and validation services or attaching modern visualization systems One way to fulfill sophisticated XML driven database require ments is to use monolithic software but another important approach is to stitch together loosely coupl
79. 0 Version 0 Available 0 2 0 0 3 0 Depends on http expath org lib http client gt cxan search http expath http client exist EXPath HTTP Client for eXist expath http client saxon EXPath HTTP Client for Saxon gt cxan install expath http client saxon gt cxan install google apis gt I Besides those two tools the website and the client the most valuable part of CXAN is the collection of packages itself CXAN is not a brand new idea and is similar in spirit to systems like Debian s APT system and its famous apt get com mand CTAN for TeX and LaTeX or CPAN for Perl also with a website at http cpan org and a client to look up and install packages locally 1 2 Webapp and Servlex The EXPath Webapp Module defines a web container using XSLT XQuery and XProc to implement web applications It defines how the HTTP reguests are dis patched to those components based on a mapping between the reguest URI and the components It also defines how the container communicates with the components basically by providing them with an XML representation of the HTTP request and by receiving in turn an XML representation of the HTTP response to send back to the client The purpose of this module is to provide the developer with a low level yet powerful way to map HTTP requests to XML components without need for any other technology It is defined independently on any processor and can actually be implemented by all kind of
80. 20 year old TV set expect to do their internet surfing from smartphones many of which don t yet support XSLT So the problem throughout this period has been the same XSLT penetration on the browser has not at any stage been ubiquitous enough to encourage a significant level of adoption by web site developers And this of course led to the second problem because web developers weren t using XSLT in large numbers the technology didn t progress The failure of browser vendors to implement XSLT 2 0 is widely bemoaned by XSLT aficionados but who can blame them If we paid for our software things might be different So long as it s free browser vendors have no incentive to cater to minorities They are driven by the monthly figures on market share and as a result they have to cater to the masses But it s not really the failure to implement XSLT 2 0 that s the biggest issue it s the failure to modernize the role of XSLT and its relationship to the rest of the browser environment which hasn t exactly stood still since 1999 The browser in 1999 was still largely what its name indicates a way of displaying static content on a PC screen and navigating between pages by following hard coded hyperlinks The way XSLT 1 0 is implemented in the browser reflects that level of maturity in the evolution of the web It translates XML to HTML and hands the HTML over to the browser to be displayed period What s the story on forms on interaction on AJAX
81. 3 Springer Verlag Berlin Heidelberg 4 Liip AG Bxe the wysiwyg xml editor https fosswiki liip ch display FLX Editors BXE May 2007 August 19 2010 5 T O Reilly and J Battelle Web squared Web 2 0 five years on In Web 2 0 Summit O Reilly Media Inc and TechWeb 2009 6 SDL Structured Content Xopus The web based wysiwyg xml editor http xopus com March 2010 Retrieved August 17 2010 7 H S Thompson Efficient implementation of content models with numerical occurrence constraints In XTech 2006 Amsterdam The Netherland May 2006 IDE Alliance 8 H S Thompson and R Tobin Using finite state automata to implement W3C XML schema content model validation and restriction checking In Proceedings of XML Europe London 2003 IDE Alliance 9 H S Thompson and R Tobin Xsv an xml schema validator http www cogsci ed ac uk ht xsv status html December 2007 Retrieved January 10 2011 10 W3Schools Xml schema tutorial http www w3schools com schema default asp 2010 Retrieved December 20 2010 12 JSON for XForms Adding JSON support in XForms data instances Alain Couthures lt alain couthures agencexml com gt Abstract XForms was originally specified for manipulating XML only instances of data but designing forms with XForms can benefit from JSON support espe cially for integration with existing AJAX environments This is possible when defining how to map any JSON o
82. 3 2 4 JSON Datatypes For each Javascript datatype the most approaching XSD type is automatically asso ciated XForms bindings have to be used to adjust more precisely the effective datatype Example a string A br42 c new Date 2011 3 26 d true would be serialized as lt exml anonymous xmlns exml http www agencexml com exml xmlns xsi http p www w3 org 1999 xMLSchema instance xmlns gt lt a gt stringA lt a gt lt b xsi type xsd double gt 42 lt b gt lt c xsi type xsd dateTime gt 2011 03 26T00 00 00Z lt c gt lt d xsi type xsd boolean gt true lt d gt lt exml anonymous gt 3 2 5 JSON Named Arrays Arrays are modeled with an extra attribute Empty arrays reguire another attribute because if not there would be an ambiguity for an array with just the empty string as element Extra XPath functions might be helpful e is array node which might be defined as count node exsi maxOccurs un bounded 0 e is non empty array node which might be defined as count node exsi maxOccurs unbounded and xsi nil true 0 e array length node which might be defined as count node exsi maxOccurs unbounded and xsi nil true Example a stringA 42 b C wee would be serialized as 18 JSON for XForms lt exml anonymous xmlns exml http www agencexml com exml xmlns xsi http p www w3 org 1999 xMLSchema instance xmlns ex
83. Devices with NETCONF and YANG lt value gt milk lt value gt lt value gt cof fee lt value gt lt choice gt lt element gt lt optional gt lt element gt lt nma notification gt lt nma notifications gt lt start gt lt define name coffee machine state supply levels percent nma default 0 gt lt data type unsignedByte gt lt param name minInclusive gt 0 lt param gt lt param name maxInclusive gt 100 lt param gt lt data gt lt define gt lt grammar gt lt start gt lt define name ietf inet types ipv6 address gt lt data type string gt lt param name pattern gt trimmed lt param gt lt data gt lt define gt lt define name ietf inet types ipv4 address gt lt data type string gt lt param name pattern gt trimmed lt param gt lt data gt lt define gt lt grammar gt A 2 RELAX NG Schema lt xml version 1 0 encoding utf 8 gt lt grammar xmlns http relaxng org ns structure 1 0 xmlns cm http example com coffee xmlns nma urn ietf params xml ns netmod dsdl annotations 1 datatypeLibrary http www w3 org 2001 XMLSchema datatypes ns urn ietf params xml ns netconf base 1 0 gt lt include href home lhotka Development pyang schema relaxng lib rng gt lt start gt lt element name rpc reply gt lt ref name message id attribute gt lt element name data gt lt interleave gt lt grammar ns http examp
84. Florescu Oracle lt dana florescu oracle com gt Ghislain Fourny ETH Zurich lt gfourny inf ethz ch gt Donald Kossmann ETH Zurich lt donaldk inf ethz ch gt Abstract Over the years the HTML based Web has become a platform for providing applications and dynamic pages that have little resemblance to the collection of static documents it had been in the beginning This was made possible by the introduction of client side programmable browsers Because XML and HTML are cousins XML technologies can be almost readily adapted for client side programming In the past we suggested to do so with XQuery and imple mented it as a plugin However using a plugin was seen as an insurmountable obstacle to a wider adoption of client side XQuery In this paper we present a version of XQuery in the Browser without any plugin needing only JavaScript to interpret XQuery code This enables use even on mobile devices where plugins are not available Even though our current version is still considered to be at an alpha stage we were able to deploy it successfully on most major desktop and mobile browsers The size of the JS code is about 700KB By activating compression on the web server reducing the transfered data to less than 200 KB as well caching on the client using the XQuery engine does not cause noticable overhead after the initial loading 191 XQuery in the Browser reloaded In addition we are already reaching a large level of compl
85. HTML page to return to the client The function looks like the following lt xsl function name app hello xslt gt lt the representation of the http request given by servlex gt lt xsl param name request as element web request gt lt xsl param name bodies as item gt lt compute the message based on the param who gt lt xsl variable name who select Srequest web param name eg who value gt lt xsl variable name greetings select concat Hello Swho gt lt first return the description of the http response gt lt web response status 200 message 0k gt lt web body content type application xml method xhtml gt lt web response gt lt then return the body of the response an html page gt lt html gt lt head gt lt title gt lt xsl value of select Sgreetings gt lt title gt lt head gt lt body gt lt p gt lt xsl value of select Sgreetings gt lt xsl text gt in XSLT lt xsl text gt lt p gt lt body gt lt html gt lt xsl function gt The sequence returned by the function here an element web response and a HTML element is used by Servlex to send a response back to the client with the code 200 Ok the content type application xml and the HTML page as payload The application source code is structured as follows but this is up to the developer hello world src hello xproc
86. ITI Series Institut Teoretick Informatiky Institute for Theoretical Computer Science 2011 519 xXMiprague XML Prague 2011 Conference Proceedings Institute for Theoretical Computer Science ITI Charles University Malostransk nam sti 25 118 00 Praha 1 Czech Republic http iti mff cuni cz series XML Prague 2011 Conference Proceedings Copyright 2011 Jiri Kosek Copyright 2011 MATFYZPRESS vydavatelstvi Matematicko fyzikalni fakulty Univerzity Karlovy v Praze ISBN 978 80 7378 160 6 s XMIprague XML Prague 2011 Conference Proceedings Lesser Town Campus Prague Czech Republic March 26 27 2011 lt oXygen gt XML Editor Content Authoring lt oXygen gt XML Author is the e Visual XML Authoring E DITA and DocBook Ready e Single Source Publishing e CMS Integration e Highly Configurable and Extensible The Complete XML Development and Authoring Solution XML Development lt oXygen gt XML Editor is an ed nt Intelligent XML Editing Visual Schema Modeling XSLT and XQuery Debugging XML Databases Integrated Tools www oxygenxmi com AVENEL SY lt oXygen gt is available in two editions e lt oXygen gt XML Author for the content authors starting from 199 USD e lt oXygen gt XML Editor for developers containing the complete development environment starting from 64 USD Academic 349 USD Professional Both editions can run as a standal
87. L database to maintain the metadata about packages here eXist but that could be any one usable from XProc like Qizx MarkLogic and many others a set of XProc pipelines each one implementing a particular page of the website and a site wide transform to apply a consistent layout accross all pages 47 CXAN a case study for Servlex an XML web framework P Dispatcher j f SS Tags FA Page lt page gt Standard XPath error lt title gt CXAN lt ttitle gt lt para gt error code QName Welcome to CXAN sage What happens when the user points his her browser to the address http cxan org The HTTP request hits Servlex which knows this request is aimed at the CXAN web application In the general case Servlex can host several applications each of them with its own name In the case of CXAN the Servlex instance is dedic ated to the single CXAN application It gets the web descriptor for CXAN and looks for a component matching the URI It finds the Home component based on the following element in the descriptor lt servlet name home gt lt xproc uri http cxan org website pages home xproc gt lt url pattern gt lt servlet gt Servlex then builds the web request element with the relevant information the headers the request URI the raw URI as well as a parsed version presenting the domain name the port number the parameters etc It uses Cal
88. L markup OSIS allows for two potentially overlapping structures Document structure BSP and verse structure BCV Document structure is dominated by book sections and paragraphs BSP additionally with titles guotes and poetic material While verse structure is indicated by book chapter and verse numbers BCV Because these two systems can overlap and because XML does not allow for overlapping elements OSIS defines a milestone mechanism for both document and verse structure elements 10 The milestone approach involves using empty elements to make the start and end of each structural item rather than the more intuitive approach of enclosing content within an element e g lt chapter sID Gen 1 1 gt content lt chapter eID Gen 1 1 gt rather than lt chapter sID Gen 1 1 gt content lt chapter gt 162 EPUB Chapter and Verse Milestones do solve the overlapping structures problem but they are much less convenient to work with than nested enclosed content The semantics of milestones are hard to represent in Schema An XML validator cannot validate whether OSIS milestones are used properly It cannot validate e that an element is consistently either milestoned or not e that for each element with an sID that there is a paired element with an eID e that each paired sID eID have the same attribute value 10 Processing milestones is relatively straightforward when the whole document is
89. N is a new open source project that provides a set of libraries and REST endpoints to enable the MarkLogic Server to become an advanced JSON store Behind the scenes the JSON is represented as XML and the JSON centric queries are resolved using MarkLogic s XML centric indexes In this paper we present the design of the project discuss its pros and cons and talk about the interesting uses for a fully queryable highly scalable JSON store Note The MLJSON project is in pre release and details are subject to change Source code is available at https github com isubiker mljson 1 JSON and MarkLogic Server A quick refresher JSON stands for JavaScript Object Notation It s a lightweight data encoding and interchange format that s native to JavaScript but now widely utilized across programming languages It s commonly used for passing data between web servers and web browsers specifying configuration data and exchan ging data between decoupled environments MarkLogic Server is a document centric transactional search centric structure aware schema agnostic XQuery and XSLT driven high performance clustered database server MarkLogic uses XML as a native data type with indexes optimized to run ad hoc queries against XML documents with ad hoc schemas By putting a JSON facade on top of MarkLogic Server and having MarkLogic store the JSON internally as an XML representation it s possible to have a JSON store with all the c
90. OSIS and from OSIS to XHTML for inclusion in the EPUB is entirely by XSLT 2 0 Additional steps such as copying graphic files and zipping up the EPUB are done using Ant We stuck with just XSLT 2 0 so the eventual users of the open source project only need to know one technology and if they re to be generating OSIS XML they should be developing some familiarity with working with XML markup Every so often we consider doing parts of it in Perl but that decreases the pool of people who can work on the whole project Bible experts who know XSLT 2 0 are going to be rare but Bible experts who know both XSLT 2 0 and Perl are going to be even rarer We used Ant to run the processing because it is widely available it can be con figured using text files and it s useful for non XML actions such as copying graphics files and zipping directories Ant enables us to choose which parts of the processing need to be run A complete build and test sequence can be invoked using ant clean process validate epubs epubs check This sequence takes a couple of minutes to produce and verify up to four variants of a single set of SFM files 1 2 The source files SFM stands for Standard Format Markers a misnomer if ever there was one SFM codes and their usage has become so fragmented that there s now Unified Standard Format Markers 16 to standardise SFM and encourage interoperability Our translations didn t use USFM SFM codes appear at the begin
91. PARQL does not allow empty sequences as parameters and suppresses errors in certain logical operations e g TRUE OR Error becomes TRUE We use the empty sequence e g generated by OPTIONAL expressions as a placeholder for error and put additional checking code to capture wrong input values For AND OR NOT and effective Boolean value we create helper functions that interpret correctly or catch the error on XQuery 3 0 respectively 3 4 Modifiers SPARQL solution modifiers either affect the unordered result sequence by imposing an order projecting variables or limiting the number of results In any case there is a straightforward translation to XQuery Given the intermediate variable binding sequence projection is expressed in the final return clause by only showing refer enced variables The SPARQL ORDER BY maps directly to the order by in an XQuery 87 Translating SPARQL and SQL to XQuery FLWOR expression both working on sequences Result size LIMIT and OFFSET are handled by placing positional predicates on the result sequence e g position lt 11 for LIMIT 10 DISTINCT is pushed into the query plan affecting operators on patterns as well as the custom XQuery functions implementing SPARQL identity semantics REDUCED is currently translated into a NO OP since dropping some du plicates is only an optimization 3 5 Query Forms SPAROL supports four query forms SELECT ASK CONSTRUCT and DESCRIBE We show the trans
92. POST amp permission foo read amp permission bar read 33 A JSON Facade on MarkLogic Server Set collections POST jsonstore xgy uri http foo bar amp collection foo amp collection bar Set document quality jsonstore xgy uri http foo bar guality 10 POST If you don t like the jsonstore xgy path or don t want to expose it publicly you can add a URL rewriter rule to beautify the HTTP paths The other REST endpoint is jsonquery xqy It accepts a query constraint as a g parameter written in JSON and returns the resulting JSON documents For example Table 4 Query Fetch Query by price jsonquery xqy g key price value 15 comparison lt It returns a result that s of course encoded as JSON count 21 results book The Great Gatsby author F Scott Fitzgerald price 12 99 The MLJSON user never sees XML The documents are stored as JSON and queried using JSON with results returned as JSON 7 Discussion What are the pros and cons of the MLJSON design We think it provides an easy and approachable storage model well suited to those starting a new application from scratch and a new alternative to those with a pre existing investment in JOON as a storage format On the other hand people familiar with MarkLogic and the full expressiveness of XQuery and XSLT will find the simple store and retrieve model of MLJSON restrictive Also there are
93. Prague is a conference on XML for developers markup geeks information managers and students In its sixth year XML Prague focuses especially on integ ration of XML with new web technologies and ever increasing capabilities of modern web browsers The conference provides an overview of successful XML technologies with the focus being more towards real world application versus theoretical expos ition XML Prague conference takes place 26 27 March 2011 at the Lesser Town Campus of the Faculty of Mathematics and Physics Charles University Prague XML Prague 2011 is jointly organized by the XML Prague Organizing Committee and by the Institute for Theoretical Computer Science The full program of the conference is broadcasted over the Internet see http xmlprague cz XML fans from around the world are encouraged to take part on line Remote and local participants are visible to each other and all have got a chance to interact with speakers This is the sixth year we have organized this event Information about XML Prague 2005 2006 2007 2009 and 2010 was published in ITI Series 2005 254 2006 294 2007 353 2009 428 and 2010 488 see http iti mff cuni cz series Petr Cimprich amp Jirka Kosek amp Mohamed Zergaoni XML Prague Organizing Committee The Institute for Theoretical Computer Science is supported by project 1M0545 of the Czech Ministry of Education xiii XIV Client side XML Schema validation Aleksejs Gorem
94. R Alexander Milowski eee 135 EPUB Chapter and Verse Tony Graham and Mark Howe 149 DITA NG A Relax NG implementation of DITA George Bina 167 XQuery Injection Eric van der Vlist caked aia iui aie ana aise 177 XQuery in the Browser reloaded Thomas Etter Peter M Fischer Dana Florescu Ghislain Fourny and Donald Kossmann ssssesesesesessesesesesesrsrsesesesrerersrsrsrrsesrerrrereses 191 Declarative XQuery Rewrites for Profit or Pleasure John Snelson 211 vii viii General Information Date Saturday March 26th 2011 Sunday March 27th 2011 Location Lesser Town Campus of Charles University Lecture Halls 55 and S6 Malostransk n m st 25 110 00 Prague 1 Czech Republic Organizing Committee Petr Cimprich Ubiqway James Fuller MarkLogic Vit Janota Tomas Kaiser University of West Bohemia Pilsen Jirka Kosek xmlguru cz amp University of Economics Prague Pavel Kroh pavel kroh cz Mohamed Zergaoui Innovimax Programm Committee Robin Berjon freelance consultant Petr Cimprich Ubiqway Jim Fuller MarkLogic Michael Kay Saxonica Jirka Kosek chair University of Economics Prague Uche Ogbuji Zepheira LLC Petr Pajas Google Felix Sasaki German Research Center for Artifical Intelligence John Snelson MarkLogic Eric van der Vlist Dyomedea Priscilla Walmsley Datypic Norman Walsh MarkLogic Mohamed Zergaoui Innovimax Produced By XML Prague cz
95. Request but while testing we cannot load anything due to the same origin policy as mentioned in Section 3 4 e Similarly we cannot test module import as it also depends on file access e We do not support normalize unicode which accounts for 34 tests Simil arly we do not support 197 test cases are schema related and 46 on static typing The tests cases that are actually failing are distributed across diverse test groups Some problems arise from the fact that the catalog XML file is over 10 megabytes in size giving us an XML parser error when executed in the browser There are some remaining problems with date arithmetics and time zones 206 XQuery in the Browser reloaded 5 2 Supported Platforms XQuery in the Browser runs on all modern and standards compliant browsers These are namely later versions should always work Firefox 3 6 and 4 also Mobile Google Chrome 7 Internet Explorer 9 Safari 5 Opera 11 also mobile the Android browser mobile Chrome the iPhone browser mobile Safari We do not yet support Internet Explorer 8 or lower because it has a much lower compliance to W3C standards Due to the limitations of GWT using document write XOIB currently handles only HTML and HTML5 DOMs but not XHTML DOMs which is somewhat ironic given that XHTML is conceptually conceptually much closer to the XML world 5 3 Performance 5 3 1 Runtime Please note that these numbers are all approximated as there are
96. XMLHttpRequest based test e events an XMLReader based test with an empty event handler registered for XML events e no handler an XMLReader based test with no event handler registered The tests were run and the VMSize was measured after a run to 10 iterations of loading the XML document The measured memory usage is shown in Figure 1 The memory usage for the XMLReader interface is minimal and grows very little in comparison to the document size For the XMLHttpRequest based tests the memory usage increases dramatically such that the iteration of 10 document load requests for the 100MB test failed to complete The memory usage of the XMLReader based tests stays very low but slowly in creases This may be due to garbage collection in Javascript environment that has yet to happen or it may be a small leak within the implementation or browser In theory the memory usage should stay relatively constant regardless of the size of the document as the testing code does not retain the document information The average parsing time was also measured and is shown in Figure 2 In contrast to the memory consumption the constant need to callback to the scripting environ ment for event delivery shows a 2 75 times penalty in parsing time When no event handler for the XML is registered the parsing time drops back down to consistent levels with the XMLHttpRequest implementation A modified version of the XMLReader and XMLItemEvent interface was tested
97. XPath 2 0 or XOuery 1 0 in these expressions this opens a door for attacks that are similar to XOuery literal injections Again if you use request parameters you ll be safe If not the sanitization to apply is the same than for XOuery injection except that the XML tool or API that you ll be using should take care of the XML entities 3 5 4 evaluate injection Extension functions such as saxon evaluate or eXist s util eval are also prone to attacks similar to XOuery injection if user input is not properly sanitized The conseguences of these injections may be amplified by extension functions that provide read of write access to system resources but even vanilla XPath can be harmful with its document function that provides read access to the file system as well as network resources that may be behind the firewall protecting the server These function calls need to be secured using similar technigues adapted to the context where the function is used 188 XQuery Injection Bibliography 1 Wikipedia Code injection http en wikipedia org wiki Code_injection 2 The Web Application Security Consortium Web Hacking Incident Database http projects webappsec org w page 13246995 Web Hacking Incident Database 189 190 XQuery in the Browser reloaded Riding on the coat tails of JavaScript Thomas Etter ETH Zurich lt etterth student ethz ch gt Peter M Fischer ETH Zurich lt peter fischer inf ethz ch gt Dana
98. a the webapp map expath web xm1 all public components are associated an import URI The webapp map can then use those absolute URIs to reference components making it independent on the physical strcuture of the project The webapp descriptor looks like the following lt webapp xmlns http expath org ns webapp descriptor xmlns app http example org ns my website name http example org my website abbrev myweb version 1 3 0 gt lt title gt My example website lt title gt lt resource pattern style css media type text css gt lt resource pattern images png media type image png gt lt servlet gt 41 CXAN a case study for Servlex an XML web framework lt xproc uri http example org ns my home xproc gt lt url pattern gt lt servlet gt lt servlet gt lt xguery function app other page gt lt url pattern other gt lt servlet gt lt servlet gt lt xslt uri http example org ns servlets xsl function app yet page gt lt url pattern yet gt lt match group 1 name id gt lt url gt lt servlet gt lt servlet gt lt xslt uri http example org ns catch all xsl gt lt url pattern gt lt servlet gt lt webapp gt Besides some metadata like the webapp name and its title the webapp descriptor is basically a seguence of components each associated with a URL pattern A URL is a regex When a r
99. abash to call the corresponding pipeline connecting the web request document to the pipeline port request As you can see the pipeline is identified by an absolute URI The packaging support configured on Calabash with the Servlex own repository takes care of resolving this URI correctly within the repository The case of the Home component is very simple it simply returns the following abstract page description lt page menu home gt lt title gt CXAN lt title gt lt image src images cezanne jpg alt Cezanne gt 48 CXAN a case study for Servlex an XML web framework lt para gt CXAN stands for lt italic gt Comprehensive XML Archive Network lt italic gt If you know CTAN or CPAN resp for La TeX and Perl then you already understood what this website is all about providing a central place to collect and organize existing libraries and applications writen in XML technologies like XSLT XQuery and XProc lt para gt lt page gt In the basic case a component must return a full description of the HTTP request to return to the user using an element web response But for all webpages the HTTP response will always be the same except the content of the returned body that is the page itself a status code 200 Ok and a content type header with text html Besides all pages share the same structure and a consistent layout So instead of repeating this information for every pages in every pipeline the
100. accessing default values is initiated and then on each start element callback if there is a default attribute for that element that is not present then it is added to the attributes list with the specified property set to false to identify it as a default attribute 3 DITA implementation in Relax NG 3 1 Schema development The first step in creating the Relax NG schemas for DITA 1 2 was to start with the DITA 1 2 DTDs and apply Trang to get equivalent schemas However there are many DTDs and Trang will generate the same DTD multiple times and sometimes with slightly different content depending on how the entities in the DTD get expan ded Thus after a first step of applying Trang on each main DTD the ent and mod modules get automatically converted it is needed a second step that requires mer 170 DITA NG A Relax NG implementation of DITA ging the multiple instances of the same schema and putting the schemas in a folder structure similar with the one that is used for the DITA DTDs The rough schemas obtained after the conversion and merging need to be changed to properly use Relax NG constructs for extensions and redefinitions Having the schemas in XML format in the case of the Relax NG with XML syntax it is possible to develop Schematron rules that can check that the definitions in the schemas are consistent 3 2 Automatic conversion to Relax NG compact Trang can convert without loss from Relax NG XML syntax to Relax NG co
101. ace for the SPAROL and SQL to XQuery translator http www xql2xquery org Website is protected Please use ethz xquery 12 Graham Klyne et al Resource Description Framework RDF Concepts and Abstract Syntax 13 Jim Melton SOL XQuery and SPAROL What s wrong with this picture XTech 2006 14 Matthias Nicola and Tim Kiefer Generating SOL XML Query and Update Statements CIKM 2009 15 Eric Prud hommeaux Alexandre Bertails A Mapping of SPAROL Onto Conventional SOL http www w3 org 2008 07 MappingRules StemMapping 16 Eric Prudhommeaux et al SPAROL Query Language for RDF 17 Sausalito XOuery Application Server Version 0 9 6 2009 http sausalito 28msec com 18 SOL Part 14 XML Related Specifications SQL XML ISO IEC 9075 1 2008 IDT 19 SPAROL Ouery Results XML Format W3C Recommendation 15 January 2008 http www w3 org R rdf spargl XMLres 20 XML Ouery Use Cases W3C Working Group Note 23 March 2007 http www w3 org TR xquery use cases 98 Configuring Network Devices with NETCONF and YANG Ladislav Lhotka CESNET z s p o lt lhotka cesnet cz gt Abstract This paper gives an overview of the open standards for the NETCONF protocol and associated framework for configuration data modelling NETCONF is an XML based communication protocol that allows for secure management of network devices from remote manager applications The YANG language for configuration data modelli
102. after the selected element If the XML documentis valid the user can edit the document We use the restric tion methodology that is a user can make only a subset of the actions allowed by the schema and so the document always stays schema valid When a user selects an W3C XML component the API checks the FSM from the compiled schema to provide prompts for the allowed actions Results DOM Validation RL Figure 2 Application Programming Interface 4 Editing Functionality In the XML context researchers are just beginning to look at update languages and validation 2 so we explored some new methods to cope with editing issues We considered two validation methodologies incremental and allowed changes only or restriction Client side XML Schema validation In 2 and 3 authors present incremental validation techniques of XML docu ments with respect to an W3C XML Schema definition When a valid document is updated it has to be verified that the new document still confirms to the imposed constraints Brute force validation from scratch is not practical because it requires reading and validating the entire document following each update The core of the incremental approach consists in checking the XML document for the validity after the performed transaction In spite of the inherent flexibility of this approach in cluding the possibility to rename components and to add or remove whole sub trees in practice the meth
103. agement http calibre ebook com 5 Converting SFM Bibles to OSIS http www crosswire org wiki Converting SFM_Bibles_to_OSIS 6 Cory Doctorow Content Selected essays on Technology Creativity Copyright and the Future of the Future San Francisco Tachyon 2008 pp124 125 7 EPUB specifications http www idpf org specs htm 8 EpubCheck http code google com p epubcheck 9 Johannes Gutenberg http en wikipedia org wiki Johannes_Gutenberg Printed_books 10 OSIS Milestones http www crosswire org wiki OSIS_Bibles OSIS_Milestones 11 Navigation Control File NCX http www niso org workrooms daisy Z39 86 2005 html NCX 165 EPUB Chapter and Verse 12 The Good Book Business why publishers love the Bible http www newyorker com archive 2006 12 18 061218fa_fact1 13 Open Packaging Format OPF 2 0 1 v1 0 1 http www idpf org doc_library epub OPF_2 0 1_draft htm 14 The OSIS Website http bibletechnologies net 15 OSIS 2 1 1 User Manual 06March2006 http bibletechnologies net utilities fmtdocview cfm id 28871A67 D5F5 4381 B22EC4947601628B amp methods title 16 Unified Standard Format Markers http paratext ubs translations org about usfm 166 DITA NG A Relax NG implementation of DITA George Bina Syncro Soft oXygen XML Editor lt georgelioxygenxml com gt Abstract DITA DocBook and TEI are among the most important frameworks for XML documents While the l
104. ame value pairs The pairs here are represented as XML elements with a name corresponding to the map name and a value to the map value The type is provided as an attribute Some elements i e the empty array the booleans and the null value don t need child text nodes Below is object and h another encoding example this time using a JSON array instead of olding just simple values import module namespace json http marklogic com json at mljson json json lib json xqy ToXML hello world null false true 9 99 The result lt json type array gt lt item type string gt hello world lt item gt item type array gt lt item type object gt lt item type null gt lt item boolean false gt lt item boolean true gt lt item type number gt 9 99 lt item gt lt json gt In an array there are no names so the items are represented by lt item gt elements 28 A JSON Facade on MarkLogic Server Name Escaping Using JSON names as XML element names makes it easier for MarkLogic to execute efficient XPath and full text queries against this data as we ll see later but what if the name isn t a legal XML name As of this writing an illegal name character will generate an error The plan is for MLJSON to support an escaping technique that enables safe storage of any string with reliable round tripping An underscore will
105. amp amp gt Samp amp amp amp apos gt lt xsl variable name queryStart gt user mail lt xsl variable gt lt xsl variable name queryEnd gt lt xsl variable gt query 184 XQuery Injection lt xsl variable name query select encode for uri concat SqueryStart P SusernameSanitized SqueryEnd gt lt xsl variable name userInformation select doc concat http localhost 8080 orbeon exist rest db app users gt query Squery gt To use request parameters simply write something such as Example 6 XQuery Using Query Parameters in XSLT lt xsl variable name usernameNormalized gt select lower case normalize space username gt lt xsl variable name query gt declare namespace request http exist db org xquery request user mail request get parameter mail 0 lt xsl variable gt lt xsl variable name userInformation select doc concat http localhost 8080 orbeon exist rest db app users gt mail encode for uri usernameNormalized amp amp query encode for uri aguery gt Here again the choice to normalize spaces and convert to lower case depends on the context 3 4 XForms The problem is very similar in XForms with the difference that XForms is meant to deal with user input and that the chances that you ll hit the problem are significantly bigger The rule
106. an HTML string 2 1 2 Creating a document from scratch The various node classes can be used as factories for creating entities documents and other nodes 37 Akara Spicy Bean Fritters and XML Data Services from amara import tree doc tree entity doc xml append tree element None u spam doc xml write lt xml version 1 0 encoding UTF 8 gt n lt spam gt 2 2 The XML bindery Some of that xml children N stuff is a bit awkward and Amara includes a friendlier API called the XML bindery It is like XML data bindings you might have heard of but a more dynamic system that generates object attributes from the names and construct in the XML document from amara import bindery MONTY XML lt monty gt lt python spam eggs gt What do you mean bleh lt python gt lt python ministry abuse gt But I was looking for argument lt python gt lt monty gt doc bindery parse MONTY XML m doc monty pl doc monty python or m python pl is just the first python element print print pl xml attributes None u spam print pl spam for p in doc monty python The loop will pick up both python elements p xml write Importantly bindery nodes are subclasses of amara tree nodes so everything in the amara tree section applies to amara bindery nodes including the methods for re serializing to XML or HTML Amara bindery uses iterators to provide access to multiple child elements w
107. anIfElseTrue if condition then then else b true gt not condition or then where rw subtype then xs boolean declare option rw rule rw BooleanIfThenTrue if condition then true else b else gt condition or else where rw subtype else xs boolean Bibliography 1 XQuery 1 0 An XML Query Language Second Edition 2 XQuery 1 0 and XPath 2 0 Formal Semantics Second Edition 3 XSL Transformations XSLT Version 20 4 Oracle Database Performance Tuning Guide 10g Release 2 10 2 Using Optimizer Hints 5 The Glorious Glasgow Haskell Compilation System User s Guide Version 6 12 2 Rewrite rules 6 Writing an XSLT Optimizer in XSLT Michael Kay http www w3 org TR 2010 REC xquery 20101214 lt http www w3 org TR 2010 REC xquery semantics 20101214 http www w3 org TR 2007 REC xslt20 20070123 lt http download oracle com docs cd B19306_01 server 102 b14211 hintsref htm gt http www haskell org ghc docs 6 12 2 html users_guide rewrite rules html http conferences idealliance org extreme html 2007 Kay01 EML2007Kay01 html 224 Declarative XQuery Rewrites for Profit or Pleasure 7 Towards an Exhaustive Set of Rewriting Rules for XQuery Optimization BizQuery Experience Maxim Grinev Sergey Kuznetsov 7 http www ispras ru en modis downloads rewriting extended pdf 225 226 Ji Kosek ed XML Prague 2011 Conference Procee
108. antly worse performance with ARQ clearly outperforming the XQuery engines Although there are several approaches to translate SQL to XQuery all of them suffer from quite severe restrictions One ap proach and an overview of the competing methods are given in 14 For the opposite direction the Pathfinder Relational XQuery Processor 9 has achieved a very high degree of correctness and performance 96 Translating SPARQL and SQL to XQuery 7 Conclusion and Future Work In this work we tackled the problem of aligning SPAROL SOL and XQuery se mantics and execution by providing a translation from SPARQL and SQL to XQuery The translation is complete correct and universally usable as it does not need any schema or specific workloads An initial performance evaluation shows that in areas on which we already performed optimizations the XQuery translation beats the native SPARQL implementations whereas in other areas it still lags behind When comparing to SQL storage seems to be the relevant difference For both translations current XQuery optimizers seem to exploit many optimizations when queries are reasonably simple but fail once a certain level of complexity is exceeded We see the following avenues for future work Our translation should be further validated both by formal verification and testing against the official DAWG SPAROL test suite We aim to incorporate additional optimizations in order to reduce intermediate results for pat
109. applications and work only on specific platforms e g Altova XMLSpy Oxygen XML Editor Syntext Serna To our knowledge there is only one client side application that is supported and provides comprehensive editing functionality called Xopus However it is commercial software with all ensuing consequences no open source free system with these capabilities exists Therefore our goal has been to develop a system mainly from scratch using the algorithms and concepts of Web 2 0 WYSIWYG and XML validating This paper explores client side XML validation and editing techniques The continuous restriction validation technique is proposed Based on analysis and research a client side W3C XML Schema validator and editor for XML documents is designed and implemented The system can be used for validation of XML documents accord ing to W3C XML Schema identification of possible elements attributes to insert de lete as well as for physical deletion of elements from an XML tree and text editing The application is written on JavaScript using good object oriented practises and could be easily improved and integrated in future The rest of this paper is organised as follows After reviewing related works and preliminaries in Section 2 we introduce the system architecture in Section 3 Section 4 describes XML Schema compliant editing In Section 5 we discuss implementation details and make a brief analysis and in Section 6 we summarise our conclusions and
110. arison lt It s also possible to specify how many results will be returned using the position constraint key book position 1 to 10 To understand how the JSON is evaluated internally the following table shows the internal XPath associated with each JSON query JSON Internal XPath key foo value bar json foo bar key foo value bar guux json foo bar guux innerKey foo value bar quux json foo bar quux key foo json exists foo key foo value key id json foo id 0596000405 value 0596000405 key book or key id json exists book id 0596000405 value 0596000405 key other id or other id 0596000405 value 0596000405 key price value 8 99 comparison lt json price lt 8 99 key book position 1 to 10 json exists book 1 to 10 These examples just scratch the surface There s also a fulltext constraint that exposes MarkLogic s capabilities regarding full text range metadata property value and geospatial indexing fulltext or equals key greeting string Hello World weight 2 0 caseSensitive false diacriticSensitive true punctuationSensitve false whitespaceSensitive false stemmed false wildcarded true minimumOccurances 1 maximumOccurances null 31 A JSON Facade on Ma
111. arrives that runs on a new stack frame without excessive inversion of the Saxon code After issuing the call the calling code can continue as normal leaving a place marker element in the result tree for the result of the apply templates instruction when the document arrives and has been fully processed the resulting tree fragment can be stitched into the result tree in place of the place marker element The ability to make the transform ation asynchronous in this way seamlessly to the user demonstrates one of the benefits of programming in a declarative language However the limitations of the JavaScript threading model are such that it s hard to see how to handle the general case for example where the result of processing an external document is used to form the value of a variable and this variable is then used by the original processing thread So far so good we can run a stylesheet create HTML output add the HTML output to the tree maintained by the browser and have it rendered That s essentially what today s in browser XSLT 1 0 implementations do We ve already delivered some benefits access to the much richer functionality of XSLT 2 0 as well as browser independence the same XSLT engine and therefore the same XSLT stylesheets run in any browser that supports HTML5 and JavaScript But as I said in the first section to make XSLT on the browser a success I think we have to go further In particular we have to handle interac
112. as to find the website find the package download it and invoke the repository manager with this package file to install it locally And if the package depends on another the user has to find the dependencies and install them also Recursively All that process could be automated CXAN tries to solve that problem by providing two complementary components The first component is the website The CXAN website is aimed at gathering all known XML packages at organizing them in a stable distribution and at maintaining that distribution over the time Every package in CXAN is given a unique ID a ab breviation string The second component is the CXAN client The client is a program that manages parts of this stable distribution in a local repository The client can install packages on the local machine by downloading them directly from the CXAN website and resolving automatically the dependencies There is a command line client to maintain a standard on disk repository but every processor can define its own client or an alternate client for instance to provide a graphical interface in an XML IDE The website is organized as a large catalog of XML libraries and applications that you can navigate through tags authors and categories or that you can search 36 CXAN a case study for Servlex an XML web framework using some keywords or among the descriptions It is located at http cxan org The following screenshot shows the details of the
113. at Javascript was becoming capable of being used as a system programming platform to implement other languages In the twelve months since XML Prague 2010 we have seen announcements of an XProc engine running on the browser 3 and the XQIB XQuery engine running on the browser has been re engineered to run as native Javascript 4 I have also produced a prototype implementation of XSLT 2 0 on the browser which I want to say more about today All these products were produced from existing Java engines by cross compiling the Java to Javascript using the Google Web Toolkit GWT 5 Although no detailed performance data is available all deliver a responsiveness which feels perfectly adequate for production use At first sight the architecture seems improbable Javascript was hardly designed as a system programming language it doesn t even offer integer arithmetic GWT translates all integer arithmetic into floating point arithmetic and emulates a long 64 bit integer using two 64 bit double values The code generated for common system programming operations such as bit masking and shifting looks outrageously 127 XSLT in the Browser inefficient And yet there is so much processor power available on today s typical desktop machine that this really doesn t matter it turns out to be fast enough It s worth observing that one well respected client side XForms engine XSLT Forms 6 uses an even more improbable architecture it is i
114. atabase is queried according to parameters entered by our users but there are cases where you may want to copy user input within XQuery direct element constructors http demo exist db org exist functions reguest get parameter 187 XQuery Injection One of the use cases for this is XQuery Update Facility which update primitives may contain direct element constructors in which it is tempting to include input fields values Here again you re safe if you use request parameters but you need to sanitize your input if you re doing direct copy The danger here isn t that much delimiters but rather enclosed expressions that let your attacker include arbitrary XQuery expressions The lt also needs to be escaped as it would be understood as a tag delimiter as well of course as the amp That makes 4 characters to escape amp must be replaced by samp lt must be replaced by slt must be replaced by So a 75 must be replaced by 3 5 3 XUpdate injection XUpdate is safer than XQuery Update Facility since it has no support for enclosed expressions That doesn t mean that and lt do not mean to be escaped but since XUpdate documents are well formed XML document the tool or API that you ll be using to create this document will take care of that if it s an XML tool Unfortunately XUpdate uses XPath expressions to gualify the targets where updates should be applied and if you use a database like eXist which supports
115. atest versions of DocBook and TEI use Relax NG as the schema language DITA is still using DTDs There were some fragile at tempts to get DITA working with Relax NG but it takes more than writing a Relax NG schema to have this working DITA NG is an open source project that aims to provide a fully functional framework for a Relax NG based imple mentation of DITA DITA NG provides the Relax NG schemas for DITA 1 2 and also support for default attribute values based on Relax NG a defaultValue annotations this is the critical part that makes DITA work The presentation covers an overview of the Relax NG schemas how DITA specializations can be done using Relax NG a lot simpler than with DTDs the support for default attribute values for Relax NG and includes a demo of the complete workflow of working with DITA based on Relax NG Keywords DITA XML authoring editing schema Relax NG 1 Introduction The XML documents can be roughly divided in data oriented and content oriented The content oriented documents contain a lot of mixed content basically they contain text annotated with markup that is used to identify different roles for that text While XML Schema is very useful for the data oriented XML documents it is not the same case with content oriented documents There are a few characteristics that makes it not suitable for these documents like for instance the extension mechanism that can add new elements only at the end of an existing co
116. ave an interface that invokes the transformation through markup of this kind in a skeletal HTML page rather than relying on Javascript coding My design aim is that it should be possible to use XSLT without using any Javascript The principal output of the stylesheet is of course a tree of nodes and this is in serted into the HTML page at a location identified here by an id attribute href body XSLT 2 0 allows a stylesheet to produce multiple result trees and each of these can populate a different part of the HTML page using the href attribute of lt xsl result document gt in the same way As well as identifying the attachment point of the result document by means of an ID value it can also be selected by means of an XPath expression for example lt xs1 result document href select id table tr n gt which allows a sequence of result documents to populate the rows of an HTML table In the current implementation the document function fetches an XML document using a synchronous XmlHttpRequest This is of course unsatisfactory and in a production release it will be necessary to make the request asynchronous Initial experiments suggest that it should be reasonably easy to implement this provided 130 XSLT in the Browser the function is used in a context such as lt xsl apply templates select document xyz xml gt this can translate into an XmlHttpRequest with a callback function to be invoked when the document
117. be used to initiate an escape sequence It will be followed by four hexadecimal numbers defining the character it represents If an underscore appears in the name before it s es caped it will be escaped like any other special character Any characters not allowed at the start of an element name such as a digit will be escaped also underscore is allowed at the start of an element name An empty string will be handled using a special rule that maps it to a single underscore Remember all escaping and unescaping will happen automatically and transparently Table 1 Example Escape Sequences Element Tag Name Notes lt a gt lt _0031 gt Element names can t start with a digit lt gt SH lt 005F gt All underscores need to be specially escaped lt 0024 gt foosbar lt foo 0024bar gt ifoosbar lt 0031f00 0024bar gt 2 OEM lt x 003Ahtml gt 5 Ouerying JSON Now that we have a technigue to store JSON documents inside MarkLogic we need a way to guery the documents Using XPath and XOuery are certainly viable options but they reguire exposing the internal details of the storage format to the user It s better to define a guery syntax that maintains the JSON facade The MLJSON query syntax fittingly lets you specify query constraints through a declarative JSON structure For example this query syntax finds stored JSON 29 A JSON Facade on MarkLogic Server documents that
118. been defined for use with NETCONF e Secure Shell SSH which is mandatory for all implementations 5 e Simple Object Access Protocol SOAP 6 e Blocks Extensible Exchange Protocol BEEP 7 e Transport Layer Security TLS 9 Other features of the NETCONF protocol such as datastore locking or error handling while important are not relevant for the main topics of this paper 3 YANG Data Modelling Language When the NETCONF working group was chartered in the IETF the priority was to finish the NETCONF protocol as soon as possible Therefore it was deliberately decided to leave the large and potentially contentious area of data modelling outside 102 Configuring Network Devices with NETCONF and YANG the scope of the working group In the meantime various approaches to data modelling were tried Some used standard XML schema languages such as W3C XML Schema Definition XSD or RELAX NG for this purpose while others de veloped new specialized data modelling languages In April 2008 the NETMOD NETCONF Data Modelling working group was chartered and at the same time one of the specialized data modelling languages named YANG was selected as the base for further development The standard XML schema languages were rejected for the following main reasons e Configuration data modelling has to address not only syntax but also semantics of the modelled data e W3C XSD is very complex and difficult to learn e T
119. betically numbered chapters can stop and resume again after one or more numerically numbered chapters Some books don t have chapters only numbered verses Unnumbered verses can appear outside any chapter and alphabetically numbered verses can follow numerically numbered Verses can be merged such that one sentence is multiple verses or split into a and b parts Verse parts can be non consecutive Content can also appear in verse form and be repeated in tabular form as if Moses really wanted to put a table there but didn t have the technology at the time One translation helpfully sprinkles throughout as markup to indicate pos sible line breaks another translation uses as meaningful text in footnotes Although SFM doesn t include a way to indicate editions Catholic and Protestant versions of a translation may have been printed from one SFM but a reference in a footnote in a shared book to a verse in a Catholic only book may have dif ferent text in the printed Protestant Bible or the reference and the punctuation e g and around the reference may be elided in the Protestant Bible 2 3 One step at a time Processing is performed in stages by a chain of stylesheets that each do a single task to one book at a time This allows us to swap or insert stages to handle the non standard parts of each translation s SFM The initial expectation was for ten or so stylesheets There are currently
120. bject into an XML document The key point of the proposed conversion is to allow an intuitive use of XPath with a minimal number of extra attributes and elements XForms developers should manipulate JSON data without having to mentally convert it in XML This is implemented in XSLTForms and demonstrated with the integration of an external JSON API in an XForms page Keywords xml xforms json xsltforms 1 Introduction XForms is a condens of good programming patterns for forms design XForms uses concepts also found in entreprise proofed architectures such as ASP Net and J2EE extra elements are mixed within an host language such as HTML to dramat ically reduce the need of programming instructions and improve both productivity and quality XForms is indeed a precursor for HTML5 which now includes some of its good ideas but just at controls level XForms was initially specified to be integrated in a full XML environment at client side Web developers are in practice heavily using JSON for exchanging data between browsers and servers and having to use XML instead to benefit from XForms is certainly a problem for them XForms is using XPath expressions for selecting nodes and checking constraints and XPath cannot be used directly on JSON objects XPath is a rich query language itself while JSON doesn t have this possibility natively So it is proposed to store JSON objects within internal XML documents the conversion being made automat ical
121. cent gt lt element gt lt optional gt lt optional gt 117 Configuring Network Devices with NETCONF and YANG lt element nma implicit true name cm milk gt lt ref name coffee machine state supply levels percent gt lt element gt lt optional gt lt optional gt lt element nma implicit true name cm coffee gt lt ref name coffee machine state supply levels percent gt lt element gt lt optional gt lt interleave gt lt element gt lt optional gt lt optional gt lt element name cm temperature gt lt data type unsignedByte gt lt element gt lt optional gt lt interleave gt lt element gt lt optional gt lt interleave gt lt nma data gt lt nma rpcs gt lt nma rpc gt lt nma input gt lt element name cm put the kettle on gt lt optional gt lt element name cm recipe nma default espresso gt lt choice gt lt value gt espresso lt value gt lt value gt turkish lt value gt lt value gt cappuccino lt value gt lt choice gt lt element gt lt optional gt lt element gt lt nma input gt lt nma rpc gt lt nma rpcs gt lt nma notifications gt lt nma notification gt lt element name cm low level warning gt lt optional gt lt element name cm ingredient gt lt choice gt lt value gt water lt value gt 118 Configuring Network
122. chanism for the EPUBs but that is used less often 2 3 3 Pros and cons of using XSLT 2 0 e Using XPaths made some hard logic easy e g determining whether a text node is the correct place to insert a verse end milestone lt true only if Stext is child of the right kind of element and within a chapter gt lt xsl function name t versable as xs boolean gt lt xsl param name text as text gt lt xsl seguence select exists Stext preceding o chapter 1 and empty Stext ancestor o title not canonical true and empty Stext ancestor o note and empty Stext ancestor o speaker and empty Stext ancestor o w and exists for Selement in Stext ancestor return if namespace uri Selement eg namespace uri So ns and local name Selement Sversable elements then element else gt lt xsl function gt e Stylesheets are simpler and more concise than if XSLT 1 0 was used 158 EPUB Chapter and Verse e Because XSLT works by matching patterns moving a node from one part of the document to another is really a matter of firstly matching to the new location for the node then selecting from that node to find the node to copy to the new location and secondly all but repeating the same logic to match the node in its original location so it is explicitly not processed lt xsl template match 0 div exists t opening chapter start mode move up gt
123. common XML standards as well as fresh ideas for expressing XML pattern processing based on long experience in standards based XML applications Some of these features include XPath and XSLT a lightweight dynamic data binding mechanism XML modeling and processing constraints by example using Examplotron Schematron assertions XPath driven streamable pro cessing and overall low level support for lazy iterator processing and thus the map reduce style Akara does not enforce a built in mechanism for persist ence of XML but is designed to complete a low level persistence engine with overall characteristics of an XML DBMS Akara despite its deliberately low profile to date has played a crucial role in several marquee projects including The Library of Congress s Recollection project and The Reference Extract project a collaboration of The MacArthur Foundation OCLC and Zepheira In Recollection Akara runs the data pipeline for user views and is used to process XML MODS files with catalog records In RefExtract Akara processes information about topics and related Web pages to provide measures of page credibility Other users include Cleveland Clinic Elsevier and Sun Microsystems This paper introduces Akara in general but focuses on the innovative methods for XML processing in stand alone code and wrapped as RESTful services http akara info 53 Akara Spicy Bean Fritters and XML Data Services 1 Introduction Akara s dev
124. consoled by the discov ery that other Bible EPUBs suffer from the same problem As soon as we moved from desktop EPUB software to embedded readers we discovered that the manufacturer s interface makes a significant difference to the user experience For example desktop software tends to display the contents list in a separate pane which makes moving around our translations quite easy By com parison the Sony Reader requires one button press plus two screen taps just to get to the closed contents page which must then be opened and scrolled This discovery prompted us to provide a system of links to enable Sony Reader users to get from verse to chapter to book to contents page as easily as possible Also the Sony Reader back function is hidden off a submenu which makes visiting a glossary something of a one way trip The iPhone screen is small enough to make hitting links on superscript verse numbers a challenge We also tried the Archos eReader 70 which appears to have no mechanism for selecting links in a document despite displaying them which made our Bible EPUBs of limited use to anyone wishing to get beyond the first half of the book of Genesis On the basis of our experience it seems to us that E book publishers should either use quite basic XHTML and CSS or produce one EPUB version per device For the moment we have opted for the first option 164 EPUB Chapter and Verse 4 Conclusion Comparing Gutenberg s Bible wit
125. continue with next edges e ifa candidate is equivalent to SE or EA we can always insert it after Otherwise we check candidate edges CE If CE has EA then the candidate is valid for in sertion The example of this algorithm is described in the next subsection For Insert Before algorithm we perform the same actions but over the element before the selected If the selected element is first we can add any element that comes before the selected element in the FSM or text if the parent element is of mixed type or selected element itself if its occurrence is less then maxOccurs field All element insertion operations require a valid XML document The content and attributes of inserted elements are determined by the default and fixed fields or restriction values for simple type or simple content elements by the required children for complex content elements and by the attributes of complex type elements In addition required attributes and content must be added to created child elements recursively The Insert Into 1 action requires checking maxOccurs field of all possible chil dren elements within the selected element and comparing the field with real children occurrences RCO If the RCO lt maxOccurs we add the elements into the array of insertion candidates If the selected element is empty we can insert any element from the array otherwise we need to find a place for the candidates for each child element we apply Insert After alg
126. cuments e facilitate exchange of data models and schemas with other IETF working groups e help XML savvy developers and users understand the constraints of a YANG data model 4 1 Overview of the Mapping Procedure As was explained in Section 3 a single YANG module specifies grammatical and semantic constraints for several different objects datastores with configuration and operational state data as well as various NETCONF protocol messages In ad dition a concrete data model may consist of multiple YANG modules In order to cope with this variability the mapping procedure is divided into two steps see Figure 1 1 Inthe first step a collection of input YANG modules that together define a data model is transformed to the so called hybrid schema 2 Inthe second step the hybrid schema then may be transformed in different ways to obtain DSDL schemas for desired object types YANG module s T hybrid schema get reply RPC notification Figure 1 Structure of the mapping The hybrid schema uses RELAX NG syntax for specifying grammatical and datatype constraints and also for representing YANG groupings as named pattern definitions Semantic constraints are mapped to various annotations that are attached to RELAX NG elements The annotations are XML attributes and elements belonging to the namespace with URI urn ietf params xml ns netmod dsdl annotatio
127. d ev event xforms value changed gt lt xf toggle ev event DOMFocusIn case show autocompletion gt lt xf input gt lt xf switch gt lt xf case id show autocompletion gt lt xf repeat id results nodeset instance iresults 2 p is non empty array and amp x300 gt lt xf trigger appearance minimal gt lt xf label gt lt xf output value gt lt xf label gt lt xf action ev event DOMActivate gt 22 JSON for XForms lt xf lt xf lt xf setvalue ref instance isearch search value current gt toggle case hide autocompletion gt action gt lt xf trigger gt lt xf repeat gt lt xf case gt lt xf case id hide autocompletion gt lt xf switch gt 5 3 How It Works Each time a character is added in the search field after a delay the request is sent to the server by serializing the leaf nodes of the search instance using a GET method The response is converted into XML data instance and a refresh is performed The returned entries are listed so one of them can be selected filtering is required for empty array and for erroneous answer 5 4 The Full Form lt html xmlns http www w3 org 1999 xhtml xmins Xx f http www w3 org 2002 xforms xmlns ev http www w3 org 2001 xml events gt lt head gt lt tit ley lt xf model gt lt xf Hi KIPEDIA OpenSearch Test Form lt title gt i
128. d the XMLReader instance For ex ample for data binding into JSON a JSON data binder could be provided directly 146 Efficient XML Processing in Browsers to the XMLReader As such the jump between C and Javascript can be avoided and that should improve the overall processing time The simple ability to load and parse large data sets allows the browser environ ment to extend into application scenarios where it might not have gone before For example a worker for a RDF search engine could load process and extract triples from XHTML documents by streaming them through a local RDFa processor Such a client application can take advantage of all kinds of new local services provided by the advanced browser environment as well as take advantage of the security constraints and protection provided by that same environment The implementation work described in this paper will be provided as a patch to WebKit that can be easily used run WebKit directly within the Safari browser For the more adventurous the patch can be built into other browsers like Chrome Further research on uses ancillary processors and improvements are intended Whether these proposed interfaces or their descendants are just a local enhancement or something to be standardized is up to the web community to decide Bibliography 1 Kesteren A XMLHttpRequest August 2010 http www w3 org TR XMLHttpRequest 2 Kesteren A XMLHttpRequest Level 2 August 2010 http
129. d to via the addEventListener method are the new model for processing the result When an XML response is received by the instance the XML parser is invoked in exactly the same way as for any XMLHttpRequest based request That is when the media type of the response entity is recognized as XML the XML parser is invoked If the media type is not an XML media type the parser will not be invoked and no response document is processed The difference is that instead of building an XML DOM a sequence of events are produced for each major facet of the XML being parsed That is events are produced for the start end document start end element characters data comments and processing instructions These events are encapsulated in a new interface called XMLItemEvent interface XMLItemEvent Event const unsigned short START DOCUMENT 1 const unsigned short END DOCUMENT 2 const unsigned short START ELEMENT ot const unsigned short END ELEMENT 4 const unsigned short CHARACTERS 5 const unsigned short PROCESSING INSTRUCTION 6 const unsigned short COMMENT 7 readonly attribute unsigned short itemType 138 Efficient XML Processing in Browsers readonly attribute DOMString prefix readonly attribute DOMString localName readonly attribute DOMString namespaceURI readonly attribute DOMString uri readonly attribute DOMString value returns the attribute value for a specific
130. der Web developers just need to learn Javascript and JSON Javascript is definitely easier to learn for Web developers mandatory for programming rich graphical ef fects by the way and yet good enough to develop anything Javascript is still heavily improved and there are plenty of libraries to enrich it Many non Web developers used to consider Web as some kind of underworld inhabited by unqualified self taught developers deserving smaller salaries if any Nevertheless AJAX has enabled Web developers to become key actors for new ap plications XML complexity proned by only a part of the non Web developers sounds useless to them whilst they are proud of JSON as the notation that extended their leadership JSON sounds like a revenge notation 2 5 Other Data Formats In fact any structured data format can be considered for XForms For example vCard CSV with or without titles formatted log files could be treated the same way It means that propositions for JSON support in XForms should be extensible and the use of json in terms should be avoided 3 JSON Objects Internal Storage In Browsers 3 1 Constraints 3 1 1 Reversibility JSON objects have to be serialized identically to the original ones when not modified by controls This does not mean that every XML document can be mapped into the proposed representation in JSON and back in XML 15 JSON for XForms 3 1 2 XPath Full Support XForms developers shouldn t have
131. dible translate Pali item 2 xml lang y 7 xref baseTopic xml test x3 x xtre 4 v eul ul 358038b5 2cc7 4f2b 810 y WA LZ SSS C Ten Gr ND AOZER a t x U 0065 Modified 5 Conclusions and further work The current state of development permits the complete workflow for working with DITA based on Relax NG The advantages are mainly on the clarity of the schemas and on the ease of developing DITA specializations without going through the parameter entities hell it is very difficult to correct errors when the DTDs use parameter entities Support for xml model PI for schema association Right now only the oxygen processing instruction is supported for Relax NG schema association As the W3C has a recommendation that specifies how an xml model processing instruction can be used to specify this association the parser component that adds default attribute values should use also that to find the associated Relax NG schema Support for specifying the Relax NG schema from outside as a parser property Right now the Relax NG schema is detected from a PI It should be possible to specify the schema also as a parser property so that the document does not need to have an explicit reference to a schema A similar support is available for XML Schema for example 174 DITA NG A Relax NG implementation of DITA Automatic conversion to DTDs Right now the DITA NG provides automatic conversion from Relax NG XML
132. dings Vydal MATFYZPRESS vydavatelstv Matematicko fyzik ln fakulty Univerzity Karlovy v Praze Sokolovsk 83 186 75 Praha 8 jako svou 354 publikaci Ob lku navrhl prof Jaroslav Ne et il Z p edloh p ipraven ch v syst mu DocBook a vys zen ch pomoc XSL FO a programu XEP vytisklo Reprost edisko UK MFF Sokolovsk 83 186 75 Praha 8 1 vyd n Praha 2011 ISBN 978 80 7378 160 6
133. dren is a sequence of the child nodes of the entity including the top element monty doc xml children 0 You might be wondering about the common xml prefix for these methods The higher level Bindery data binding API builds on amara tree It constructs object attribute names from names in the XML document In XML names starting with xml are reserved so this Amara convention helps avoid name clashes You can navigate from an node to its parent assert m xml parent doc Access all the components of the node s name including namespace information assert m xml local u monty local name i e without any prefix assert m xml gname u monty qualified name e g includes prefix assert m xml prefix None assert m xml gname u monty qualified name e g includes prefix assert m xml namespace Non assert m xml name None u monty The universal name or expanded name A regular Python print tries to do the useful thing with with each node type pl m xml children 0 print pl xml children 0 lt amara tree element at 0x5e68b0 name u python 0 namespaces 1 attributes b 1 children gt print pl xml attributes None u spam eggs 56 Akara Spicy Bean Fritters and XML Data Services A Notice the difference between the treatment of elements and attributes To deserialize a node to XML use the xml write or xml_encode method The former writes to an outpu
134. e item gt f e As well as being useful for XQuery library writers rewrite rules in query options can be useful to XQuery power users who wish to give hints force the use of indexes or fill in for cases where the implementations optimization behaviour is inadequate For instance a power user with knowledge of their XML data might know that the bibioentry element only ever occurs at a path of article bibliography biblioentry in which case they might choose to use the following rewrite 219 Declarative XQuery Rewrites for Profit or Pleasure biblioentryRewrite oiblioentry gt article bibliography biblioentry Or maybe they know that an index is available for biblioentry elements in which case they might use biblioentryIndex article bibliography biblioentry id e gt key biblioentry e 4 Complications Simple as rewrite rules seem there are many complications with their design that need work Its my hope that several XOuery implementations will find it useful to implement this kind of rewrite notation but if they were to do so there would be additional standardization issues that go beyond what the XOuery specifications already control One such issue is that currently XOilla inserts type checks in its expression tree before the rewrite rule phase occurs This simplifies rewrite rules because the im plicit type casts and checks are treated as part of the argument expression for a function
135. e Ns Exauc is moved inside the chapter Psalm titles are part of the text of the Bible unlike other titles which are modern day conveniences and as such are not canonical The Psalm title can and does contain verse milestones whereas the XSLT would not insert a milestone into a non canonical title Most translations use SFM markup to represent some accented characters 157 EPUB Chapter and Verse 2 3 2 Testing Initially testing used XSpec run against test files However we discontinued with XSpec at a time when the stages were being moved around a lot the rearrangement meant the table stage in particular could not correctly process its current input yet the XSpec tests which ran against test data were still all passing Running the XSpec tests against the live data wasn t that much of an option since we were working on different translations at different times and a test that ran against a different translation could be just as irrelevant as a test that ran against test data had been The primary test for the OSIS has been taking snapshots of the OSIS and after a change in the stylesheets or more rarely the SFM comparing the snapshot against the current OSIS using a difference checking stylesheet that would for example ignore differences in the date stamps in the two files We are also moving into using Schematron tests to make assertions about the OSIS XML There is a similar snapshot and compare me
136. e XQuery implementations op timization to understand the new functions An example XQuery module of rewrite rules can be found in Appendix B Consider the following XQuery 3 0 higher order function declare function map Sf as function item as item Sseq as item as item if empty seg then else f head seg map f tail seq be The map function provides an abstraction for performing some action on every member of a given sequence It is a very common function in popular functional languages like Haskell where it is frequently used This often leads through function inlining or programmer inefficiencies to nested invocations of the map function which it can be important to eliminate to remove intermediate results As a library writer implementing a function like map I might want to provide rules to handle unnesting of my function 5 declare option rw rule fn MapMapFusion map f map g e gt map function a map f g a e The rewrite rule identifies a nested call to the map function and uses an anonymous function to compose the two functions f and g Even though the above rewrite removes the large intermediate result the map function is still called on the result of the g function If that function only returns a singleton item the inner map can be completely eliminated with another rewrite rule declare option rw rule fn MapSingleton map f e where rw subtype
137. e actions needed to respond to a user input event This style of processing will I believe enable XSLT to be used to program most of the interactivity that is currently implemented in JavaScript It will need some further extensions it s not obvious to me yet for example how the stylesheet should update a single attribute of the HTML DOM nor how it should issue POST requests to the server It will also need the ability to interoperate with JavaScript in both directions None of these obstacles appear to be insuperable and the existing event based nature of the XSLT language appears to make it ideally suited to handling an event based user interface One area where JavaScript programming of a user interface gets very messy and error prone is when implementing a structured dialog such as a flight reservation a question and answer wizard where the system takes the user through a series of 132 XSLT in the Browser pages managing the URLs of the successive pages and the effect of the back button as it does so This area is currently seeing much exploration and innovation sites like Google and Twitter are experimenting with new ways of using fragment iden tifiers in URIs to capture state see for example Jeni Tennison s article 8 I hope in a later paper to show how XSLT can be used to tackle this problem The key I believe is to think of a dialog as a pair of interleaved documents one representing the communication from the system
138. e sake of this example let s say you are lazy and store them as plain text I ll spare you the details but the real meat in you XML database realm will then be to return the password and roles for a user with a given login name If you are using an XML database such as eXist with its REST API you will end up opening an URL with a Java statement such as Example 1 Naive REST URL construction new URL http localhost 8080 orbeon exist rest db app users query gt user mail 27 username 27 2 2 Attack Let s put a black hat and try to attack a site powered by an XML database that gives us a login screen such as this one Please enter your login and password Login Password Login Figure 1 Login Screen We don t know the precise statement used by the realm to retrieve information nor the database structure but we assume that the authentication injects the content of HTML form somewhere into an XQuery as a literal string and hope the injection is done without proper sanitization We don t know either if the programmer has used a single or a double quote to isolate the content of the input form but since that makes only two possibilities we will just try both The trick is 1 to close the literal string with a single or double quote 2 to add whatever is needed to avoid to raise an XQuery parsing error 3 to add the XQuery statement that will carry the attack e g to add again whateve
139. ed in the eXist configuration file and we can t be sure that the attack will work but if one of them is activated we ll be able to export the user collection An attack based on the mail module looks like the following lt foo xmlns mail http exist db org xquery mail gt let message lt mail xmlns util http exist db org xquery util gt lt from gt vdv dyomedea com lt from gt lt to gt vdv dyomedea com lt to gt lt subject gt eXist collection lt subject gt lt message gt lt text gt The collection is util serialize lt text gt lt message gt lt mail gt return mail send email message localhost lt foo gt A similar attack could send the content of the collection on pastebin com using the HTTP client module i http www w3 org TR xquery prod xquery Expr 180 XQuery Injection To inject the attack we concatenate the start container string or the attack itself and the end container string or normalize the spaces and paste the result into the login entry field The login screen will return a login error but if we ve been lucky we will receive a mail with the full content of the collection on which the query has been run If nothing happen we might have used the wrong quote and we can try again replacing the single quotes from our container string by double quotes If nothing happen once again which is the case with the naive REST URL con
140. ed components from remote and local software Akara offers a solid backbone for assembly of such heterogenous systems 12 Appendix A More background on 4Suite In order to better understand the spirit behind Akara it s useful to have historical perspective of its predecessor 4Suite which enjoyed very active development for the decade starting 1998 4Suite also spawned additional work and influence in numerous other areas for example serving as the core XML processing toolkit for Red Hat and Fedora Core distributions in the mid 2000s contributing components to the Python language serving as a reference implementation for development of RFC 3986 and thus influencing several other packages 4Suite and Akara have over the years provided several important innovations including e XML RDF triggered transforms helped inspire GRDDL e Path based RDF query mounted across an XML RDF repository Versa which inspired many others and was an input to W3C s SPAROL work e Rules based rather than type systems based data binding for XML and RDF e RDF query within XSLT e Push style data driven multiple dispatch to code e Pioneering implementations of DOM XPath XSLT XLink XPointer RELAX NG Schematron and more 7 http docs python org library sqlite3 html 79 80 Translating SPARQL and SQL to XQuery Peter M Fischer ETH Zurich lt peter fischer inf ethz ch gt Dana Florescu Oracle Corporation lt dana florescu oracle com gt Mart
141. ed in depth injections targeted on XQuery string literals What about other injections on XML based applications 3 5 1 XQuery Numeric Literal Injection It may be tempting to copy numeric input fields directly into XQuery expressions that s safe if and only if these fields are validated If not the techniques that we ve seen with string literals can easily be adapted in fact it s even easier for your attack ers since they do not need to bother with quotes That s safe if you pass these values within request parameters but you will generate XQuery parsing errors if the input doesn t belong to the expected data type Also note that request get parameter returns string values and may need to be casted in your XQuery query In both cases it is a good idea to validate numeric input fields before sending your query When using XForms this can be done by binding these inputs to numeric data types Otherwise use whatever language you are programming with to do the test If you use literals and don t want or can t do that test outside the XQuery query itself you can also copy the value in a string literal and explicitly cast it into the numeric data type you are using XQuery functions and operators The string literal then needs to be sanitized like we ve already seen 3 5 2 XQuery Direct Element Injection Literals are the location where user input is more likely copied in XQuery based applications they cover all the cases where the d
142. efined in 4 In NETCONF terminology server is the managed device and client is the usually remote manage ment application NETCONF uses XML for encoding both protocol operations and data contents Every NETCONF session is started by a pair of hello messages in which both parties advertize supported protocol version s and optional capabilities If the server supports the YANG data modelling language described below it will also advertize the data model that it implements Within a session the client and server commu nicate by using a remote procedure call RPC mechanism For instance the client has two RPC methods at its disposal for querying the server s datastore e get method asks for both configuration and operational state data e get config method asks for configuration data only The NETCONF message carrying the simplest form of the get method looks as follows lt rpc message id 123 xmins urn ietf params xml ns netconf base 1 0 gt lt get gt lt rpc gt An imaginary Internet enabled coffee machine could honour this reguest by sending the following reply 100 Configuring Network Devices with NETCONF and YANG lt rpc reply message id 123 xmlns urn ietf params xml ns netconf base 1 0 gt lt data gt lt configuration xmlns http example com coffee gt lt ipv4 gt lt address gt 192 0 2 1 lt address gt lt subnet mask length gt 24 lt subnet mask length gt lt ipv4 gt lt ipv6 gt lt
143. eguest is received by the webapp container the webapp is identified by the context root that is the first level of the URI Then all the patterns in the corresponding webapp descriptor are tried to be matched against the reguest path in order The reguest is dispatched to the first one that matches either a re source or a component The components can be anything among the following types Language Kind of component XProc Step Pipeline XQuery Function Main module XSLT Function Named template Stylesheet Each kind of component defines the exact way it is evaluated how the request is passed to the component and how the component gives back the response For in 42 CXAN a case study for Servlex an XML web framework stance an XQuery or an XSLT function must have exactly two parameters the first one is the web request element and the second one is the sequence possible empty of the entity content aka the request bodies The result of calling such a function must in turn give an element web response and possibly several subsequent items representing the response body An XProc pipeline is evaluated the same way but the specification defines instead specific port names for the request and the response 3 Servlex Servlex is an implementation of the Webapp Module It is open source and available on Google Code at http code google com p servlex Under the hood it is written in Java it uses the Java Ser
144. el of data representation use cases query languages As a consequence implementations rarely cover more than one model 13 Yet there is a need to overcome this separation and to integrate data and operations One possible solution would be a common runtime for all of these formats where each language can be exploited where it is suited best 1 2 Problem Statement In order to overcome the differences between the models we investigate if and how one language can be translated into another In this paper we focus on a translation from both SPARQL and SQL to XQuery which has seen little attention so far XQuery is an interesting target since it is the most expressive language and its implementa tions are now reaching maturity The translation is required to express the semantics correctly to cover all expressions and to create code that can be executed efficiently 1 3 Contributions In this paper we present the following results e acomplete and correct translation of SPAROL to XQuery which does not require any assumptions on the schema of the data or the particular workload e a sketch of a translation of SOL92 to XQuery again with no assumption on schema or workload e a working cross compiler which takes any SPAROL or SQL92 query and turns it into an XQuery expression e initial performance results which show that even with limited optimizations XQuery is typically as fast as native SPARQL and often faster In contrast it sti
145. elopers have been involved in XML and especially in XML processing with Python since the very beginning We ve seen it all and pretty much implemen ted it all At first the motivation was that XML seemed the best hope for semi structured database technology but by now XML has become just plumbing as used in countless domains including for many unsuited uses There are many XML processing libraries in Python and even the standard library finally has a respectable one with ElementTree So why a new XML pipeline and processing project especially one as ambitious as Akara The first answer is that it s not just about XML but even focusing on the XML processing kit the fact is most XML processing tools not just in Python but in general are entirely focused on the dumb plumbing These treat XML as a tem porary inconvenience rather than as a strategic technology This is often justified because most uses of XML by far are products of poor judgment where other technologies would have been far more suited But for those cases where XML is well suited briefly characterized as where traditional granular data combines with rich prosaic expression the current crop of tools is inadeguate Akara s developers want to be able to treat with XML above the level of plumbing to deal with it at the level of expression Used correctly XML is not an inconvenience and bears fruit when handled as richly and naturally as possible because the data in XML
146. ement These limitations are not crucial since they only slow down the editing speed but for a full function tool they would all need to be investigated Client side XML Schema validation Another restriction is the stability of the FSM on which the API is based In our case this is not an issue because we assume the FSM will not change during an editing session This does rule out parallel development of schema and document Insertion of an element with required content requires the construction of a valid skeleton sub tree to preserve overall validity this has not been implemented yet 4 5 Example In this subsection we provide a simple constructed example of the described al gorithms To illustrate them we consider the following simple W3C XML Schema lt xml version 1 0 gt lt xs schema xmlns xs http www w3 org 2001 XMLSchema gt lt xs element name address gt lt xs complexType gt lt xs sequence gt lt xs element name town type xs string gt lt xs element name street type xs string maxOccurs 4 gt lt xs element name flat type xs decimal minOccurs 0 gt lt xs element name room type xs string minOccurs 0 gt lt xs sequence gt lt xs complexType gt lt xs element gt lt xs schema gt We can see that the element address is of complex type and contains two required elements town and street and two optional elements flat and room In addition ele ment street can occur
147. enced 2 YANG allows NETCONF servers to omit empty containers from a reply to a get or get config request If such a container has descendant leaf nodes with default values then the omitted container effectively becomes part of the default content Consequently the DSRL schema must provide element maps not only for every leaf with a default value but also for all ancestor containers that may also be missing 4 5 Validating instance documents The DSDL schemas generated in the second mapping step can be used for validating instance documents of the type for which they were created Standard validators 112 Configuring Network Devices with NETCONF and YANG such as xmillint Jing or the reference implementation of ISO Schematron can be used for this purpose We are not aware of any implementation of DSRL so we wrote an XSLT stylesheet which transforms a subset of DSRL that is used by the mapping to XSLT The stylesheet is included in the pyang distribution see Section 4 6 Instance document validation proceeds in the following steps which are also illustrated in Figure 2 e The XML instance document is checked for grammatical and datatype validity using the RELAX NG schema e Default values for leaf nodes have to be applied and their ancestor containers added where necessary This step modifies the information set of the validated XML document e The semantic constraints are checked using the Schematron schema
148. ent chemistry color for value name in xgllib getData chemistry Selement chemistry name where fn exists value col and fn exists element and fn exists value_name return lt result gt lt var name col gt value_col lt var gt lt var name element gt value element lt var gt lt var name name gt value_name lt var gt lt result gt let SGRAPH 1 SGRAPH O fn matches var name name ium for node in SGRAPH 1 order by node var name col return Snode return xallib formatSparglXml 89 Translating SPARQL and SQL to XQuery Sresult position 5 to 14 col 4 Mapping and Translating SQL to XQuery Given that XQuery was designed to also handle relational data see Use Case R described in 20 and that the expressions can be fully nested the translation of selection projection inner joins ordering sub queries as well as updates is straightforward The translation only requires an adaptation of predicates and path expressions to the concrete serialization of relational data into XML For group by and outer join we consider both explicit nested for loops and the specialized con structs for XQuery 3 0 Null values are mapped to an empty sequence in evaluation and empty elements in the results since the Boolean evaluation rules are a close match Yet many differences in the data model and semantics make a fully correct translation rather complex
149. ent rdf ID H gt lt name gt hydrogen lt name gt lt number gt 1 lt number gt lt Element gt lt Gas rdf ID H2 gt lt element rdf resource H gt lt weight gt 2 lt weight gt lt Gas gt lt rdf RDF gt 2 2 The SPAROL Ouery Language SPARQL 16 is often referred to as the query language for RDF The basic operation is graph pattern matching in particular triple patterns in which subject predicate and or object may be variables These patterns can be combined using the operators AND UNION OPT and FILTER yielding solution sequences actually unordered bags which then can be changed by solution modifiers such as DISTINCT ORDER BY REDUCED LIMIT and OFFSET SPAROL defines four query forms SELECT ASK CONSTRUCT and DESCRIBE Example 2 shows a SELECT query which retrieves the color of all elements ending in ium and returns the 4 to 14 color after ordering Example 2 SPARQL example query PREFIX chemistry lt http www xql2xquery org chemistry gt SELECT col WHERE 2 lement chemistry name name element chemistry color col FILTER REGEX name ium ORDER BY col 83 Translating SPARQL and SQL to XQuery LIMIT 10 OFFSET 4 2 3 XQuery XQuery is a declarative and Turing complete programming language which was originally designed to extract information and perform transformations on XML data It uses XDM as its data model which expresses sequences of ato
150. er gt 3 lt stateVersion gt lt time type string gt notime lt time gt lt xAxisOption type string gt NOTHING lt xAxisOption gt lt playDuration type number gt 15 lt playDuration gt lt iconType type string gt BUBBLE lt iconType gt lt sizeOption type string gt NOTHING lt sizeOption gt lt xZoomedDataMin type null gt lt xZoomedIn boolean false gt lt duration type object gt lt multiplier type number gt 1 lt multiplier gt lt timeUnit type string gt none lt timeUnit gt lt duration gt 27 A JSON Facade on MarkLogic Server lt yZoomedDataMin type null gt lt xLambda type number gt 1 lt xLambda gt lt colorOption type string gt NOTHING lt colorOption gt lt nonSelectedAlpha type number gt 0 4 lt nonSelectedAlpha gt lt dimensions type object gt lt iconDimensions type array gt lt dimensions gt lt yZoomedIn boolean false gt lt yAxisOption type string gt NOTHING lt yAxisOption gt lt yLambda type number gt 1 lt yLambda gt lt yZoomedDataMax type null gt lt showTrails boolean true gt lt xZoomedDataMax type null gt lt json gt Remember this isn t a format you re expected to see or utilize It s shown just to elucidate how MLJSON Works internally The root element is lt json gt That s always the case The root type is object as specified by the type attribute Objects in JSON behave like maps with n
151. er Accept to differenciate between both Internally the pipelines use a simple XML format to flow between the steps describ ing packages tags categories etc The last step in most pipelines checks the value of the Accept header if it is application xml it sends the XML back to the user as is if not it first transforms it to an abstract page This way the webapp provides almost for free a REST like API in addition to the traditional HTML website A client XProc application uses this API to provide a command line utility to the user in order to maintain packages in a local repository automatically installed and upgraded from the CXAN website there is a screenshot of this cxan comand line utility in the introduction 5 The development project How is organized the source code of this web application The project directory structure is as follows cxan website dist cxan website 1 0 0 xar cxan website 1 0 0 zip src images pages home xproc lib tools xpl page xsl xproject 51 CXAN a case study for Servlex an XML web framework expath web xml project xml The overall structure follows the EXPath project directory layout The xproject directory contains information about the project as well as the web descriptor in case of a webapp project the src directory contains the source of the project and dist is the directory were final packages are placed In addition components must contain the public URI to u
152. er accepts These algorithms enable a full implementation of W3C XML Schema However they are expensive in terms of space when handling nested numerical occurrence ranges In 7 H S Thompson defines an extended FSA with ranges to cope with this problem H Thompson and R Tobin also provided the open source XSV validator We have exploited the XSV validator in accordance with WEB 2 0 standards that is ef ficiency modularity lightness security 5 by converting its core validation code from Python to JavaScript 2 3 JavaScript Restrictions JavaScript is a client side language and was initially designed to provide dynamic websites 1 However modern business requirements expanded the use of JavaScript for content management business application systems etc We adopted the following restrictions e Platform Limitations JavaScript does not support direct access for manipulation of the user system and files except for cookie files it cannot use databases and access other domain files It uses a sand box scripting strategy when the code is executed in a certain environment without the risk to damage the system as well as the same origin policy so that a script in one domain has no access to another domain s information To overcome these limitations it is necessary to Client side XML Schema validation use auxiliary server side applets e g PHP Java applets etc or ad hoc techniques e g iframes dynamic lt scr
153. erwise not allowed Note that while YANG makes an explicit difference between leaf and container elements it has no means for modelling XML attributes The scoping rules for node identifiers also enforce another simplification of the generic XML model nodes with identical names cannot be nested In particular this constraint eliminates recursive structures The selection of YANG built in types is roughly comparable to that of XSD Datatype Library 16 However YANG allows for deriving new named datatypes from the built in ones perhaps in multiple steps by specifying additional restric tions or facets in the XSD terminology A default value may be specified for a non mandatory leaf node as well as for a datatype In the latter case the default value applies to all leaf nodes of that datatype unless they define their own default value Semantic constraints may be specified using the must statement Its argument is an XPath expression that has to evaluate to true possibly after the result is con verted to a boolean value The context for the XPath expression is the data tree in which the must statement appears i e a datastore RPC operation or notification Also all missing leafs that define a default value have to be conceptually added with the default value before the XPath evaluation takes place Some parts of a YANG module can be declared as conditional and their presence or absence depends on either e static parameters known as
154. ery patternSPA patternxQu triplePatternspa foreach subjName subjVars patternspa a for SsubjName in xgllib getSubj triplePatternspa foreach predName predVars patternspa filterspa for SpredName in xgllib getPred subjName foreach objName objVars patternepa for SobjName in xallib getObj predName where foreach constant constants patternspa subjName predName objName SsubjName constant SpredName constant SobjName constant foreach filterCondition filterXgu and filterCondition return lt result gt foreach varName vars patterngpa lt varName gt data varName lt varName gt lt result gt 3 2 Graph Pattern Combination 3 2 1 Optional Patterns The purpose of an optional pattern is to supplement the solution with additional information If the pattern within an OPTIONAL clause matches the variables defined by that pattern are bound to one or many solutions If the pattern does not match the solution remains unchanged The optional pattern is implemented in XQuery by a binary function which implements a left outer join over the intermediate graph representations Since the OPTIONAL keyword is left associative the rule can be applied repeatedly to handle multiple consecutive optional patterns 3 2 2 Alternative Graph Pattern In an alternative graph pattern two possible patterns are evaluated and the union of both is taken as a result T
155. ery request gt user mail request get parameter mail 0 UTF 8 String usernameNormalized URLEncoder encode username trim toLowerCase gt UTE 8 reader parse new InputSource new URL http localhost 8080 orbeon exist rest db app users mail gt usernameNormalized 8 guery query openStream To guery is now a fixed string that could be stored in the eXist database or encoded in a static variable 3 3 XPath 2 0 Environments In environments that relies on XPath 2 0 such as XSLT 2 0 XProc XPL the same patterns can be used if we replace the Java methods by their XPath 2 0 eguivalents In XSLT 2 0 it is possible to define a sanitize function similar to the one we ve created in Java but this isn t the case for other host languages and we ll skip this step To sanitize user inputs in an XPath 2 0 host language we need to add a level of escaping because the amp character is not available directly but through the amp amp entity reference The XQuery query combines simple and double quotes that are not very easy to handle in a select attribute even if the escaping rules of XPath 2 0 help a lot and the query pieces can be put into variables for convenience That being said the user inpu can be sanitized using statements such as Example 5 XQuery Sanitized in XSLT lt xsl variable name usernameSanitized select lower case normalize space replace replace username
156. ess to default values This was done by implementing a visitor that walks over the schema patterns and collects the default values in a map It is possible to have this map because this is a feature from the DTD compatibility specification and the same attribute from the same element cannot have different default values This gives an easy and guick access to the default values 2 3 Integrate with an XML Parser Xerces Xerces has an internal API called Xerces Native Interface XNI All the parser components work at this level This mainly specifies a pipeline where the different components are placed starting with the scanner and then containing DTD or XML Schema validators XInclude handler etc There parser processing pipelines are created by a parser configuration class We create such a parser configuration that injects in the XNI pipeline a component that adds the default values similar with how these default values are added by the DTD or XML Schema validators Thus all the further processing will see no difference no matter if the values were added by the standard Xerces DTD or XML Schema validator components or by our Relax NG default values processor component The component looks for a processing instruction that should appear before the root element and that should specify the associated Relax NG schema Right now it supports the oXygen PI used to associate Relax NG schemas with documents If a schema is detected then the support for
157. eteness and compliance more than 98 6 percent correct tests at the 1 0 2 XQuery Test Suite We have not yet done formal testing on Update and Full text but plan to do so in the near future Keywords XML XQuery Browser 1 Motivation Currently there is a growing perception that the Web and XML communities are drifting apart One general concern is that up to date XML technologies such as XSLT 2 0 or XQuery 1 0 are not seeing any support in the browsers thus negating much of their potential Web pages are based on HTML XML and HTML are both derived from SGML and therefore have many similarities This is why programming languages for XML can also be used for HTML with some adjustments We have proposed the use of XQuery as a client side programming language XQuery seamlessly supports HTML navigation and updates and as it is already used on the server business logic database querying moving code between the layers is almost straightforward Last year at XML Prague 2010 8 the XQuery in the Browser plugin 7 was presented as a possible solution It provides full XQuery support on the client side by embedding the Zorba XQuery engine 6 into anumber of contemporary browsers While the applications and usability were convincing using a binary plugin was seen as insurmountable obstacle to a wider adoption since even well established plugins like Flash or Java are no longer available on major platforms e g on the growing number o
158. eviously element Seventname as added through b addEventListener xs string listener as from the element s Swhere xs QName b alert message as xs string Displays a message box with the content Smessage Additionally we also support the EXPath HTTP library for data retrieval It has the same limitations as fn doc which are described in the next paragraph 197 XQuery in the Browser reloaded 3 4 Functionality with different semantics As there is no file system in the browser the semantics of module import with a file location hint and of fn doc Suri as xs string are defined as follows If a rel ative URI is provided the library will download the file automatically using an XMLHttpRequest If an absolute URI is given the library will also try to retrieve it This may fail due to security constraints namely the same origin policy which only allows re quests coming from one page to access the same host at the same port using the same protocol as was used on that page This policy may be circumvented with HTTP headers As we use the JavaScript classes for regular expressions not all options are supported The option for dot all matching is currently not available because it is not supported by the JavaScript RegEx classes but an emulation will probably be provided in the future 3 5 Not implemented functionality We have chosen not to implement schema support to keep the code smaller and because we were
159. extreme variations when testing The dominant numbers for a certain browser have been chosen They where taken on a system with average performance for the year 2010 Phenom II X6 2 8 GHz Windows 7 x64 In order to eliminate network overhead the page was served on localhost using Apache First some performance numbers from a simple test page lt DOCTYPE html gt lt html xmlns http www w3 o0rg 1999 xhtml gt lt head gt lt title gt X0IB Sample page lt title gt lt meta charset UTF 8 gt lt link href style css rel stylesheet type text css gt lt script type text javascript gt var time start Number new Date lt script gt lt script type text javascript src mxaueryjs mxaueryjs nocache js gt lt script gt lt script type application xquery gt insert node lt div gt let x b dom hl return xs string x lt div gt as last into b dom body lt script gt lt head gt lt body gt lt body gt lt html gt 207 XQuery in the Browser reloaded The time is measured from the first script tag to the end of the query execution Therefore some final rendering might not be measured Table 3 Load times Browser Firefox 3 6 Firefox 4 Chrome 8 Internet Ex plorer 9 Time ms 230 200 120 140 These values seem very high but when considering that there are many optimiza tions which can be done they are quite good We have to take into cons
160. f Handling this server side allowed us to re use the compilation phase of XSV an existing W3C XML Schema processor e Our focus is on supporting client side instance authoring environments where schema change is infrequent As well as loading XML documents and schemas client side applications can initiate validation and query validation state via an API In our existing prototype the ap plication is a simple WYSIWYG editor as shown on Figure 1 Our implementation so far focuses on the validation framework the editing tool is essentially a test application The validation framework starts from a JavaScript implementation of the validation core of the XSV validator 9 The JavaScript im plementation uses the same validation algorithm and structure Editing functionality is described in section 4 The testbed UI consists of XML and W3C XML Schema URL input fields The application requests reflected schema that is an object model of the corresponding schema from the server and the validation engine starts working If the document is not valid the application displays the errors and further editing is not possible Client side XML Schema validation R XML Document Is XML Schema 3 Figure 1 Implemented System Architecture Otherwise the system offers the following editing operations deletion of attributes and elements insertion of attributes and elements into the selected element and insertion of elements before or
161. f mobile devices Instead browser vendors have been investing significantly into the quality of their JavaScript implementations 1 3 9 achieving orders of magnitude better performance Asa result JavaScript has become a viable platform for implementing XQuery Since writing an XQuery engine from scratch is a major effort we opted for translating MXQuery 4 an existing Java based engine using Google s Web Toolkit 2 A similar albeit independent approach has been taken by Michael Kay 10 The target language is XSLT 2 0 instead of XQuery yet the overall approach design and results are very similar to ours 2 Current approaches for client side programming 2 1 Container based approaches Java Flash Silverlight For a long time the most popular approach of programming complex applications in the browser has been to use a self contained runtime environment like Java Flash 192 XQuery in the Browser reloaded or Silverlight While such an approach provides high performance and effective developer support it does not integrate well with HTML To make matters worse the runtimes have to be downloaded installed and updated separately on most platforms On most mobile devices they are not available at all on desktop system privileged user rights are often required for installation 2 2 Javascript DOM Events Frameworks JavaScript is nowadays by far the most commonly used programming language for client side website p
162. features or arbitrary dynamic conditions specified in an XPath expression subject to the same rules as for the must statement 104 Configuring Network Devices with NETCONF and YANG An example of a feature is the availability of permanent storage in the configured device which may enable certain functions Active features are declared in NETCONF hello as a part of the data model specification Reusable collections or subtrees of data nodes may be defined in YANG by means of groupings They are used in a similar way as for instance named patterns in RELAX NG with three notable differences 1 YANG groupings can only contain entire data node definitions not just arbitrary statements 2 When a grouping is used its contents may be modified for example by adding new data nodes at any location inside the data hierarchy defined by the grouping 3 The names of groupings are qualified with the namespace of the module where each grouping is defined and may also be scoped Moreover the names of data nodes defined by a grouping always belong to the namespace of the module in which the grouping is used not the one in which it is defined These naming issues are further discussed in Section 4 2 1 From the viewpoint of standard XML schema languages the single most exotic feature of YANG is the augment statement Its main purpose is to modify an existing module from outside by adding new data nodes at any location inside the hierarchy
163. fn QName para Hello World lang en 1 cts element range guery fn QName price lt 15 cts element pair geospatial query fn QName location fn QName latitude fn QName longitude cts point 53 12 cts circle 10 53 12 cts box 5 6 3 4 cts polygon 53 12 57 15 53 12 coordinate system wgs84 1 cts collection query recent 6 REST Interface We ve now seen how JSON is mapped to XML and how the XML is queried using a JSON query syntax To glue it all together MLJSON exposes a REST web service interface to handle the loading deleting modifying and querying of JSON docu ments held in MarkLogic The first REST endpoint is jsonstore xgy It provides basic insert fetch and delete capabilities Table 2 Sample REST URLs Insert a document jsonstore xgy uri http foo bar PUT Delete a document jsonstore xgy uri http foo bar DELETE Get a document GET jsonstore xgy uri http foo bar You can also use the jsonstore xgy endpoint to set properties name value associ ations held as document metadata set collections a named grouping assign permissions security and dictate guality inherent search relevance Table 3 Setting Attributes on Documents Set property POST jsonstore xgy uri http foo bar amp property foo bar Set permissions jsonstore xgy uri http foo bar
164. for instance and do not need to be considered during the rewrite Another issue is how to express generic rewrites on parts of FLWOR expressions Important rewrites like where predicate push back and loop lifting need to be made against FLWOR expressions but the clauses of a FLWOR expression are not themselves expressions in the XQuery grammar This means that rewrite rules as currently described must match an entire FLWOR expression which reduces their applicability Further complications lie in other XQuery language irregularities like reverse axis predicates and a lack of literal xs boolean values The rewrite rules in XQilla currently rely on predicates to detect such reverse semantics predicates and treat the pattern true as matching the literal xs boolean value rather than a function call If a rewrite wanted to match against an implicit operation like effective boolean value or xs untypedAtomic type promotion portability suffers XQilla allows a pattern of boolean to match its effective boolean value operation but no such standard XQuery function exists for xs untypedAtomic type promotion A likely solution is to create a rewrite rule specific pseudo function to match against that operation if desired 2 Moving the execution of a predicate to earlier in the query to reduce the amount of data that gets processed 220 Declarative XQuery Rewrites for Profit or Pleasure 5 Conclusion There are still a great many im
165. ggered by an event we get different results The execution is now much faster on all browsers This demonstrates that the per formance even at this early stage is already sufficient for dynamic websites 5 3 2 Download Size For loading a page two files have to be loaded First the dispatcher file mxqueryjs nocache js which is under 6 kB in size This will then select the file with the actual code depending on the browser version This file is about 700 kB in size By enabling gzip compression on the server the transferred data can be reduced to 200 kB In addition this code can be cached making subsequent access almost instantaneous 6 Conclusion We have shown that it is possible to build an XQuery engine on top of JavaScript without major performance or functionality penalties This allows to use XQuery for browser programming XQuery already has a large user base which comes mainly from the database and XML communities This enables them to write web applications in a language which is familiar to them 7 Future work We consider the following directions for future work e Integration of the Javascript branch back to MXQuery mainline for better long term maintenance e Improved browser support investigating if Internet Explorer 8 might be a feasible target given its high market share e Performance optimizations in particular fully using indexed access to the DOM e Integration of JSON e g like the upcoming parse
166. gt lt sch rule gt The context attribute has to be set to the actual context for the rule which in this case consists of the root node nc rpc reply provided by the NETCONF Messages layer plus the entire hierarchy of nodes down to cm ipv6 In general the Schematron schema contains one sch pattern element for every input YANG module It may also contain abstract patterns which are used for map ping semantic annotations appearing inside named pattern definitions in the hybrid schema Such semantic annotations may be applied in multiple contexts possibly also in different namespaces corresponding to the places in which the containing named pattern is referenced Because of this abstract patterns use two variables e pref represents the namespace prefix of the module in which the named pattern is referenced e start represents the initial part of the context path corresponding to the place in which the named pattern is referenced Such an abstract pattern may look like this lt sch pattern abstract true id example sorted leaf list gt lt sch rule context start Spref sorted entry gt lt sch report test preceding sibling Spref sorted entry gt Duplicate leaf list entry lt sch value of select gt lt sch report gt lt sch assert test not preceding sibling pref sorted entry 6gt gt Entries must appear in ascending order lt sch assert gt lt sch rule gt lt sch pattern gt
167. gt by lt a href s gt linked b poet lt a gt lt p gt uri name puri A routine to generate HTML of the information for a single work Notice how amara lib iri joinis used to construct links getters u work get work u poet get poet ust a way to package up the reusable routines for generating poet and work info ytop ge up 8 sP Return an index of works or poets works for work id name poet in POETDB u work iteritems works append getters entitytype work id return n join works Go through the index of works and return an aggregate HTML from the fragments 10 Handling HTTP POST The above example handles HTTP GET and of course POST is a big part of the Web It s best known for Web forms though Akara is not specialized for such usage in the way more mainstream Web frameworks are CherryPy Django etc You can use Akara to handle Web forms but more often Akara users will be dealing with data services often using requests directly POSTed to the endpoint This is a common pattern for open Web APIs such as those of social networks Since POST on the Web is generally used in cases where state of Web resources are changing this is usually the area where you need to deal with some sort of persistence in your application You ll see an example of that in this section moving from the in memory data structure of the previous section to something more serious You ll also see an example of ho
168. h the Monk Bibles that preceded it Cory Doc torow writes Luther Bibles lacked the manufacturing quality of the illuminated Bibles They were comparatively cheap and lacked the typographical expressiveness that a really talented monk could bring to bear when writing out the Word of God none of the things that made the Gutenberg press a success were the things that made monk Bibles a success By the same token the reasons to love ebooks have precious little to do with the reasons to love paper books 6 Given our experiences with current e readers it is just as well there are reasons beside typesetting elegance to love ebooks However the pioneers of the printing revolution also struggled to get the best out of young temperamental technology Despite the frustrations and the constraints of this fledgling medium we have found standard XML tools such as XSLT 2 0 to be flexible and productive in this application domain We hope that our work can form the basis of an open source project which can enable the Pixel Bible to do things that were never possible with the Luther Bible and Monk Bible Bibliography 1 Adobe Digital Editions http www adobe com products digitaleditions 2 Amazon com Announces Fourth Quarter Sales up 36 to 12 95 Billion http phx corporate ir net phoenix zhtm c 97664 amp p irol newsArticle amp ID 1521089 3 About Bible Tech Group http bibletechnologies net AboutBTG dsp 4 calibre E book man
169. have an object at the top of the hierarchy containing a name of foo with a value of bar key foo value bar It matches this JSON structure abc xyz foo bar It s equivalent and internally gets mapped to the following XPath json foo 5 Shar You don t generally see this XPath of course It s generated as part of the jsonquery execute function declare function jsonquery execute json as xs string as element json Here s a sample call import module namespace jsonquery http marklogic com json query at mljson lib json guery xgy jsonguery execute key foo value bar The JSON guery syntax supports much more advanced expressions You can for example match a seguence of values key foo value bar auux To find a match anywhere use innerKey innerKey foo value bar guux Drop the explicit value requirement and it means any value is allowed key foo Constraints can be hierarchical where the value contains another object key foo value key id value 0596000405 Same as json foo id 0596000405 You can use or or and to enable more complex matches key book or key id value 0596000405 key other id value 0596000405 To achieve a comparison other than equality specify a comparison value 30 A JSON Facade on MarkLogic Server key price value 8 99 comp
170. he alternative pattern can be expressed in XQuery by a sequence of the results of both patterns since UNION does not specify duplicate elimination 86 Translating SPARQL and SQL to XQuery 3 2 3 Group Graph Pattern All graphs associated in a group graph pattern must match The pattern is imple mented by an XQuery function that correlates the groups on shared variables using a join and the other function capturing equality in SPAROL The intermediate results generated by SPARQL patterns are combined by means of custom XQuery functions The mapping is shown in the following table Table 2 Translation of graph pattern combinations from SPARQL to XQuery patternSPA patternXQu patternLspa xqllib optional patternLyo OPTIONAL patternRyo patternRopa patternLspa patternLlyp patternRyo UNION patternRspa patternLgp xqllib and patternLyo patternRspa patternRyo 3 3 Filter A SPARQL FILTER function can be added to graph patterns in order to restrict the result according to a Boolean condition In the running example elements whose name end in ium are filtered according to a regular expression Since comparison operations effective Boolean value and several other functions are actually defined by the related XPath 2 0 functions and operators we can use XQuery value compar ison eq neq Yet we need to consider the differences in Boolean logic and error handling S
171. he browser not the user application and the browser provides no hook that would cause this to invoke a third party XSLT engine Instead I m currently using a lt script gt element of the form lt script src stylesheet xsl type application xml xslt input source xml gt lt script gt This works but it s not ideal Firstly it s not conformant HTML5 the input attribute is not permitted though of course browsers don t reject it Secondly it seems that the browser itself fetches the XSLT code and adds it to the HTML DOM which isn t especially useful since the HTML DOM doesn t expose details such as namespaces and may do the wrong thing when the stylesheet contains HTML elements such as lt br gt So Saxon currently ignores the fact that the stylesheet has already been fetched and fetches it again using an XmlHttpRequest I m considering replacing this with a mechanism that uses a lt script gt element such as lt script id xslt transformation type application xml gt lt stylesheet href stylesheet xsl gt lt source href source xml gt lt output href body gt lt param name p value 3 gt lt initial mode name mode a gt lt script gt where the contents of the lt script gt element would not be parsed by the browser but retained as text in the HTML DOM This will reguire experimentation to establish what works across a range of browser platforms Although the details may change I m keen to h
172. he module a bit low level but as we have seen this is very easy with technologies like XSLT to create an intermediary layer of abstraction in order to be able to write the web components at a higher level of abstraction Because it provides a full consistent mapping of HTTP natively oriented towards XML it never locks its users in some restrictions because of some handy abstraction which does not fit all use cases Because it provides the full HTTP protocol information to the XML technologies it can be used to easily create websites REST web services SOAP WSDL web ser vices and everything you can do on a HTTP server And thanks to Servlex such applications can be hosted for free on Google Appengine or other cloud services like Amazon s 52 Akara Spicy Bean Fritters and XML Data Services Uche Ogbuji Zepheira LLC lt uche ogbuji net gt Abstract Akara is a platform for developing data services and especially XML data services available on the Web using REST architecture It is open source software Apache 2 licensed written in Python and C An important concept in Akara is information pipelining where discrete services can be combined and chained together including services hosted remotely There is strong support for pipeline stages for XML processing as Akara includes a port of the well known 4Suite and Amara XML processing components for Python The version of Amara in Akara provides optimized XML processing using
173. he strength of RELAX NG lies in validation rather than in data modelling After 18 months of intensive development YANG version 1 was published in Octo ber 2010 10 together with a collection of essential data types 11 The primary syntax of YANG is compact and resembles the C programming language or BSD style configuration syntax see the examples below An alternative XML syntax named YIN is also defined as a part of the standard YANG data models are structured into modules Every YANG module declares an XML namespace to which all data nodes defined by the module belong A par ticular data model to be used for a given NETCONF session consists of one or more modules that are advertized in the hello message Modules may also be further subdivided into submodules which share the namespace of their parent module The names of all modules and submodules that will be published by the IETF are guaranteed to be unique As we will see YANG is in many respects similar to the existing XML schema languages However there is an important general difference which has to do with the target of the data model While XML schema languages assume a specific instance XML document to be validated against a schema YANG data models and the asso ciated rules address several different instance objects 1 Various configuration datastores such as running candidate and possibly others The semantic rules for the running datastore slightly differ from other data
174. hed 1 which should be soon followed by models for essential system and routing parameters gt http code google com p pyang 7 http hwww yang central org twiki bin view Main DSDLMappingTutorial http www exslt org dyn functions evaluate http www ops ietf org netconf 114 Configuring Network Devices with NETCONF and YANG Bibliography 1 Bjorklund Martin A YANG Data Model for Interface Configuration Internet Draft draft bjorklund netmod interfaces cfg 00 8 December 2010 IETF work in progress http tools ietf org html draft bjorklund netmod interfaces cfg 00 2 Document Schema Definition Languages DSDL Part 1 Overview Preparatory Draft 14 November 2004 ISO IEC http www dsdl org 0567 pdf 3 M nz Gerhard Claise Benoit Aitken Paul Configuration Data Model for IPFIX and PSAMP Internet Draft draft ietf ipfix configuration model 08 25 October 2010 IETF work in progress http tools ietf org html draft ietf ipfix configuration model 08 4 Enns Rob Ed NETCONF Configuration Protocol RFC 4741 December 2006 IETF http tools ietf org html rfc4741 5 Wasserman Margaret Goddard Ted Using the NETCONF Configuration Protocol over Secure SHell SSH RFC 4742 December 2006 IETF http tools ietf org html rfc4742 6 Goddard Ted Using NETCONF over the Simple Object Access Protocol SOAP RFC 4743 December 2006 IETF http tools ietf org html rfc4743 7 Lear
175. his section we provide a description of the translation of SPARQL to XQuery For space reasons we only show the general idea and the most relevant parts of the translation The full set of rules is available at 10 We define a function spargl2xguery which takes a SPAROL query as an argument and returns the corresponding XQuery representation as a result The following translation tables show the SPARQL code INSPA in the left column and the corresponding XQuery 84 Translating SPARQL and SQL to XQuery code OUTXQu in the right column The concepts shown in this section are demonstrated by means of the sample shown in Example 2 3 1 Basic Graph Pattern Matching Matching triple patterns is the core operation of SPARQL out of which more complex graph combinations can be built These patterns in turn can be filtered and modified Triple patterns such as element chemistry name name contain specifications for subject predicate and object of an RDF triple These specifications can be either constants to match or variables which are bound Our translation maps these variables to XQuery variables and generates a sequence of bindings for all variables In the result every element contains a single value for each of the variables without eliminating duplicates see Example 3 While this is very similar to the tuple stream in an XQuery 3 0 FLWOR expression we explicitly materialize these bindings which enables this intermediate result t
176. ideration that we are analysing load times which are usually dominated by bandwith latency constraints While the script is running a browser can continue to download images included on the page To test events and dynamic pages we have also tested a modified version which does about the same as the first one but triggered by a mouse click lt DOCTYPE html gt lt html xmlns http www w3 org 1999 xhtml gt lt head gt lt title gt X0IB Sample page lt title gt lt meta charset UTF 8 gt lt link href style css rel stylesheet type text css gt lt script type text javascript src mxqueryjs mxqueryjs nocache js gt lt script gt lt script type application xquery gt declare updating function local handler Sevt loc insert node lt div gt let x b dom hl return xs string x lt div gt as last into b dom body he b addEventListener b dom input onclick xs QName local handler lt script gt lt head gt lt body gt lt input type button value Button gt lt body gt lt html gt The time is measured from the beginning of the event handler until the event handler returns 208 XQuery in the Browser reloaded Table 4 Script execution times Browser Firefox 3 6 Firefox 4 Chrome 8 Internet Ex plorer 9 Time for first 15 15 30 30 run ms Time for sub 15 15 5 10 sequent runs ms When executing the same script as tri
177. iguration gt lt interleave gt lt element name cm ipv4 gt 116 Configuring Network Devices with NETCONF and YANG lt interleave gt lt element name cm address gt lt ref name ietf inet types ipv4 address gt lt element gt lt optional gt lt element name cm subnet mask length gt lt data type unsignedByte gt lt param name minInclusive gt 0 lt param gt lt param name maxInclusive gt 32 lt param gt lt data gt lt element gt lt optional gt lt interleave gt lt element gt lt oneOrMore gt lt element nma key cm address name cm ipv6 gt lt element name cm address gt lt ref name ietf inet types ipv address gt lt element gt lt a documentation gt At least one IPv6 address is required lt a documentation gt lt optional gt lt element name cm subnet mask length nma default 64 gt lt data type unsignedByte gt lt param name minInclusive gt 0 lt param gt lt param name maxInclusive gt 128 lt param gt lt data gt lt element gt lt optional gt lt element gt lt oneOrMore gt lt interleave gt lt element gt lt optional gt lt element nma implicit true name cm state nma config false gt lt interleave gt lt optional gt lt element nma implicit true name cm supply levels gt lt interleave gt lt optional gt lt element nma implicit true name cm water gt lt ref name coffee machine state supply levels per
178. ikins Factonomy Ltd lt aleksejs goremikins factonomy com gt Henry S Thompson The University of EdinburghFactonomy Ltd lt ht inf ed ac uk gt Abstract In this paper we present a prototype Javascript based client side W3C XML Schema validator together with an API supporting online interrogation of validated documents We see this as enabling an important improvement in XML based client side applications extending as it does the existing datatype only validation provided by XForms to structure validation and supporting the development of generic schema constrained client side editors Keywords XML XML Schema JavaScript Continuous Validation 1 Introduction One key gap in the integration of XML into the global Web infrastructure is valida tion DTD validation is supported natively to different extents by different browsers and some Web protocols notably SOAP explicitly rule it out Support for more re cent schema languages is virtually non existent With the growth of interest in rich client based applications in general and the XRX methodology in particular with its emphasis on XML as the cornerstone of the client server interaction architecture this gap has become more significant and its negative impact more troublesome Client side editing and validating of XML documents in browsers using WYSI WYG text editors is a growth area of Web 2 0 These systems are typically used to create collaborative websites in knowledge
179. ime by prepending the full schema context in the form of a sequence of names of all ancestor containers Several examples of mangled names may be found in Section A 1 Another feature of YANG groupings is that their contents adopts the namespace of the module in which the grouping used In other words if module A imports and uses a grouping from module B all data nodes defined inside the grouping belong to the namespace of module A This is the so called chameleon design 17 which is also possible in RELAX NG but requires a special arrangement of the including and included schemas Consequently the mapping procedure has to follow this arrangement when generating the validation RELAX NG schemas 1 Groupings with the global scope i e those that may be imported by other modules must be collected from all input modules mapped and stored in a separate RELAX NG schema file 2 Every input module is mapped to an embedded grammar which has the local module namespace as the value of its ns attribute The schema with global definitions is then included into this embedded grammar and the grouping contents adopt the namespace declared in the ns attribute 4 2 2 Augments In order to simulate the effects of an augment statement the mapping procedure must explicitly add the new nodes as element patterns to the RELAX NG schema at the location pointed to by the argument of the augment statement Augments also influence the handling of grouping
180. in XQilla an open source C implementation of XQuery and the majority of XOilla s optimizations have been rewritten using rewrite rules As an example XQilla contained this code to optimize the effective boolean value operation fn boolean ASTNode PartialEvaluator optimizeEffectiveBooleanValue XQEffectiveBooleanValue item item gt setExpression optimize item gt getExpression const StaticAnalysis amp sa item gt getExpression gt getStaticAnalysis if sa getStaticType getMax 0 If there are no items EBV returns false ASTNode result XOLiteral create false context gt getMemoryManager item gt staticResolution context item gt release return result if sa getStaticType getMin gt 1 amp amp sa getStaticType isType TypeFlags NODE If there is one or more nodes EBV returns true ASTNode result XQLiteral create true context gt getMemoryManager item gt staticResolution context item gt release return result if sa getStaticType getMin 1 amp amp sa getStaticType getMax 1 amp amp http xqilla sourceforge net HomePage 215 Declarative XQuery Rewrites for Profit or Pleasure sa getStaticType isType TypeFlags BOOLEAN If there is a single boolean EBV isn t needed ASTNode result item gt getExpression item gt setExpression 0 item gt release
181. in Kaufmann ETH Zurich lt martin kaufmann inf ethz ch gt Donald Kossmann ETH Zurich lt donald kossmann inf ethz ch gt Abstract In our community there are three main models for representing and processing data Relations XML and RDF Each of these models has its sweet spot for applications and its own guery language very few implementations cater for more than one of these We describe a uniform platform which provides inter faces for different guery languages to retrieve and modify the same information or combine it with other data sources This paper presents methods for com pletely and correctly translating SOL and SPAROL into XOuery since XQuery provides the most expressive foundation Early results with our cur rent prototype show that the translation from SPARQL to XQuery already achieves very competitive performance whereas there is still a significant performance gap compared to SQL Keywords SQL SPAROL XQuery Common Runtime 1 Introduction 1 1 Background Today three common approaches of representing data in structured and formal form are being used Relational tables XML trees and RDF graphs While rela tional and XML data have already been applied widely for a long period RDF is 81 Translating SPARQL and SQL to XQuery now gaining popularity among others in the contexts of semantic web or social networks These three approaches not only differ in terms of their data models but also at the lev
182. instance if the component is an XQuery function the request is passed as a function parameter if the component is an XSLT stylesheet the request is passed as a stylesheet parameter The result of the evaluation of the component must be the XML representation of the HTTP response to send back to the client The HTTP request looks like the following lt web request servlet package path pkg google apis method get gt lt web uri gt http cxan org pkg google apis extra param lt web uri gt lt web authority gt http cxan org lt web authority gt lt web context root gt lt web context root gt lt web path gt lt web part gt pkg lt web part gt lt web match name id gt google apis lt web part gt 39 CXAN a case study for Servlex an XML web framework lt web path gt lt web param name extra value param gt lt web header name host value cxan org gt lt web header name user agent value Opera 9 80 gt lt web request gt As you can see this XML document contains all information about the HTTP request The HTTP method of course GET POST etc and everything related to the request URI the full URI but also its authority part the port number the context root and the path within the web application The webapp map can identify some parts in the URI using a regex and give them a name so they can be easily retrieved from within the component In the above example the map says
183. ion may only need part of the XML data returned and the current facilities will load all the data into memory unnecessarily This last problem of processing XML efficiently is the focus of this paper An alternative to the XMLHttpRequest interface is proposed in the following section that provides the ability to stream XML to a receiving application This alternate interface provides a variety of outcomes for the receiving application that include filtering subsetting and binding the XML data as well as better interoperability with JOON as a programming data structure 2 Streaming XML Interfaces There are parts of the XMLHttpRequest interface that make sense to retain Specifically the formulation of the request has proven to be quite flexible and less problematic While the ability to send large data is still an issue that is being mitigated by some new proposals for extensions see 2 Meanwhile the average case is a simple and small request that can be responded to by a large volume of data in the response To address the need to process large incoming XML documents many systems resort to using callback e g SAX 6 or event oriented interfaces While a callback interface is tempting such interfaces are often very low level and may not fit well with the typical web developer In contrast web developers are used to receiving and processing event objects by writing and registering event handler functions As such the XMLHttpRequest i
184. ions but it does not place a requirement on the schema language that can be used to work with DITA As long as the class values get in the XML documents on processing then it does not case how that happened and what schema language is used So in order to have anything working the pre requisite was to have support for default values based on a Relax NG schema Then this support needs to be integrated in the XML parser processing workflow so we can get the default values when parsing an XML document On a separate thread the DITA DTDs needed to be converted to Relax NG schemas that use the annotation mechanism to specify the default values for all the class attributes Having the Relax NG default values support integrated at the XML Parser level made possible any processing to quickly work 168 DITA NG A Relax NG implementation of DITA and for example the DITA OT processor that converts DITA to a number of output formats works out of the box by just specifying a property on the XML parser 2 Default attribute values in Relax NG As mentioned in the introduction above the support for default attribute values based on Relax NG schema is essential in order to be able to use Relax NG for DITA The Relax NG DTD compatibility specification defines an attribute annotation a defaultValue that can be used to specify the default value where the prefix a is bound to the http relaxng org ns compatibility annotations 1 0 namespace For example a schema
185. ipt gt methods e Functional Limitations A pure JavaScript provides only basic functionality and most of constructions and methods need to be implemented from scratch Many JavaScript development frameworks e g jQuery Ext JS etc expand the standard functionality but this is not always sufficient to cover all purposes e Interoperability and User Control While DOM scripting is based on W3C DOM and ECMAScript standards different browsers implement the standard in slightly different ways In addition JavaScript code is run on client side where user resources can be limited and indeed a user can disable Javascript altogether The above restrictions limit and complicate the development of full fledged JavaScript applications In our development we used PHP to overcome platform limitations Ext JS framework and in a few cases designed distinct code for different Web browsers 3 Architecture The validator operates on an XML DOM instance in the browser Schema documents are not handled directly but rather an XML serialisation of an object model of the schema components assembled from the relevant schema document s is downloaded from a schema compilation server and the component model reconstructed There are two main reasons for this approach e Schema compilation that is the construction of a set of schema components corresponding to one or more schema documents is actually substantially more complex than validation itsel
186. is likely to outlive any particular code base a long long time At the same time one desires tools that make it easy to connect stuff suited to XML with stuff that s best suited to other formats The ideal architecture supports pipelining XML processing with HTML JSON RDBMS and all that without too much coupling to code Such requirements add up to tools that encourage working with XML in as de clarative a manner as possible operating at the level of data patterns pattern dis patch and data modeling It should be very natural to overlay semantic technology over XML whether in the form of RDF or in other formats at the higher level of se mantic annotations It s also important to start by getting the details right such as proper handling of mixed content and to keep perspective with powerful generic modeling techniques such as Schematron Abstract Patterns Over a decade of XML processing has demonstrated the difficulty of pleasing the desperate Perl Python Ruby hacker without shredding the rich information expression benefits of XML The upshot is the present interest in a refactoring of the XML stack and especially accommodation of XML to a world in which JSON is firmly established on grounds previously assumed for XML Akara aims at d tente applying traditional XML standards as much as reasonable but judiciously deferring to more natural Python idioms where needed to avoid frustrating developers Akara provides these benefits but
187. it can be optimized by the browser supplier The syntax of a conforming regular expression implementation could also be simulated If we take the missing dot all option this could be emulated by replacing all dots in the search string with s S 5 Evaluation 5 1 XQuery Standard Compliance For testing an XQuery implementation the W3C provides the XQTS XQuery TestSuite an extensive testing framework We used the version 1 0 2 for our tests because it was the newest one available when we started the project 5 1 1 Testing with the XQTS The XQTS 1 0 2 consists of 15133 tests covering by mandatory and optional features of XQuery 1 0 Each test is in its own file and there is one or more valid outputs for each test There are three main categories of tests standard parse errors runtime errors Standard tests have to return a result which can then be compared to the expected result The error tests expect a certain error code which indicates what triggered the error These tests are described in an XML file Because we wanted continuous testing we had to find a way to automate it GWT supports JUnit tests and our build server Hudson also offers good JUnit support For these reasons we converted the XML file to JUnit test cases using an XSLT stylesheet 205 XQuery in the Browser reloaded Another problem remained GWT runs tests in a windowless browser environ ment It is not possible to load any data from outside due to
188. ith the same name from amara import bindery MONTY XML lt quotes gt lt quote skit 1 gt This parrot is dead lt quote gt lt quote skit 2 gt What do you mean bleh lt quote gt lt quote skit 2 gt I don t like spam lt quote gt lt quote skit 3 gt But I was looking for argument lt quote gt lt guotes gt doc bindery parse MONTY XML ql doc guotes guote or doc guotes guote 0 D8 Akara Spicy Bean Fritters and XML Data Services print ql skit print gl xml attributes None u skit XPath works too b gl xml select u skit for q in doc quotes quote The loop will pick up both g elements print unicode q Just the child char data from itertools import groupby from operator import attrgetter skit_key attrgetter skit for skit quotegroup in groupby doc quotes quote skit key print skit unicode q for g in quotegroup 2 2 1 Creating a bindery document from scratch WHen creating a document from scratch the special nature of bindery specializes the process a bit involving the bindery entity base class from amara import bindery doc bindery nodes entity base doc xml_append doc xml element factory None u spam doc xml write lt xml version 1 0 encoding UTF 8 gt n lt spam gt The xml append fragment method is useful for accelerating the process a bit from amara import bindery doc bindery nodes entity base doc xml append fragment
189. json function in XSLT 3 0 and the ability to call Javascript and be called by Javascript 209 XQuery in the Browser reloaded e Truly asynchronous HTTP and the ability to access more browser state such as headers cookies or indexed storage e Further streamlining modularization and dynamic loading as well as the ability to centrally host the library Bibliography 1 Google Chrome s Need For Speed http blog chromium org 2008 09 google chromes need for speed_02 html 2 Google Web Toolkit http code google com webtoolkit 3 The New JavaScript Engine in Internet Explorer 9 http blogs msdn com b ie archive 2010 03 18 the new javascript engine in internet explorer 9 aspx 4 MXQuery XQuery Engine http www mxquery org 5 XQuery in the Browser Web site JavaScript Edition samples http www xqib org js 6 Zorba XQuery Processor https www zorba xquery com 7 G Fourny M Pilman D Florescu D Kossmann T Kraska D McBeath XQuery in the Browser Proceedings of the 18th International Conference on World Wide Web WWW 2009 Madrid Spain April 20 24 2009 2009 8 G Fourny M Pilman D Florescu D Kossmann T Kraska D McBeath XQuery in the Browser talk XML Prague 2010 March 12 13 2010 Czech Republic 9 A Gal M Franz Incremental Dynamic Code Generation with Trace Trees Technical Report No 06 16 Donald Bren School of Information and Computer Science University of California Irvine N
190. ke it much easier to extend the range of optimizations supported by XQilla in the future An example module of rewrite rules from XQilla is included in Appendix B 3 Types of Rewrite Clearly not all potential rewrites of XQuery expressions are correct that is produce the same results after the rewrite as before Of those that are correct even fewer are beneficial by any metric The rewrite rule language itself provides no checks on correctness or benefit Indeed correctness proofs are still the subject of much research with no general 216 Declarative XQuery Rewrites for Profit or Pleasure solution In practice the rule author will need to be responsible for the correctness of their rules Even so rewrite rules that change semantics and correctness can be useful under certain circumstances Similar to the use of operator overloading in C re write rules have potential uses in creating domain specific languages operating outside the bounds of specified XQuery semantics Consider the following rewrite AddComplex a b where rw subtype va element my complexNumber and rw subtype b element my complexNumber gt element my complexNumber lement my real a my real b my real lement my imaginary a my imaginary b my imaginary This rewrite rule allows two elements representing complex numbers to be added together using the regular operator an operation that sho
191. koning with the client where we see if our changes have broken SFM print origination 2 2 With much wisdom comes much sorrow Our initial appraisal of the task based on reviewing one book from each of four translations was that e Bibles contain books in a well known order e Books contain numbered chapters e Chapters contain paragraphs line groups and sections e Sections contain paragraphs line groups and sections e Paragraphs and lines contain verses e Individual verses are numbered consecutively e Verses contain inline markup such as highlighting and footnote references The reality after five translations is e Bibles contain introductions testaments glossaries and other notes External notes can be applicable to a chapter a verse or to part of the hierarchical structure Ecclesiastes 1 18 152 EPUB Chapter and Verse of the Bible that OSIS doesn t address such as The Pentateuch first five books of the Bible or The minor prophets Bibles differ in which books they contain and the hierarchy of the books The biggest difference is between Catholic and Protestant Bibles but there are differ ences within those types plus an ecumenical Bible that is equally far from both Protestant and Catholic hierarchy Chapters can start in the middle of a section and or in the middle of a paragraph can have more than one number and can be numbered alphabetically or numer ically within the same book Alpha
192. lassic MarkLogic benefits scale and speed rich full text support and enterprise database features The JSON format is wonderfully simple It contains objects a sequence of name value pairs arrays a sequence of values and values which can be a string 25 A JSON Facade on MarkLogic Server number object array true false or null A precise definition can be found at http json org 2 Design Considerations There were two main considerations when designing the MLJSON library 1 Approach things from a JSON angle Craft the XML to match the JSON not vice versa 2 Make good use of MarkLogic indexes Craft the XML so that it works well with MarkLogic indexes For example the names in name value pairs are represented as XML element names because that works well with MarkLogic s indexing of XML element structure Similarly JSON hierarchies are implemented as XML hierarchies to match MarkLogic s native abilities to handle hierarchical XML The overall goal of course has been to expose the power and features of MarkLogic but against JSON structures instead of XML documents 3 Conversion The MLJSON library includes two functions to convert JSON to XML and back again A user of the REST endpoints wouldn t normally call these functions but we ll look at them to understand the MLJSON underpinnings The jsonToXML function accepts a JSON string and returns an XML representation declare function json jsonToXMI
193. lation of SELECT here since it is the most common form The result of the SELECT form is a list of value combinations which are bound in varListSPA Table 3 Translation of SELECT query from SPARQL to XQuery resultSPA patternXQu nsListSPA nsListyoy SELECT varListspa let result patternyg WHERE order by orderListyo patternsp return result positionxo ORDER BY orderListsp limitOffset spa 3 6 Translation of the Running Example In Example 5 we show the result of the automatic translation of the SPAROL example to XOuery as introduced in Section 2 2 First the namespaces reguired for the guery are declared and the data of the involved collections is assigned to variables which are named according to GRAPH x with x E N These placeholders represent the different intermediate results The first variable GRAPH_ 0 contains the result of a basic graph pattern as de scribed in Section 3 1 A for loop is generated for each variable because the SPAROL semantics adds one result for each possible binding of values to variables We start with the bindings for the subject parts element In a next step we look at the predicate and object steps for each subject This can be done efficiently because RDF XML nests the predicates and objects into the subjects Since there are two variables for the objects name and col we create two additional loops For each possible binding the value of
194. le com coffee gt lt include href coffee machine gdefs rng gt 119 Configuring Network Devices with NETCONF and YANG lt start gt lt interleave gt lt element name cm configuration gt lt interleave gt lt element name cm ipv4 gt lt interleave gt lt element name cm address gt lt ref name ietf inet types ipv4 address gt lt element gt lt optional gt lt element name cm subnet mask length gt lt data type unsignedByte gt lt param name minInclusive gt 0 lt param gt lt param name maxInclusive gt 32 lt param gt lt data gt lt element gt lt optional gt lt interleave gt lt element gt lt oneOrMore gt lt element name cm ipv6 gt lt element name cm address gt lt ref name ietf inet types ipv6 address gt lt element gt lt optional gt lt element name cm subnet mask length gt lt data type unsignedByte gt lt param name minInclusive gt 0 lt param gt lt param name maxInclusive gt 128 lt param gt lt data gt lt element gt lt optional gt lt element gt lt oneOrMore gt lt interleave gt lt element gt lt optional gt lt element name cm state gt lt interleave gt lt optional gt lt element name cm supply levels gt lt interleave gt lt optional gt lt element name cm water gt lt ref name coffee machine state supply levels percent gt lt element gt
195. lement map gt lt dsrl maps gt 123 124 XSLT in the Browser Michael Kay Saxonica lt mike saxonica com gt Abstract This paper is in three parts past present and future The first part discusses why the success of XSLT in the browser has so far been underwhelming The second part discusses progress in porting Saxon to the browser The third and final part describes a vision for the future potential of client side XSLT 1 The Past XSLT 1 0 1 was published in November 1999 A year before the spec was finished Microsoft shipped an implementation as an add on to Internet Explorer 4 which became an integral part of Internet Explorer 5 In their eagerness they shipped a language usually now referred to as WD xsl which had many differences from the W3C specification that emerged the following year and although they soon followed it up with a conforming implementation the WD xsl dialect is still some times encountered to this day This illustrates nicely the paradox of the browser world things can sometimes move very guickly but they can also move very slowly indeed The speed at which Microsoft moved demonstrates that implementation in the browser was very much part of the thinking of the designers of the language The introduction to the XSLT 1 0 specification states that the language was not designed as a general purpose transformation language but as a language whose primary purpose is converting XML
196. lement represented by an array of XMLName instances The event also has a getAttributeValue method which will return an attribute value or undefined depending on whether the start element has the specified attribute This XMLItemEvent interface has been purposely flattened rather than creating a complex interface hierarchy of events This should promote a simpler view for scripting languages and lessen the overhead of the interfaces within the implement 139 Efficient XML Processing in Browsers ation The consequence is that some attributes or methods may not have meaning for certain kinds of XML events When that occurs the returned value is undefined Finally names in XML documents are represented by the XMLName interface The use or construction of instances of XMLName is limited to places where it cannot be avoided A user who knows the attribute for which they need a value can just pass a simple string containing the name of the attribute to the getAttributeValue method to retrieve the value Only when you need to enumerate the names of attrib utes or enumerate namespace declarations is this interface used interface XMLName readonly attribute DOMString prefix readonly attribute DOMString localName readonly attribute DOMString namespaceURI bi The attributes of the XMLName interface always return a string value for which the interpretation of an empty string is possibly context specific e An empty string for a prefi
197. lessandro Vernet Orbeon for the time he has spent discussing these issues with me and for suggesting to rely on query string parameters 1 Code Injection Wikipedia defines code injection as mailto vdv dyomedea com 177 XQuery Injection Code injection is the exploitation of a computer bug that is caused by processing invalid data Code injection can be used by an attacker to introduce or inject code into a computer program to change the course of execution The results of a code injection attack can be disastrous For instance code injection is used by some computer worms to propagate SOL injection is arguably the most common example of code injection since it can potentially affect any web application or website accessing a SQL database in cluding all the widespread AMP systems The second well known example of code injection is Cross Site Scripting which could be called HTML and JavaScript injection According to the Web Haccking Incident Database SOL injection is the number one attack method involved in 20 of the web attacks and Cross Site Scripting is number two with 13 suggesting that code injection technigues are involved in more than 1 out of 3 attacks on the web If it s difficult to find any mention of XQuery injection on the web it s probably because so few websites are powered by XML databases but also because of the false assumption that XOuery is a read only language and that its expression p
198. ll trails SOL implementations due to the simpler and more mature relational storage 1 4 Outline This paper is organized as follows Section 2 gives a short introduction to SPARQL XQuery and SQL establishes a running example and outlines the challenges of the translation A detailed description of the translation from SPARQL to XQuery is shown in Section 3 a summary of the translation of SQL to XQuery in Section 4 Section 5 describes the implementation of the translator and presents some initial correctness and performance results Section 6 presents related work The paper is concluded in Section 7 by a summary and directions for future work 82 Translating SPARQL and SQL to XQuery 2 Fundamentals 2 1 RDF The Resource Description Framework RDF is W3C Recommendation 12 for the specification of metadata models for the Semantic Web An RDF document expresses information about a resource on the web by a list of subject predicate object triples which correspond to a directed graph There are different formats for the serialization of RDF data The most common format is RDF XML 3 which stores RDF data with an XML syntax Other formats like Notation 3 N Triples and Turtle are more suitable for human readability All formats are equivalent semantically and can be converted into one another easily Example 1 shows a simple RDF document Example 1 RDF XML Example Periodic table and composite elements lt rdf RDF gt lt Elem
199. localName or localName gt namespaceURI pair DOMString getAttributeValue DOMString localName DOMString namespaceURI Returns an array possibly empty of XMLName instances where the b prefix namespaceURI are specified readonly attribute Array namespaceDeclarations Returns an array possibly empty of XMLName instances one for b each attribute readonly attribute Array attributeNames The XMLItemEvent interface extends the Event so that any instance can be processing and propagated within the browsers event handling system Each of the attributes of this event have the following definition itemType a code that indicates what kind of XML the event represents e g a start element prefix a prefix used on element name only available for start end element localName the local name of an element name only available for start end element or a target of a processing instruction namespaceURI the namespace URI of an element name only available for start end element uri the URI of the document only available for start end document value character content or the data of a processing instruction namespaceDeclarations any namespace declarations on a start element repres ented by an array of XMLName instances Each XMLName instance will use only the prefix and namespaceURI attributes attributeNames the names of all the attributes excluding namespace declara tions on a start e
200. ly without any loss when serializing back to JSON JSON support in XForms is also interesting for cross domain requests where security reasons force Javascript developers to use HTML src element instead of XMLHttpRequest An XForms implementation can hide this so there is no different way for a Web developer to treat cross domain requests 13 JSON for XForms 2 XForms Context 2 1 MVC Design XForms is based on elements for defining models controls events and actions X HTML is the favorite host language for XForms but this could also be SVG for example Within a model there are instances of XML data which will be used for building controls and outputs and some of them will be edited Bindings allow to add rich constraints to data such as types calculations conditional read only or relevant status Controls are referencing nodes to output values and modify them according to constraints Events are dispatched so actions are locally performed 2 2 XPath in XForms XForms Recommendation defines extra specific XPath functions for accessing dif ferent instances for encrypting values for context retrieving XPath expressions are evaluated according to a current context as it is done in XSLT stylesheets for example The default context for an instance is at the document element so XPath expressions in XForms usually don t need to start with my_document_element and in most forms the document element name is in fact
201. management systems for personal note taking and so on anywhere consistent precise and clear document structure and design are crucial The WYSIWYG environment is more accessible to users without technical knowledge providing easy tools to create and edit complex yet still valid XML documents The development of such such applications requires powerful editing tools which enforce continuous compliance with XML Schema based lan guage definitions and good performance to ensure high efficiency and usability rates Client side XML Schema validation Web 2 0 promotes the principle of software as a service which means that no installation is needed and a user can work with an application from any Web browser regardless of deployment platform These capabilities widen the usability of applic ations making it possible to use the program on Windows Linux Mac OS mobile phones laptops and any devices with Internet access During the past few years Web 2 0 has become a popular technique that fre quently dictates the demands for modern Internet business However client side XML validators mainly remain out of reach The principal obstacles are the com plexity of development of effective continuous XML Schema validation and style sheet generation algorithms considerable limitations of client side platforms and complex testing requirements There are a number of visual WYSIWYG editors available on the market Most of them are stand alone
202. markup producing something like XHTML but without the namespace 3 Generating XML and HTML Amara supports the traditional well known SAX like approach to generating XML output startElement output text output endElement But this is generally awkward and unfriendly e g the code block structure does not reflect the XML output structure so it can be really hard to debug when you trip up the order of output constructs so in this tutorial we ll focus on structwriter a rather more natural approach The struct in this case is a specialized data structure 60 Akara Spicy Bean Fritters and XML Data Services that translates readily to XML For now just the one example which does cover most of the key bits import sys datetime from amara writers struct import from amara namespaces import tags u xml u python u atom w structwriter indent u yes w feed ROOT E ATOM NAMESPACE u feed XML NAMESPACE u xml lang u en E u id u urn bogus myfeed E u title u MyFeed E u updated datetime datetime now strftime SY Sm SdTSH M 3SZ E u author E u name u Uche Ogbuji E u uri u http uche ogbuji net E u email u uche ogbuji net E E u link u href u blog E u link u href u blog atom1 0 u rel u self E E u id u urn bogus myfeed entryl E u title u Hello world E u updated gt datetime datetime no
203. maxOccurs unbounded gt lt exml anonymous exsi maxOccurs unbounded gt lt exml anonymous exsi maxOccurs unbounded xsi nil true gt lt exml anonymous gt lt exml anonymous gt lt exml anonymous exsi maxOccurs unbounded gt lt ce gt stringCl lt c gt lt d gt stringD1 lt d gt lt exml anonymous gt lt exml anonymous exsi maxOccurs unbounded gt lt e gt stringC2 lt c gt lt d gt stringD2 lt d gt lt exml anonymous gt lt exml anonymous gt lt exml anonym and for default context e 1 1 equals stringA e 1 2 equals 42 e is array 2 equals true e is array 3 1 equals true e c d stringD1 equals stringCl e c stringC2 d equals stringD2 3 2 7 XPath Engine Proposed Enhancements When possible XPath engine modifications can simplify expressions for JSON ob jects e asin MySQL non XML names in expressions could be quoted with character backquote to avoid predicates with fullname calls e name and local name functions could be extended to include fullname and local fullname functions support and return _ just when true e name and local name functions could be modified to return empty string instead of respectively exml anonymous and anonymous e used for exml anonymous could be simplified as just e used for exml anonymous could be written 20 JSON for XForms e u
204. mic values or XML nodes Support for graph structures is limited as there is no standard way to define links across the tree hierarchy and no expressive operations on these links exist 2 4 SOL SOL is the most popular language as an interface to a relational database manage ment system DBMS and has been extended to suit many additional purposes It provides expressions for data definition DDL data manipulation DML access privileges DCL and transaction control TCL Given the complexity of the lan guage we only consider the SQL92 DML subset in this work 2 5 Challenges and Opportunities At the level of the data model the differences between the relational tree sequence and graph models are already attenuated by the serializations in particular the RDF XML mapping Similarly type system differences are resolved by SQL XML mapping which is described in 18 and the shared XML Schema XPath 2 0 type system SPAROL XOuery Both SQL and SPARQL use three valued Boolean logic explicitly addressing null values and errors respectively In contrast to this XQuery uses two valued Boolean logic does not represent null values and will only provide error handling in the next version 3 0 The process of graph pattern matching in SPAROL is quite different from the path navigation style interaction For this reason emulation is required which is less concise and possibly less efficient 3 Mapping and Translating SPAROL to XQuery In
205. mpact syntax However it does not maintain the same folder structure So converting from Relax NG XML to Relax NG compact needs to be done though a script that invokes the conversion on each main schema then moves all the files in the correct folder structure removing duplicates and then updates all the schema references to match the folder structure This was implemented in an Ant build script 3 3 Sample schemas comparison with DTDs The DITA specification defines a base module and then a few standard specializa tions The standard specializations however are in no way different from user spe cific specializations Thus if we look at a standard specialization we can see also how a user specific specialization will look like and we can compare the DTD based specialization with the Relax NG one Let s look at some of the parts that define the Relax NG schemas and compare them with how the same things are defined in the DTDs 3 3 1 The domains attribute This is again a default attribute that appears on the root element and specifies what domains are included in that schema In the case of DTDs it is defined like this lt ENTITY included domains amp concept att amp hi d att 6ut d att amp indexing d att Shazard d att 6abbrev d att amp pr d att amp sw d att amp ui d att 171 DITA NG A Relax NG implementation of DITA with careful definitions for each of these entities There is no automated check to test
206. mplemented entirely in XSLT 1 0 Again despite the obvious intrinsic inefficiencies its interactive per formance is quite sufficient to meet the requirements of most applications Let s look at what was needed to implement XSLT 2 0 using Saxon 7 in the browser I m calling it Saxon Client Edition or Saxon CE using the GWT cross compiler 1 The first step was to get rid of all unnecessary code The source code of Saxon EE 9 3 is around 400K lines of Java Although the generated Javascript will be cached the time taken to view the first screen on a web site is critical and I de cided to aim to get this down to around 40K Comments don t matter and the first decision was to start with the functionality of Saxon Home Edition which immediately brings it down to 143K lines of non comment code At the time of writing I have reduced it to 60K lines which produces a little under 900Kb of highly compressed Javascript The target of 40Kb looks quite feasible Saxon 5 0 which was the first full implementation of XSLT 1 0 delivered in December 1999 was just 17K lines and XSLT 2 0 is probably only twice the size of XSLT 1 0 as a language Looking at some sample code the implementation of the substring function it s fairly clear that one can cut it down from 200 lines of code to 20 by removing a few optimizations of marginal value diagnostics and code for special situations such as streaming Getting the code down to 80K was fairly easy by cut
207. n index of works or poets works for work id name poet in POETDB u work iteritems works append getters entitytype work id return n join works try return getters entitytype eid except KeyError return not_found baseuri Again focusing on the key new bits from amara lib iri import join Amara comes with a lot of URI and more generally IRI internationalized URI functions which are more RFC compliant than the urllib equivalents including the join function which constructs URI references from hierarchical path components from akara import request respons The response object allows you to manage HTTP request status headers and such def not found baseuri ruri reguest uri reguest environ response code 404 Not Found return u lt p gt Unable to find s try the lt a href s gt index of works lt a gt lt gt p gt S ruri baseuri Just a little utility function to provide a 404 response with some information useful to the end user request uri is a Python stdlib function to reconstruct the request URI from a WSGI environment baseuri request environ SCRIPT NAME Here you construct the URL to access this service def get work wid uri join baseuri work wid 76 Akara Spicy Bean Fritters and XML Data Services name poet POETDB u work wid puri join baseuri poet poet return lt p gt Poetic work lt a href s gt s lt a
208. nd it difficult to develop and maintain two rather different configuration systems and eventually concentrated on their propri etary CLI as the up to date and authoritative configuration vehicle Consequently 99 Configuring Network Devices with NETCONF and YANG network operators were left with no choice other than to develop various kludges for sending configuration scripts generated from a database through SSH to the command line interfaces of their devices Such methods however ingenious they might be are necessarily fragile and unable to react properly on various errors that the scripts may trigger The NETCONF working group of the Internet Engineering Task Force IETF was chartered in May 2003 with the aim of standardizing an open and vendor neutral protocol for network device configuration This paper first describes essential features of the NETCONF protocol using several examples and then deals with the recent IETF efforts in the data modelling area namely the YANG language for modelling configuration and operational state data manipulated by the NETCONF protocol Finally the standardized mapping of YANG data models to Document Schema Definition Languages DSDL is dis cussed in some detail As a side effect this will allow us to compare the main features of the YANG language with analogical features of the existing XML schema lan guages 2 NETCONE Configuration Protocol NETCONF is an XML based client server protocol d
209. nd query execution although it could also happen on demand during query execution if the implementation wants to support just in time optimization Rewrite rules continue to be applied to the query or expression until no rewrite rule matches Importantly rewrite rules are applied to the results of other rewrite rules utilizing co dependent rules often allows simpler and more generic rewrite rules to be written This power also implies the possibility of unterminated rewrite loops so care must be taken to avoid this situation when creating the rewrite rules During the query rewriting phase function and variable inlining are handled automatically Similarly if a sub expression is determined not to depend in any way the dynamic context for execution that sub expression is replaced with its result 1 Where the rw prefix is bound to the namespace URI http xgilla sourceforge net rewrite 214 Declarative XQuery Rewrites for Profit or Pleasure as a literal or constant folded These optimizations are interleaved with rewrite rule application meaning that rule authors can realistically rely on them being performed on the result of their rule application Any other optimizations that the implement ation performs could also be performed during the query rewriting phase although there might be good reasons for not doing so 2 6 Using Rewrite Rules in XQilla Asa test of the power of the rewrite rules language it has been implemented
210. nder much of the benefit of REST In many Web applications rather than calculating a greeting on the fly we re instead gathering information and even modifying some well known referenceable resource In such cases the common convention is to use hierarchical URLs to rep resent the different resources As an example say we re developing a database of poets and their works Each poet would be a distinct resource e g at http localhost 8880 poetdb poet ep To get this somewhat more sophisticated behavior we take advantage of the common WSGI convention of Python The following complete Akara module im plements the poet database from wsgiref util import shift path info from akara services import simple service Tj from akara import request POETDB SERVICE ID http example org my services poetdb Cheap DBMS POBTDB u poet u ep u Ezra Pound u 45 Usura Place Hailey ID u co u Christopher Okigbo u 7 Heaven s Gate Idoto Anambra Nigeria u work u cantos u The Cantos u poet ep u mauberley u Hugh Selwyn Mauberley u poet ep u thunderpaths u Paths of Thunder u poet co simple service GET POETDB SERVICE ID poetdb text html def poetdb entitytype shift path info reguest environ 73 Akara Spicy Bean Fritters and XML Data Services eid shift path info request environ info POERTDB entitytype eid if entity
211. ng is described and its features compared to those offered by the existing XML schema languages Finally the standardized mapping of YANG data models to the DSDL schema languages RELAX NG Schematron and DSRL is discussed in some detail and the procedure for in stance document validation is outlined Keywords network configuration data modelling NETCONF NET MOD DSDL RELAX NG Schematron DSRL 1 Introduction Configuration of network devices and especially those that are part of the Internet infrastructure often requires considerable expertise and effort Manual configuration via a web or command line interface CLI may be a good option for a home WiFi router but large installations of routers in backbone networks call for robust and automated configuration approaches Historically SNMP Simple Network Management Protocol was designed to cover the entire range of network management tasks including remote configuration While SNMP has been widely implemented and deployed nowadays it is almost exclusively used for monitoring purposes such as gathering statistics from devices but rarely for configuration A number of reasons were identified to have contributed to the failure of SNMP in the configuration area 15 Perhaps the most important of them was the mismatch between the high level human oriented syntax of com mand line interfaces and the low level data oriented paradigm of SNMP As a result most device vendors eventually fou
212. ning of lines preceded by a They only occur at the start of lines except for some translations where they don t Here s some example SFM markup from the start of Genesis Vid GEN SER01 SB4 Gen se ES 21 06 2010 h GEN ESE mt Gen se s Dieu cr e l univers et l humanit eam http www la bible net 150 EPUB Chapter and Verse p v 1 Au commencement Dieu cr a le ciel et la terre We were provided with Balises indicating which markup codes were used in a particular translation but these lists were soon shown to be incomplete Even when the codes on the Balises were used they were used in different combinations one translator even found three different ways within one Bible to markup tables SFM books include markup for the major and minor parts of the book s title except even the most common version of those codes can be used in either order SFM books may also include an optional introduction where the end of the intro duction is indicated by a ine code except some translations have introductions without an ine The translations also included inline markup for indicating bold or italic emphasis or superscript text The inline markup was usually followed by a single letter code except when it wasn t and a single end code usually x closed all open highlights except when it didn t Inline markup didn t cross the SFM codes at
213. nput and deliver output within this environment In the first place this means being able to use both the HTML DOM and the XML DOM as input which is done using wrappers much as the Java and NET product support different DOM flavours Similarly it means being write a result tree in a form that can be attached to either an XML or HTML DOM It means providing an invocation API at the moment the only thing I have implemented is invocation using the HTML lt script gt element but some kind of Javascript API is likely to supplement this And it is necessary to implement the document function in terms of XmlHttpRequest calls to the server All this is working and it already provides as much capability as most of the browser XSLT 1 0 implementations But I want to go beyond this and that s the topic of the next section Experience with GWT has been very positive indeed I have not encountered any cases where the Javascript it generates produces different results from the native Java or where the generated code even needs to be examined The compilation process translating Java to Javascript is slow but the time is well spent because it results in superbly fast Javascript code Unfortunately it s not possible to invoke a transformation by simply sending the XML document to the browser with an lt xml stylesheet gt processing instruc 129 XSLT in the Browser tion at the start The interpretation of this is under the control of t
214. ns l Metadata and documentation strings are considered an important component of YANG data models so they are also carried over to the hybrid schema Dublin 108 Configuring Network Devices with NETCONF and YANG Core elements are used for metadata and the documentation element of the RELAX NG DTD Compatibility annotations 14 is used for documentation The hybrid schema is an intermediate product of the mapping and cannot be directly used for any validation As a matter of fact it is not even a valid RELAX NG schema because it contains multiple schemas demarcated by special elements belonging also to the above namespace On the other hand a person familiar with RELAX NG should be able to get a relatively precise idea about the data model from reading the hybrid schema The second step of the mapping is specific for every target object datastore or NETCONF message Its task is to extract relevant parts from the hybrid schema and transform them into three DSDL schemas RELAX NG schema is for the most part simply assembled from appropriate fragments of the hybrid schema e Schematron schema is generated from semantic annotations e DSRL schema is generated from annotations specifying default contents Appendix A shows the hybrid schema for our coffee machine data model as well as the three validation schemas obtained in the second mapping step 4 2 Mapping Grammatical Constraints and Datatypes YANG statements defining da
215. nstance id isearch mediatype application json gt action opensearch format json search lt xf instance gt lt X f instance id iresults mediatype application json gt lt xf instance gt lt xf bind nodeset search constraint instance iresults 2 p upper case upper case current gt lt xf submission method get replace instance instance iresults gt separator s amp validate false action http en wikipedia org w api php gt lt xf setfocus ev event xforms ready control search gt lt xf model gt lt sty le type text css gt search label float left width 4em results border lpx solid black width 15em margin left 4em 23 JSON for XForms results xforms value hover background color 418ad5 lt style gt lt head gt lt body gt lt hl gt WIKIPEDIA OpenSearch Test Form lt h1 gt lt p gt Please enter a subject in the following field The value is not case b sensitive but it has to exist in the results of the corresponding search lt p gt lt xf input id search ref search incremental true delay 500 gt lt xf label gt Subject lt xf label gt lt xf send ev event xforms value changed gt lt xf toggle ev event DOMFocusIn case show autocompletion gt lt xf input gt lt xf switch gt lt xf case id show autocompletion gt lt xf repeat id results nodeset instance iresults 2 p
216. nt A gt A C B where rw 1s constant B declare option rw rule rw AddSubTransFold A B C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double gt A C B where rw is constant A gt A C B where rw 1s constant B declare option rw rule rw SubAddTransFold A B C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double gt A C B where rw is constant A gt A B C where rw is constant B declare option rw rule rw SubSubTransFold A B C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double gt A C B where rw is constant A gt A B C where rw is constant B Only for xs decimal since otherwise 0 messes things up declare option rw rule rw MulZeroFold e 0 gt 0 where rw subtype e xs decimal declare option rw rule rw MulOneFold e 1 gt e where rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double declare option rw rule rw DivOneFold e div 1 gt e where rw subtype e xs decimal or rw subtype this xs float or rw subtype this xs double declare option rw rule rw AddZeroFold e 0
217. ntent model or the re strictions on redefinitions Thus for content oriented documents the choices are either Relax NG or to stay with good old DTDs The latest version of content oriented vocabularies like DocBook and TEI made the move to Relax NG This however did not happened for DITA which stays with DTDs DITA provides also an XML Schema implementation but only few people use that most probably because of the extension 167 DITA NG A Relax NG implementation of DITA limitations the majority use DTDs Now an evident guestions is Why does not DITA use Relax NG like all the others Io understand the issues that prevented the use of Relax NG for DITA we need to see what stays at the core of DITA functionality The main promise of DITA is that it facilitates easy interchange of information between not necessarily related entities That is I can have my own DITA specializations my own specific elements and if I pass this to someone else that also uses DITA then although they do not have a specific support for my elements that are still able to process my documents That is possible because the processing is not done at element name level but by looking at a hierarchical value specified in a class attribute that is present on any element and specifies how that element is derived from the standard DITA classes The value is hierarchical as it contains values starting from the most generic type to the actual type This is similar as concep
218. nterface has been modified by removing all the whole document response attributes and by adding an event handler for XML events This new interface is defined as follows interface XMLReader event handler attributes attribute EventListener onreadystatechange attribute EventListener onabort attribute EventListener onerror attribute EventListener ontimeout attribute EventListener onxml 136 Efficient XML Processing in Browsers state const unsigned short UNSENT 0 const unsigned short OPENED 1 const unsigned short HEADERS RECEIVED 2 const unsigned short LOADING 3 const unsigned short DONE 4 readonly attribute unsigned short readyState void open in DOMString method in DOMString url raises DOMException void open in DOMString method in DOMString url in boolean async raises DOMException void open in DOMString method in DOMString url in boolean async in DOMString user raises DOMException void open in DOMString method in DOMString url in boolean async in DOMString user in DOMString password raises DOMException void setRequestHeader in DOMString header in DOMString value raises DOMException void send raises DOMException void send in Document data raises DOMException void send AllowAny DOMString data raises DOMException void abort boolean parse in DOMSt
219. o be used as an argument to general functions Example 3 Variable Binding Sequence lt result gt lt var name color gt silvery lt var gt lt var name name gt aluminium lt var gt lt result gt lt result gt lt var name color gt metallic white lt var gt lt var name name gt uranium lt var gt lt result gt A basic graph pattern contains a set of triple patterns which all need to hold and can be joined correlated by using the same variable In our translation one triple can yield up to three nested for iterations since one loop is generated for each dif ferent variable Given the subject predicate object nesting in RDF XML we start by retrieving subjects using a path expression and bind the variables specified on subjects Nested into an iteration over all subjects we retrieve the predicates below them bind the variables and again iterate over these variables if necessary Objects are handled ina similar nested fashion In general constants correlations and other direct filter conditions are expressed as part of the where clause since all possible combinations are generated by the loops Whenever possible we push these predic ates down in order to minimize the size of the intermediate result Wherever ne cessary access to named graphs as external resources is mapped to doc or collection calls 85 Translating SPARQL and SQL to XQuery Table 1 Translation of basic pattern from SPARQL to XQu
220. o the XQuery expressions can leak out of the literal strings where they are supposed to be located 1 http demo exist db org exist functions xmldb get child collections 182 XQuery Injection 3 1 Generic How To The most common way to block this kind of attacks is to escape the dangerous characters or sanitize user inputs before sending them to the XQuery engine In an XQuery string literal the dangerous characters are 1 The s that can be used to make references to predefined or digital entities and needs to be replaced by the samp 2 The quote either simple or double that you use to delimit the literal that needs to be replaced by sapos or amp quot And that s all These two replacements are enough to block code injections through string literals Of course you also need to use a function such as encode for uri so that the URL remains valid and to block injections through URL encoding The second way to block these attacks is to keep the values that are entered through web forms out of the query itself When using eXist this can be done by encoding these values and sending them as URL query parameters These parameters can then be retrieved using the re guest get parameter extension function Which of these methods should we use There is no general rules and it s also a matter of taste That being said Sanitizing is more portable reguest get parameter is an eXist specific function that cannot be u
221. oc xml uri inputsource uri 0 resolver restricted resolver doc amara parse uri ecount doc xml select u count return str ecount 7 2 Hello World The following Akara module implements a simple Hello world service from akara services import simple service HELLO SERVICE ID http example org my services hello simple service GET HELLO SERVICE ID hello def helloworld friend None return u Hello friend decode utf 8 Returns a Unicode object Save this as hello py and make it available in PYTHONPATH and update the akara conf of an Akara instance to load the module If the instance is at localhost 8880 you can invoke the new module as follows curl http loca Hello Uche Or if you prefer put hi lhost 8880 hello friend Uche ttp localhost 8880 hello friend Uche into your browser to get the nice greeting Go ahead and play around with URL basics e g http docs python org tutorial modules html the module search path 72 Akara Spicy Bean Fritters and XML Data Services curl http localhost 8880 hello friend UchetOgbuji Hello Uche Ogbuji Which in this case behaves just like http localhost 8880 hello friend Uche 200gbuji 8 Introducing WSGI and working with URL path hierarchy The above approach works fine if you are creating very simple dynamic query services but it gets very tempting to do too much of that and to squa
222. od does not conform to our goal because it exploits the schema information too late For our target user population and functionality we need an approach which preserves validity at all times by constraining what updates are possible This point is crucial for WYSIWYG editing since the consequences of restyling an invalid document are unpredictable at best We called our method restriction validation meaning that the user could perform only allowed actions and the document is always valid Before editing the document the user is offered a selection of possible validity preserving actions over the selected component 4 1 Update Operations To talk about the algorithm first we need to define the update language As we have already said each action must conform to the controlling schema We cover the following update actions e Insert After where X is a selected element and Y is an element or text results in inserting Y immediately after X This operation is not defined for the root node e Insert Before where X is a selected element and Y is an element or text results in inserting Y immediately before the X This operation is not defined for the root node Insert Into 1 where X is a selected element of complex type and element Y is an element or text results in inserting Y into X in a place that conforms to W3C XML Schema If X already has Y then insert immediately after the last Y 2 where X is a selected element and Y i
223. of thumb here again is never inject a user input in an XQuery without sanitizing it or moving it out of the query using request parameters When using an implementation such as Orbeon Forms that supports attribute value templates in resource attributes it may be tempting to write submissions such as Example 7 Unsafe XForms Submission lt xforms submission id doSearch method get resource http localhost 8080 orbeon exist rest db app users query gt user mail instance search instance result replace instance gt Unfortunately this would be exactly similar to the unsafe Java realm that we ve used as our first example 185 XQuery Injection To secure this submission we can just adapt one of the two methods used to secure XSLT accesses This is especially straightforward with the Orbeon implement ation that implements an xxforms variable extension very similar to XSLT variables You can also go with FLOWR expressions or use xforms bind calculate definitions to store intermediate results and make them more readable but it is also possible to write a mega XPath 2 0 expression such as this one Example 8 XForms Submission Sanitized lt xforms submission id doSearch method get resource http localhost 8080 orbeon exist rest db app users gt _query encode for uri concat user mail lower case normalize space replace replace instance search gt Samp S amp amp amp
224. on times varied quite significantly for different queries and scale factors as shown in Figures 1 8 We are omitting a number of resuls in graphs which are available on the XOBench Benchmarking Service 7 e We were not able to gather any result from MonetDB After rewriting the queries to conform to mandatory static typing we ran into non implemented functions or accessors While a full rewrite for this functionality might have been possible the effort would have been significant e XQilla performed very similar to Zorba sometimes slightly worse we therefore omitted the results to make the graphs more readable e BerkeleyDB and eXist performed similar or slower than the other XML databases We again omitted the graphs to improve readability For a simple triple match query Figure 1 all XQuery engines outperform ARQ significantly and scale better indicating that the translation of triple and basic graph pattern matching is quite efficient For the XQuery processors the total cost is completely dominated by XML parsing thus making Zorba slower than Saxon The XQuery databases do not need to pay the price for parsing and only need to deal with a bigger data size As a result they maintain a lead of 3 orders of magnitude over ARQ even without the support of manually added indexes When comparing 91 Translating SPARQL and SQL to XQuery the semantically same query translated from SQL and comparing the results against MySQL the
225. one application or as an Eclipse IDE plugin on Windows 7 Vista XP 2000 Mac OS X Linux and Solaris Syncro Soft Srl All rights reserved lt oXygen gt is a registered trademark of Syncro Soft in U S and other countries Other product names may be trademarks or registered trademarks of their respective owners Table of Contents G n ral INLOFI AFO Selene tas diode ddd alse hte dla i ta tet OKA k Ea Sie oOo ae ix DPONSOTS ao KA ot E A o S DO eae ations xi I BA E e Cs 23x esas siki za ne EEE wars rennet meas dua AA eee oa Zda oe ane aaa hi ela xiii Client side XML Schema validation Aleksejs Goremikins and Henry S Thompson ssessssesssessseesssesesessssesessrsesesessnsenesessnsenensssese 1 JSON for XForms Alain Couthures sccccisccciseseccrssvecsisseiccssosccavsvsaassndecsssnssassoteccssscesostecs 13 A JSON Facade on MarkLogic Server Jason Hunter and Ryan Grimm 25 CXAN a case study for Servlex an XML web framework Florent Georges 35 Akara Spicy Bean Fritters and XML Data Services Uche Ogbuj e 53 Translating SPAROL and SQL to XQuery Peter M Fischer Dana Florescu Martin Kaufmann and Donald Kossmann 0 81 Configuring Network Devices with NETCONF and YANG TEA SLO TY ooo niae E ARE A A E A nota za budnd tues te 99 XSLT in the Browser Michael Kay sb ce aucesisterstaatagmimeos aweonenveGencanegriserdvonnt sacs 125 Efficient XML Processing in Browsers
226. onent that adds default attribute values and tried to edit a DITA based on Relax NG document in oXygen As expected everything works the same level of support that is available for DITA based on DTD or XML Schema is present with nothing missing including being able to correctly render the documents though CSS in the Author editing mode 173 DITA NG A Relax NG implementation of DITA om lt sampleConcept xml Users george Documents workspace dita ng samples sampleConcept xml lt oXygen gt XML Editor Dee eB fa Q ey gt E h b gt ant test Saxon EE XPath 2 0 concept conbody p ol li B a lt 840 amp Oy o E s Fl ns ss E We ih is H i i i G DB i VY B GIG CGS iBZUa TS MOH EE HBSS F ooo DITA Maps Manager 5ampieConceptxmi x 4 bw OOO Attributes IC 7 A E u 4 B B ty concept conbody p ol li azza x E Attribute Value i 1b Sample audience E DITA Topic Map Sample E ETSI 1 item aa z 2 item 2 er conkeyre baseTopic xmli test x3 D 4 test ee e first conrefend e second dir DITA works with Relax NG id Vv colspec colname cl colnum 1 colwidth 1 0 importance colspec colname c2 colnum 2 colwidth 7 7 otherprops rea B Project BE oram 4 Data So x I Test platform 000 Outline 2 Test 2 product A Element t 3 Test 3 props A rev v ol iteml Super i status li item 1 Almost incre
227. onverted into boolean logic XQuery itself is used as the predicate language for maximum familiarity and in order to avoid designing an additional predicate language The matched expres sions are exposed to XQuery as a new atomic type named rw expression returned using the expression reference syntax already seen Expression items can be queried by a library of built in functions including the rw subtype and rw never subtype functions used to match the inferred type of the expression against a SeguenceType the rw is constant function used to check that the expression does not depend in any way on the dynamic context and the rw uses focus function which checks if the expression depends on the focus context item position or size 2 4 Multiple Cases and Case Predicates It s also possible to include multiple result cases in rewrite rules Each case can have its own predicate and cases are examined in order until an applicable case is found fn FnEmptyFold empty e gt false where rw subtype e item gt true where rw subtype e empty seguence This example uses the inferred type of the argument to the empty function to avoid execution of the expression where possible A complete grammar for the rewrite rule notation is available in Appendix A 2 5 Rewriting phase Rewrite rules are applied to a query or expression during the rewriting phase This typically happens between static analysis a
228. orba 1 BaseX gt E EE e Sedna 0 1 T T T T T 1 10 50 100 500 1000 5000 scaling factor Figure 8 BERLIN SQL Query 8 As a conclusion the results indicate that our translation is working quite well for most cases outperforming the SPAROL reference implementation for most of the queries and scaling similarly or better than the SQL results albeit at a higher initial cost Certain join queries still seem to be problems in both translations necessitating further investigation and more translation hints 6 Related Work A significant number of translations from SPARQL to SQL have been performed which achieve both completeness and efficiency as summarized in 15 Even though SPARQL and XQuery are significantly closer than SPAROL and SQL we are only aware of two works that have tackled such a translation In 4 the authors aim to query XML sources from SPARQL without having to use XQuery explicitly In order to do so they require the presence of XML schema which is mapped to an equivalent OWL schema for RDF SPARQL queries written against this OWL schema are then translated into XQuery expressions on the corresponding XML schema The trans lation is incomplete and based on the publicly available information it is not possible to verify its correctness 8 embeds SPAROL into XQuery and provides a translation to XQuery without assuming schema knowledge Again this translation is incom plete and their evaluation shows signific
229. orithm see above and in addition for the first child Insert Before algorithm if Insert After allows element from the candidates we can insert it immediately after the child or before for the first child the accepted Client side XML Schema validation elements and positions are saved for further insertions If there are many insertion positions take one which has the lowest in the tree insertion position The Insert Into 2 action firstly finds all possible attributes for the selected ele ment Secondly checks whether the attribute already exists in the element Lastly if an allowed attribute does not exist one can perform the insert action Since W3C XML Schema attribute does not have bounded occurrence ranges it either exists or not we cannot insert any of the existent attributes If the attribute has default or fixed fields then the newly inserted attribute has the value of these fields other wise the value is an empty string or other type determined minimal allowed value The Delete action for elements requires the number of occurrences NoO of se lected element in the document and supposed number of occurrences in W3C XML Schema that is minOccurs field If the NoO gt minOccurs we can delete the element In case of attributes we need to check whether the attribute use is optional can delete required cannot delete The Edit action is trivial we need to get the text type or simple type from the schema and check
230. ort for XSLT in many browsers it suffers from several drawbacks making it unsuitable as a general complete solution for Web client developement The main problem is that current browsers only support XSLT 1 0 often even in incompatible dialects Since XSLT 1 0 is more than 10 years old a lot of important functionality available in XSLT 2 0 is missing In addition XSLT in the browser runs completely independently from the Web site it belongs to An XSL Transformation just receives an XML node and outputs an HTML XML document or text depending on the output type set but does not in teract with the Web site 194 XQuery in the Browser reloaded Example 4 An XSL transformation lt xsl stylesheet xmlns xsl http www w3 org 1999 XSL Transform version 1 0 gt lt xsl template match gt lt xsl copy of select b text gt lt xsl template gt lt xsl stylesheet gt In the browser an XML Transformation can be launched from a JavaScript program Example 5 A JavaScript sample which will execute an XSL transformation xsl holds a stylesheet document xml holds the XML document xsltProcessor new XSLTProcessor xsltProcessor importStylesheet xsl resultDocument xsltProcessor transformToDocument xml Another variant to use XSLT in the browser is by including lt xml stylesheet href stylesheet xsl type text xsl gt at the beginning of the page then the page itself is being used
231. ovember 2006 10 M Kay Compiling Saxon using GWT http saxonica blogharbor com blog archives 2010 11 16 4681337 html 210 Declarative XQuery Rewrites for Profit or Pleasure An optimization meta language for implementers and users alike John Snelson MarkLogic Corporation lt john snelson marklogic com gt Abstract One of the big challenges for any emerging database product is the maturity of its query optimizer This is even more of a problem with XQuery 1 which unlike SQL hasn t yet had the benefit of forty years of optimization research Any efforts to advance the state of the art in optimizing XQuery are therefore important as steps towards fulfilling its promise as a new database paradigm This paper introduces a novel meta language for efficiently specifying rewrites to the expression tree of an XQuery program The applications of this language are wide ranging including use by XQuery implementers to efficiently communicate and execute optimizations use by XQuery library writers to extend the optimization semantics of the host implementation with a deep understanding of the library functionality and use by XQuery end users to provide optimization hints and greater insight into the program written or data operated on This paper also discusses the use of this language to replace and extend the op timization layer in XQilla an open source implementation of XQuery 1 Introduction 1 1 The Optimization Problem XQue
232. ow possible We implemented the algorithms for controlling an XML editing process in compliance with W3C XML Schema The application provides an opportunity to detect the possible elements attributes to insert or delete as well as physically add or delete items from an XML DOM tree and edit text The developed framework could be easily extended as we have used modular development methodology popular Ext JS libraries and well considered interfaces We resolved the problem 11 Client side XML Schema validation of detection of selected elements within FSM which implies linking XML DOM XML Schema FSM and graphical tree representation In future work we plan to supplement the validation engine to support all W3C XML Schema constructions explore the valid subtree insertion problem and add support of a wide range XSL and or CSS transformations 7 References 1 Adobe Systems Inc The Mozilla Foundation Opera Software et al Ecmascript 4th edition language overview October 7 2007 2 D Barbosa A Mendelzon L Libkin L Mignet and M Arenas Efficient incremental validation of xml documents In Proceedings of the 20th International Conference on Data Engineering ICDE04 pages 671 683 Boston Massachusetts USA 2004 3 B Bouchou M Halfeld and F Alves Updates and incremental validation of xml documents In 9th International Workshop on Database Programming Languages pages 216 232 Potsdam Germany September 200
233. ower if limited and that the conseguences of XOuery injection attacks would remain limited This assumption must be revised now that XML databases start implementing XOuery Update Facilities and have extensive extension function libraries which let them communicate with the external world 2 Example of XOuery Injection 2 1 Scenario If you develop an application that requires user interaction you will probably need sooner or later some kind of user authentication and if your application is powered by an XML database you may want to store user information in this database In the Java world Tomcat comes with a number of so called authentication realms for plain files SOL databases or LDAP but there is no realm to use an XML database to store authentication information That s not really an issue since the realm interface is easy to implement This interface has been designed so that you can store the passwords either as plain text http en wikipedia org wiki Code injection http en wikipedia org wiki SOL injection S http en wikipedia org wiki AMP_ 28solution_stack 29 7 http en wikipedia org wiki Cross site_scripting http projects webappsec org w page 13246995 Web Hacking Incident Database http www w3 org TR xguery update 10 10 http tomcat apache org tomcat 6 0 doc realm howto html 178 XQuery Injection or encrypted Of course it s safer and recommanded to store encrypted passwords but for th
234. p expath org ns samples servlex hello xsl function app hello xslt gt lt url pattern xslt gt lt servlet gt This component represents a function app hello xslt which is defined in the stylesheet with the public URI http expath org ns samples servlex hello xsl Servlex constructs then the following reguest document lt web reguest servlet xslt path xslt method get gt lt web uri gt http localhost 8080 servlex hello world xslt who James Clark lt gt web uri gt lt web authority gt http localhost 8080 lt web authority gt lt web context root gt servlex hello world lt web context root gt lt web path gt lt web part gt xslt lt web part gt lt web path gt lt web param name who value James Clark gt lt web header name host value localhost gt lt web header name user agent value Opera 9 80 gt lt web request gt Servlex then calls the component with this request document In this case this is the XSLT function app hello xs1t An XSLT function used as a servlet must accept two parameters the first one is the element web request the second one is the se quence of bodies here empty as this is a GET request In this example this function 45 CXAN a case study for Servlex an XML web framework simply has to get the query parameter value from Srequest web param name eq who value and to format a simple HTTP response document and a simple
235. parent gt nc rpc reply nc data lt dsrl parent gt lt dsrl name gt cm state lt dsrl name gt lt dsrl default content gt lt cm supply levels gt lt cm water gt 0 lt cm water gt lt cm milk gt 0 lt cm milk gt 122 Configuring Network Devices with NETCONF and YANG lt cm coffee gt 0 lt cm coffee gt lt cm supply levels gt lt dsrl default content gt lt dsrl element map gt lt dsrl element map gt lt dsrl parent gt nc rpc reply nc data cm state lt dsrl parent gt lt dsrl name gt cm supply levels lt dsrl name gt lt dsrl default content gt lt cm water gt 0 lt cm water gt lt cm milk gt 0 lt cm milk gt lt cm coffee gt 0 lt cm coffee gt lt dsrl default content gt lt dsrl element map gt lt dsrl element map gt lt dsrl parent gt nc rpc reply nc data cm state cm supply levels lt dsrl parent gt lt dsrl name gt cm water lt dsrl name gt lt dsrl default content gt 0 lt dsrl default content gt lt dsrl element map gt lt dsrl element map gt lt dsrl parent gt nc rpc reply nc data cm state cm supply levels lt dsrl parent gt lt dsrl name gt cm milk lt dsrl name gt lt dsrl default content gt 0 lt dsrl default content gt lt dsrl element map gt lt dsrl element map gt lt dsrl parent gt nc rpc reply nc data cm state cm supply levels lt dsrl parent gt lt dsrl name gt cm coffee lt dsrl name gt lt dsrl default content gt 0 lt dsrl default content gt lt dsrl e
236. pecialized semantics for the XML received For example an XML Reader based application can load an Atom feed selecting only specific entries and when the atom content element is detected a small XML DOM could be build for the content This would allow mixing of Javascript objects and XML data In this example an auto binding actor wraps the XMLReader interface For each XML element it receives it makes a determination of whether it has local structure or is just a slot value If it is a slot value a property is set with a coded version of the name If the element has children or attributes the element is converted into a sub object with its own properties For example the following XML lt order count 1 gt lt item gt lt price gt 12 50 lt price gt lt quantity gt 2 lt quantity gt lt description gt The lt strong gt best price lt strong gt lt description gt lt item gt lt order gt would get automatically translated into Count LE item price 12 50 quantity 2 description strong best price By adding hints the JSON binder can turn certain XML contexts into other data types For example the element price is a float the element quantity and attribute count are integers and the element description is mixed content We d like the outcome to be coune s Ly 145 Efficient XML Processing in Browsers item price EA oy quantity
237. pens when we register an event handler b addEventListener b dom input onclick xs QName local somehandler Inside XQuery in the Browser there is one DOM event handler function which handles all incoming DOM events and dispatches them to the relevant XQuery functions It also aggregates the subscriptions from the XQuery code and pushes the relevant subscriptions into the DOM so that only the necessary events are being generated To keep track of which event handlers are currently registered there is a Hashmap which has the following declaration static HashMap lt NodeAndEventName List lt String a QName gt gt handlers NodeAndEventName is a helper class to provide means to use a node reference and an event name as a key So when we get an event in our universal event handler 203 XQuery in the Browser reloaded we get the target node and the event name from the DOM event object Then all functions in the list are invoked An advantage of having only one function is that it is easy to unsubscribe events When e g b removeEventListener b dom input onclick local somehandler is called the universal event handler gets the list of handlers for the pair element onclick It removes local somehandler from the list If the list is empty now we can remove the event handler in the browser using the JavaScript function element removeEventListener To resolve the function name and to call a function
238. provements that could be made to the rewrite rule notation presented In the future it seems likely that they might benefit from multiple distinct rewriting phases 5 and from a mechanism for describing free variable substitutions 7 They are yet to touch on cost based query optimization which relies on a cost analysis phase and the ability to speculatively apply multiple altern ative rewrites of any given expression The rewrite rule notation presented in this paper has shown its potential effect iveness both for XQuery implementers and XQuery library writers or power users Hopefully it has also been useful for communicating a number of interesting optim izations that are often applied by XQuery implementations This is just the start of a conversation on effective writing and communication of XQuery optimizations which I hope will continue and result in better language tools for everyone A EBNF for XQuery Rewrite Rules This appendix uses the same notation as the XQuery specification 1 Non terminals not explicitly defined by this paper are references to non terminals in the XQuery grammar The XQuery non terminal PrimaryExpr is extended by this paper for the purposes of rewrite rules RWRule RWName RWPattern RWCondition RWCaset RWName QName RWPattern ExprSingle RWCase gt RWResult RWCondition RWResult ExprSingle RWCondition where ExprSingle PrimaryExpr RWExprWildca
239. r is needed to avoid to raise a parsing error 5 to open again a literal string using the same quote Let s take care of the syntactic sugar first We ll assume that the XQuery expression is following this generic pattern lt URL gt query lt PATH gt lt SUBPATH gt lt entry value gt 179 XQuery Injection Our entry value can follow this other pattern or lt ATTACK gt or After injection the XQuery expression will look like lt URL gt _query lt PATH gt lt SUBPATH gt or lt ATTACK gt or The inner or expression has 3 alternatives The first one will likely return false the lt SUBPATH is meant to be the relative path to the user name and most applications won t tolerate empty user names in their databases The XQuery processor will thus pull the trigger and evaluate the attack statement The attack must be an XQuery Expr production That includes FLOWR ex pressions but excludes declarations that belong to the prologue In practice that means that we can t use declare namespace declarations and that we need to embed extension functions call into elements that declare their namespaces What kind of attack can we inject The first kind of attacks we can try won t break anything but export information from the database to the external world With eXist this is possible using standard extension modules such as the HTTP client module or the mail module These modules can be activated or desactivat
240. rams and maps they do not go through the same stages as do the books of the Bible and they are transformed to XHTML for the EPUB as a separate step from generating XHTML from the OSIS for the books The challenges in generating the EPUB files were e Making handling the hierarchy of books sufficiently flexible we knew Protestant and Catholic versions included different books but we also found different books different book groups and different titles in use in the translations e Getting all ancillary files such as graphics and external notes both into the manifest since some EPUB reading systems and recent versions of epubcheck complain about mismatches and into the spine in the right order since people don t want to see footnotes pages when scrolling between chapters but do want to see notes that are introductions to major groups of books 3 1 1 OPF Each EPUB has its own template Content opf file that contains the Dublin Core metadata which we could not get from the OSIS for the books and a minimal manifest and spine containing entries for items such as the title page copyright page and glossary The template is processed to add all the books chapters and note files twice once in the manifest and once in the spine and the figures are added once in the manifest only As a result the 57 line template Content opf for NBS expands to over 5 000 lines 160 EPUB Chapter and Verse 3 1 2 NCX Each EPUB similarly ha
241. rd RWExprWildcard QName B Arithmetic Rewrite Module from XQilla An example module of rewrite rules in this case the actual arithmetic folding rules used by XQilla to perform advanced constant folding on arithmetic expressions This module was developed by translating the existing C optimization rules to use the XQuery rewrite rule syntax xquery version 3 0 3 221 Declarative XQuery Rewrites for Profit or Pleasure Copyright c 2010 John Snelson All rights reserved Licensed under the Apache License Version 2 0 the License you may not use this file except in compliance with the License You may obtain a copy of the License at ttp www apache org licenses LICENSE 2 0 Unless required by applicable law or agreed to in writing software distributed under the License is distributed on an AS IS BASIS WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND either express or implied See the License for the specific language governing permissions and limitations under the Licens module namespace rw http xgilla sourceforge net rewrite declare option rw rule rw MulMulTransFold A B C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double gt A C B where rw is constant A gt A C B where rw is constant B declare option
242. re 7 our approach is beginning to show its limitations for the SPAROL translation Expressing each OPTIONAL expression as a nested loop with additional helper functions limits the 94 Translating SPARQL and SQL to XQuery ability of the XQuery optimizers to detect relevant information so that the XQuery implementations are now roughly as fast as ARQ Again Saxon and OizX scale best On the SQL side the query is expressed in somewhat simpler terms so that OPTIONAL is just another selection The results of the XQuery implementation are generally better but also diverge more Zorba and Sedna do not seem to detect any join optimizatins and scale fairly bad Saxon QizX and BaseX seem to detect the join and scale well when the document size increases with the former two mostly dominated by parsing MySQL however scales even better not being affected by the increasing workload This can be attributed to the very low selectivity of the query and the presence of predicates which can use the implicitly created primary key index 1000000 100000 10000 a ARQ Saxon 1000 Zorba 100 0izX 10 BaseX V 7 E X v E c S gt U x U U S i Sedna 10 50 100 500 1000 5000 scaling factor Figure 7 BERLIN SPARQL Query 8 95 Translating SPARQL and SQL to XQuery 1000000 A 100000 E 2 10000 MySQL 1000 Saxon o E 100 Z
243. res a callback function which gets sent the subtrees as they are ready The function re quires the full XML source and a pattern for the subtrees as in the following example The patterns are a subset of XPath from amara pushtree import pushtree from amara lib import U def receive nodes node print U node Will print 0 then 1 then 10 then 11 with input below return XML lt doc gt lt one gt lt a gt 0 lt a gt lt a gt 1 lt a gt lt one gt lt two gt lt a gt 10 lt a gt lt a gt 11 lt a gt lt two gt lt doc gt pushtree XML u a receive nodes Which should put out 0 1 10 11 You can also specialize the nodes sent to the callback The most common use for this feature is to deal with more friendly Bindery nodes rather than raw tree nodes from amara pushtree import pushtree from amara lib import U from amara bindery nodes import entity base def receive nodes node print U node b Will print 0 then 1 then 10 then 11 with input below return XML lt doc gt lt one gt lt a b 0 gt lt a b 1 gt lt one gt 69 Akara Spicy Bean Fritters and XML Data Services lt two gt lt a b 10 gt lt a b 11 gt lt two gt lt doc gt pushtree XML u a receive nodes entity factory entity base Which should put out same as the earlier example One can use a coroutine if you need easier state management in the push target from amara pushtree import pushtree from amara
244. rformance as for the simple retrieval query before 100000 a E 10000 g t ARQ 1000 Saxon o 5 100 r Zorba x QizX 10 BaseX gt i e Sedna 10 50 100 500 1000 5000 scaling factor Figure 3 BERLIN SPARQL Query 3 10000 A 1000 XL i f u E MySQL a s 100 Saxon S P 5 10 Zorba QizX 1 BaseX gt e Sedna 0 1 T T T T T 1 10 50 100 500 1000 5000 scaling factor Figure 4 BERLIN SOL Ouery 3 93 Translating SPARQL and SQL to XQuery For a query that uses a simple filter Figure 6 the results for both the SPAROL and the SQL translation closely resemble the result for the triple pattern specification in Q1 All optimizers detect enough relevant informtion to let the XQuery imple mentations scale better than ARQ while parsing becomes the dominant cost MySQL leads against the XML databases but the gap is closing quickly 100000 a E 10000 z t ARQ 1000 gt Saxon o 5 100 Zorba x 0izX 10 x BaseX gt i Sedna 10 50 100 500 1000 5000 scaling factor Figure 5 BERLIN SPARQL Query 6 10000 a E 1000 o m MySQL 100 Saxon o 5 10 Zr Zorba 5 O Qizx 1 BaseX gt 9 Sedna 0 1 10 50 100 500 1000 5000 scaling factor Figure 6 BERLIN SOL Ouery 6 For a guery with a join and multiple OPTIONAL statements Figu
245. ring Network Devices with NETCONF and YANG The NETCONF protocol is extensible which means that new operations even vendor or device specific may be defined and used along with the standard NETCONF operations For instance the following request can be used to instruct our coffee machine to start preparing a beverage lt rpc message id 125 xmins urn ietf params xml ns netconf base 1 0 gt lt put the kettle on xmlns http example com coffee gt lt recipe gt cappuccino lt recipe gt lt put the kettle on gt lt rpe gt The server s reply could in this case be just an acknowledgement lt rpc reply message id 125 xmlns urn ietf params xml ns netconf base 1 0 gt lt ok gt lt rpc reply gt An optional but very useful capability are event notifications 8 that may be used by the server to send asynchronous reports about various events Our coffee machine could implement the following notification lt notification xmlns urn ietf params xml ns netconf base 1 0 gt lt eventTime gt 2011 03 26T00 01 00 01 00 lt eventTime gt lt low level warning xmlns http example com coffee gt lt ingredient gt milk lt ingredient gt lt low level warning gt lt notification gt The NETCONF protocol was designed as independent of transport protocols en coding rules and other provisions specific to each transport protocols are standard ized separately So far four transport protocols have
246. ring str vesponse DOMString getAllResponseHeaders raises DOMException DOMString getResponseHeader in DOMString header raises DOMException readonly attribute unsigned short status getter raises DOMException readonly attribute DOMString statusText getter raises DOMException Extension 137 Efficient XML Processing in Browsers void overrideMimeType in DOMString override EventTarget interface void addEventListener in DOMString type in EventListener listener in boolean useCapture void removeEventListener in DOMString type in EventListener listener in boolean useCapture boolean dispatchEvent in Event evt raises EventException l The semantics of the open send abort getAllResponseHeaders getResponseHeader overrideMimeType addEventListener removeEventListener and dispatchEvent methods as well as the readyState status statusText and all the on event EventListener attributes are exactly the same as with the XMLHttpRequest interface The other methods and attributes from the XMLHttpRequest interface have been re moved as they are related to the whole document response interpretation What has been added is the parse method and the onxml attribute The parse method allows processing XML documents already loaded into strings using the same API and is provided for completeness The onxml and the subsequent XML events that are subscribe
247. rkLogic Server not contains key para string Hello World weight 1 0 hhh andNot positive contains key para string excel negative contains key para string proceed bh property contains key para string Hello World JI range key price value 15 operator lt bh geo parent location latKey latitude longKey longitude key latlong region point longitude 12 latitude 53 circle longitude 12 latitude 53 radius 10 box north 3 east 4 south 5 west 6 polygon longitude 12 latitude 53 longitude 15 latitude 57 longitude 12 latitude 53 bh collection recent filtered false score logtfidf I position 1 to 10 This evaluates using the following MarkLogic cts query structure cts or query cts element value query fn QName greeting Hello World case insensitive diacritic sensitive punctuation insensitive 32 A JSON Facade on MarkLogic Server whitespace insensitive unstemmed wildcarded lang en 2 cts not query cts element word query fn QName para Hello World lang en 1 1 cts and not query cts element word query fn QName para excel lang en 1 cts element word query fn QName para proceed lang en 1 cts properties query cts element word query
248. rogramming Its greatest advantage is that it is available in all modern browsers Because of the popularity of JavaScript a lot of resources in browser development go into optimizing its execution speed 1 3 9 and in the last few years impressive performance improvements have been achieved Another advantage is being able to manipulate the homepage directly Unlike in XSLT where the transformation operates outside the current page JavaScript allows editing the Web site directly through the DOM Document Object Model Example 1 Javascript embedded in HTML page lt html gt lt head gt lt script type text javascript gt window onload function var a document createElement div a textContent some text document body appendChild a lt script gt lt head gt lt body gt lt body gt lt html gt This Web site will display some text The user interaction with the browser is made accessible by the so called DOM events e g a button press mouse movement or keyboard input JavaScript can listen to these events and trigger actions based on them Example 2 Listening for an event var button document getElementByld buttonl button onclick function window alert button 1 was pressed 193 XQuery in the Browser reloaded Furthermore data can be downloaded as long as it comes from the same domain Example 3 Retrieving a document using XMLHttpRequest var reg ne
249. rom amara bindery model import label model examplotron model LABEL MODEL doc bindery parse TEST LABEL XML model label model print doc labels label guote None rather than raising AttributeError So even though the instance document doesn t have a quote element Amara knows from the model that this is an optional element If you try to access the quote element you get back the default value of None which can of course be overriden http xml3k org AttributeErrorf 66 Akara Spicy Bean Fritters and XML Data Services 4 1 Extracting metadata from models If the model uses inline declaration of particularly interesting parts of the document Amara provides a mechanism to extract those interesting bits as an iterators of simple tuples so one can in effect skip XML API altogether In the following ex ample the metadata extraction annotations are in the namespace given the ak prefix from amara xpath import datatypes from amara bindery model import examplotron model generate metadata from amara import bindery from amara lib import U MODEL A lt labels xmlns eg http examplotron org 0 xmlns ak http purl org xml3k akara xmlmodel gt lt label id tse added 2003 06 10 eg occurs ak resource id gt lt use ak resource for an anonymous resource gt lt quote eg occurs gt lt emph gt Midwinter lt emph gt Spring is its own lt strong gt season lt
250. rrent property values All properties are passed as parameters to all stage stylesheets and properties that are t used in a particular stylesheet are simply ignored by the XSLT processor There is an initial lines stage that is always included and is ignored if redund antly included in the stages specified in build properties that splits the text of the source SFM before each and wraps each resulting segment in a lt t line gt ele ment so the SFM can be further processed as XML in all other stages Since the lines stage runs an initial named template it is specified in the build xml file by using the lt java gt task to run the command line XSLT processor rather than using the lt xslt gt task as for the other stages The general sequence of the stages is e Convert SFM to XML e Fix up accented characters and language specific features such as inserting a no break space amp A0 after and before etc e Add structure from the outside in e Introduction and body It could be done with by Ant generating a new build xml on the fly and then Ant running Ant with the new build file Feature keys for initial template and mode and other features were recently added to Saxon in version 9 3 but that was after the project started and the operating systems of interest install an older version by default 154 EPUB Chapter and Verse e Sections e Chapter milestones e Verse start milestones e Footnotes e Poetry
251. rt an element into the DOM before all its children have been generated since changes to the DOM cause a browser to trigger a time consuming and possibly user visible repaint Replaces are done by inserting the source before the target followed by removing the target This way the implementation for replace is very short and we avoid code duplica tion Renaming of elements is a feature which is not natively supported by the browser s DOM so we use a workaround We first create a new node with the de sired name Second we set all attributes from the old node on the new node Third we set the parent of all children of the old node to the new node Finally we replace the old node with the new node 4 2 5 Handling Events In order to be able to use XQuery in the Browser to implement interesting web ap plications we need access to the DOM events These include keypresses mouseclicks and similar functionality For this we provide functions to add and remove event handlers in our browser namespace b more about that in Section 3 3 To add a function to handle certain events on an element we need three arguments the ele ments on which the event should be observed the event name and the identifi er name of a function which will be called when the event is triggered Since MXOuery does not support Function Items yet we have opted to take QNames as function identifers For a complete sample see Section 3 2 We will now take a look at what hap
252. rt is very handy to represent an XProc pipeline The component handles two different processing on a GET request it gets the package information from eXist and builds the corresponding abstract page on a PUT the request must con tain at least the package itself the XAR file The webapp analyzes it saves it on the disk and update the information in eXist Of course that means doing some checks is the uploaded content a proper EXPath package does the package already 50 CXAN a case study for Servlex an XML web framework exist in the database if yes is the new package consistent with the existing inform ation The advantage of XProc is that the developer can use whatever technolgy that best suits his her needs XProc itself for the flow control XSLT to transform documents and XQuery to query or update a database We have seen so far how the website serves HTML pages to a browser But as we Saw earlier this is only one of the two goals of the CXAN website The second goal is to enable tools to communicate with it in order to search for packages download them display information about them and upload new packages for instance from the command line Indeed semantically this is exactly what it already does with webpages Tech nically instead of serving HTML pages it just has to serve XML document carrying the same information Because the same set of URI is used in both cases it uses the REST principle based on the HTTP head
253. rw rule rw MulDivTransFold A div B C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double gt A C div B where rw is constant A gt A C div B where rw is constant B duration div duration decimal declare option rw rule rw DivMulTransFold A B div C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double and not rw subtype A xs duration and not rw subtype B xs duration and not rw subtype C xs duration gt A div C B where rw is constant A gt A B div C where rw is constant B declare option rw rule rw DivDivTransFold A div B div C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double and not rw subtype A xs duration and not rw subtype B xs duration and not rw subtype C xs duration 222 Declarative XQuery Rewrites for Profit or Pleasure gt A div C div B where rw is constant A gt A div B C where rw is constant B declare option rw rule rw AddAddTransFold A B C where rw is constant C and rw subtype this xs decimal or rw subtype this xs float or rw subtype this xs double gt A C B where rw is consta
254. ry blends a database query language transformation presentation language and general purpose functional language As such it stands to inherit richly from the bodies of research done in both database optimization and functional program ming language optimization as well as new avenues of research in XML and XPath based indexes and optimizations With so much optimization heritage the reality of the situation in many XQuery databases and stand alone implementations has the potential to disappoint Many and varying optimizations remain unimplemented in different products for lack of experience time and breadth of implementation approach In reality it can take significant maturity of product in order to have researched and applied knowledge from all potential resources 211 Declarative XQuery Rewrites for Profit or Pleasure XQuery users on the other side of this equation often know what optimizations they want to be applied to their programs but lack a means of influencing the im plementation Optimization hints are common in SQL implementations 4 and rewrite rules have existed in functional languages for some time 5 but no such mechanism exists for XQuery Part of the solution it seems might be found in a small domain specific language to describe XQuery rewrites so that optimizations can easily be written shared and discussed This paper describes such a domain specific language its potential uses and short comings
255. s For example if the target node of an augment happens to come from a grouping then the grouping must be expanded at the place where it is used so that the augment may be applied to this particular use of the grouping but not to other uses of the same grouping elsewhere In practice an implementation of the mapping algorithm has to process augments from all input modules into a list of patches and apply them at the right places when traversing the schema tree 4 3 Mapping Semantic Constraints Schematron rules are constructed from semantic annotations appearing in the hybrid schema For example the hybrid schema for our coffee machine see Section A 1 contains the following RELAX NG element pattern in which the nma key declares the cm address child element as the list key 110 Configuring Network Devices with NETCONF and YANG lt element nma key cm address name cm ipv6 gt lt element name cm address gt lt ref name ietf inet types ipv address gt lt element gt lt element gt The nma key annotation is mapped to a Schematron rule which has to check that each value of the list key is unigue In the Schematron schema for a reply to the get config operation in Section A 4 the resulting rule is lt sch rule context ne rpc reply nc data cm configuration cm ipv6 gt lt sch report test preceding sibling cm ipv6 cm address current cm address gt Duplicate key cm address lt sch report
256. s 4 forms The first form asks for our name then sends it to a page im plemented in XSLT The second form sends it to a page implemented in XQuery and the third one in XProc The last form sends you to an online tool that dumps the XML request document representing the HTTP request in XML located at http h2oconsulting be tools dump If you fill in the first form with James Clark and press the button you see a simple page with the text Hello James Clark in XSLT 44 CXAN a case study for Servlex an XML web framework e gt o C C web tocatnost 8090 serviex hello world v Hello world Welcome Type down your name to see the result using different technologies Name james Clark XSLT Name XQuery Name Proc Or maybe you want to see the XML representation of the HTTP request Go to this tool or access it via the following form Name Dump When you press the button XSLT the HTML form sends a simple HTTP GET request to the URI http localhost 8080 servlex hello world xslt who James Clark When Servlex receives this request it first extracts the context root to determine which application is responsible for handling this request The string hello world helps it identifying the application and finding its webapp descriptor In this descriptor it looks for a component matching the path xs1t It finds the following match lt servlet gt lt xslt uri htt
257. s an attribute results in inserting Y into X e Delete where X is a selected element or a selected attribute results in deleting X from the XML tree We assume that X is not a root element in our development we cannot delete the root however technically it is possible e Edit where X is a selected text or attribute value and Y is a pressed key results in performing the Y action over X i e changing the text content Client side XML Schema validation In each case the change is constrained to result in a schema valid context element where X is the context element for Insert Into operations and X s parent is the context for Insert After Insert Before and Delete 4 2 Algorithm Details Insert After and Insert Before operation are similar and requires access to the FSM These actions are not defined for the root element because it is legal to have only one root The Insert After algorithm works in the following way e find the parent P of the selected element SE parent is always of complex type If P is of mixed type we can always insert text after e find FSM of P e find the SE in the FSM e find element after EA the SE e compare maxOccurs field of SE and of all SE edges in FSM with real occurrences in P RO If RO lt maxOccurs then the SE or edge is a candidate If the edge is the wild edge for wild cards then the candidate is any element that is defined by W3C XML Schema If the edge is exit edge then
258. s into XML Verse end milestones are added at a late stage since e SEM contains only verse start codes since you don t print anything at the end of a verse e OSIS best practice limits where you should put the milestones in ways that can t be enforced by the OSIS schema e Inline OSIS elements such as lt hi gt and lt w gt for glossary terms can t contain lt verse gt elements No reason is given in the OSIS documentation but it s possibly because those inline elements are also allowed in lt note gt etc where a lt verse gt would be out of place 155 EPUB Chapter and Verse 2 3 1 Example The following SFM and OSIS fragments are from La Nouvelle Bible Segond NBS Psalm 3 4 9 Cest au SEIGNEUR qu appartient le salut Que ta b n diction soit sur ton peuple Pause 4 Exauce au milieu de l hostilit 1 Du chef de cheur Avec instruments cordes Psaume De David Example 2 SFM g v 98 C est au S cEIGNEUR x qu appartient le salut fr 3 98 f ilbau S cEIGNEUR x i bgu appartient le salut x 37 39 Jon 2 10 Pr 21 31 Voir i bb n diction g 8 x q Que ta b n diction soit sur ton peuple n Pause s Exauc au milieu de l hostilit c 40 d v 1 Du chef de cho e ur Avec instruments cordes Psaume De David Example 3 OSIS lt l level 1 gt 0 lt verse osisID Ps 3 9 sID Ps 3 9 n 9 gt lt note gt lt reference type source osisRef
259. s its own template epb ncx that contains the placeholder lt navMap gt for the table of contents The specific hierarchy of books for the EPUB is transformed into lt navPoint gt elements with additional lt navPoint gt for external notes that attach to the hierarchy or to specific books Titles in the hierarchy when transformed into the NCX file point to the first book that they contain for example in the table of contents the labels for Ancien Testament La Pentateuque and Gen se all point to Gen xml Every lt navPoint gt may have a playOrder attribute indicating its place in the normal playback sequence of the book Since multiple lt navPoint gt point to the same file the stylesheet processes the generated lt navMap gt in another mode to add the correct playOrder values lt xsl template match ncx navPoint mode hierarchy gt lt xsl variable name number select count preceding ncx navPoint empty ncx navPoint 1 as xs integer gt lt xsl variable name next book number select count descendant ncx navPoint class T preceding ncx navPoint empty ncx navPoint as xs integer gt lt xsl copy gt lt xsl copy of select gt lt xsl attribute name id select concat navpoint lt xsl attribute Snumber gt name playOrder select if class eg category then Snext book number Cel se at
260. se in the package e g for an XQuery library this is its namespace URI for an XSLT stylesheet this is set using the xsl stylesheet pkg import uri element When the developer invokes the command xproj build from within the project directory it uses those informations to build automatically the package descriptor and the package itself and put the result in the dist directory The package is then ready to be directly deployed in Servlex I will not discuss here the details of the xproj program written in XProc of course this could be the subject of paper on its own The idea is to use some kind of annotations in order to configure the public import URIs within each component instead of having to maintain an external package descriptor as needed to build a proper EXPath package While that is not supported yet it is expected to create the same kind of mechanism for the web descriptor Using some annotations it is easily possible to maintain the URL mapping and the parameters accepted or re quired by a component directly within the source file the query stylesheet or pipeline By doing so the developer will not have anymore to maintain the expath web xml descriptor manually it will be generated based on those annotations 6 Conclusion The main goal of the Webapp Module is to be the glue between XML technolgies and HTTP on the server side The design choice is to provide the full power and flexibility of HTTP This choice can make t
261. sed for exml anonymous before a predicate could even be omitted 4 JSONP Support 4 1 Request Submission When a script element is programmatically added within an HTML page it is im mediately executed and specifying a src attribute for this script element allows adding parameters There is no cross domain limitation because this is a good way to load external Javascript libraries The name of a callback function has to be sent to the server to allow it to integrate it in its response 4 2 Response Processing The callback function is called by the Javascript source received for the added script element The response is the parameter of this function call So it is up to this function to convert the received JSON object into its internal XML representation Once it is converted it is treated as if it was received as an XML instance 5 The Wikipedia Search Demo 5 1 The Wikipedia Search API The Wikipedia Search API allows retrieving entry names starting with a given string A GET request has to be built such as http en wikipedia org w api php ac tion opensearch amp search myvalue amp format json amp callback mycallback The returned script is an anonymous array with two elements the first one been the given string and the second one an array containing up to 10 entries For example http en wikipedia org w gt api php action opensearch amp search prague format json callback jsoninst will return jsoninst prague
262. sed with other databases Parameters may arguably be considered cleaner since they separate the inputs from the reguest They can also be used to call stored gueries 3 2 Java Assuming that we use single guotes to delimit XOuery string literals inputs can be sanitized in Java using this function Example 2 Java Sanitize Function static String sanitize String text return text replace amp amp amp replace amp apos Each user input must be sanitized separately and the whole query must then be encoded using the URLEncoder encode method Depending on the context it may also be a good idea to call additional method such as trim to remove leading and trailing space or toLowerCase to normalize the value to lower case In the authen tication realm the Java snippet could be a http demo exist db org exist functions request get parameter 183 XQuery Injection Example 3 Authentication Realm Sanitized String query URLEncoder encode user mail gt sanitize username trim toLowerCase UTF 8 reader parse new InputSource new URL http localhost 8080 orbeon exist rest db app users guery gt query openStream To use request parameters the query and each of the parameters need to be encoded separately Example 4 Authentication Realm Using Query Parameters String query URLEncoder encode declare namespace request http exist db org xqu
263. ser wars of the 90s The result is osisbyxsl a soon to be open source solution for Bible EPUB origination 1 Introduction 1 1 Typesetting ancient and modern The 42 line Gutenberg Bible was printed in 1455 9 The King James Bible was published in 1611 Four hundred years later the Bible is still a publishing phenomen on worth an estimated half a billion dollars per year in the United States alone in 2006 when Bible sales were double the sales of the latest Harry Potter book 12 Bible societies and independent publishers produce new translations at an ever faster rate and there is considerable demand for modern translations in formats amenable to New Media Amazon reported in January 2011 2 both that it is now selling more Kindle books than paperback books and the third generation Kindle eclipsed Harry Potter and the Deathly Hallows as the bestselling product in Amazon s history 149 EPUB Chapter and Verse In 2010 we were asked by the Alliance Biblique Francaise to convert five of their French Bible translations into EPUB and also to make these translations available via an online XML feed The text was provided in SFM a TROFF like ad hoc standard in the Bible publishing world After considering various alternatives we decided to first convert the SFM into OSIS an XML vocabulary developed by the American Bible Society before generating EPUB and other delivery formats Pro cessing from SFM to
264. sert Into Insert Before v O nsert After R Remove lt xml verston 1 0 encoding utf 8 gt B Element stree B lt gt address B Element fiat lt room gt 1 lt room gt lt address gt ff ROOT gt address gt street id xml ref 11 G Reset Validate Figure 4 Implementation Example noticeable delays Therefore we consider that in general cases the performance is satisfactory For testing the validation process we used a set of 65 tests A test case provides XML W3C XML Schema and the expected result The test suite was designed for the XSV application and while we implemented the same validation engine we ex pected the same results The evaluation of the editing process is challenging because it is difficult to define test cases and evaluate the output We constructed a test suite consisting of simple and widely used W3C XML Schema structures mostly taken from 10 We performed manual testing by clicking all elements and compare the result with expected The tests were performed successfully 6 Discussion The provision of client side schema validation functionality opens up a range of improved user friendly XML based applications Existing WYSIWYG or not schema constrained editing tools are either proprietary or limited to a single built in schema our work enables open source fully general development in this space Schema based database entry tools which guarantee integrity are also n
265. shed loading The XDM returned by the guery body is then inserted in the 196 XQuery in the Browser reloaded page after the script tag Of course a main module may define functions local namespace but their scope will only be this module itself and they cannot be im ported by other module For each module the XHTML namespace is set as the default element namespace A built in module with browser specific functionality the browser module is also imported 3 3 Browser specific functions We have defined the following browser specific functions in the namespace http xqib org browserapi with the prefix b Table 1 Functions in the browser namespace Signature Side ef Semantics fecting b dom as document returns the currently displayed docu ment b getStyle Swhat as returns the value of the style with the element stylename as name stylename of element what xs string as xs string b setStyle Swhat as element sets the style with the name Sstylename as xs string Sstylename of element what to Snewvalue as xs string Snewvalue b addEventListener where as Adds an Eventhandler to the ele element Seventname as ment s Swhere which listens for the xs string listener as event with the name Seventname xs QName When the event is fired it will call the function with the QName Slistener and arity 2 b removeEventListener where as yes Removes an event listener pr
266. si http www agencexml com exi gt xmlns gt lt a exsi maxOccurs unbounded gt stringA lt a gt lt a exsi maxOccurs unbounded xsi type xsd double gt 42 lt a gt lt b exsi maxOccurs unbounded xsi nil true gt lt c exsi maxOccurs unbounded gt lt exml anonymous gt and for default context e is array a equals true e array length a equals 2 e a 1 equals stringA e a 2 equals 42 e is array b equals true e is non empty array b equals false e array length b equals 0 e is array c equals true e array length c equals 1 e c 1 equals 3 2 6 JSON Anonymous Arrays An element with a reserved name in a specific namespace has to be used Example stringA 42 c stringCl d stringD1 ey stringa d stringD2 would be serialized as lt exml anonymous xmlns exml http www agencexml com exml xmlns xsi http gt www w3 org 1999 XMLSchema instance xmlns exsi http www agencexml com exi xmlns gt 19 JSON for XForms ous exsi maxOccurs unbounded gt lt exml anonymous exsi maxOccurs unbounded gt stringA lt a gt lt exml anonymous exsi maxOccurs unbounded xsi type xsd double gt 42 lt a gt lt exml anonymous exsi maxOccurs unbounded gt lt exml anonymous exsi maxOccurs unbounded gt lt exml anonymous exsi maxOccurs unbounded xsi nil true gt lt exml anonymous gt lt exml anonymous exsi
267. stores 2 Operational state data that are always read only 3 NETCONF messages that are exchanged in both directions between the client and server 4 Custom RPC operations 5 Custom event notifications Configuration datastores and operational state data need not be implemented as XML databases but conceptually the server presents them to the client using a re stricted XML infoset model 103 Configuring Network Devices with NETCONF and YANG The building blocks of YANG data hierarchy are similar to RELAX NG e A leaf corresponds to a leaf element in XML terms Every leaf has a type and may also have a default value e A container corresponds to an element which contains leafs other containers lists or leaf lists YANG also distinguishes containers that only play an organiz ational role from those whose presence has a semantic meaning such as switching on a certain function e A leaf list represents a sequence of leafs A minimum and maximum number of elements in the sequence may be specified e A list is similar to a leaf list but represents a sequence of containers Every list declares one or more child leafs as the list key that must uniquely identify every element in the sequence e A choice represents multiple alternative content models that are allowed at a given place One of the alternatives may be designated as default e An anyxml node represents arbitrary XML content including mixed content which is oth
268. struction used in this example this might be because the application does not encode the query for URI In that case we must do it ourselves and encode the string before copying it into the entry field like the XPath 2 0 encode for uri would do And then bingo 4 1 nouveau courriel Figure 2 New message We have a new message with all the information we need to login Fichier Edition Affichage Message VBS 9884 De vdv dyomedea com A vdv dyomedea com Sujet eXist collection Date Thu 3 Feb 2011 13 04 32 0100 The collection is lt user xmlns xs http www w3 org 2001 XMLSchema gt lt mail gt vdv dyomedea com lt mai gt lt role gt orbeon admin lt role gt lt password gt secret lt password gt lt user gt Figure 3 The mail 181 XQuery Injection The second kind of attack we can try using the same technique deletes information from the database A very simple and extreme one just erases anything from the collection and leave empty document elements for Su in user return update delete Su node Note that in both cases we have not assumed anything about the database structure SQL injection attacks often try to generate errors messages that are displayed within the resulting HTML pages by careless sites and expose information about the database structure but that hasn t been necessary so far On this authentication form generating errors would have been hopeless since
269. subtree with one in part derived from the original This problem description should be familiar to many XML practitioners XSLT 3 already solves a similar problem using recursively applied templates triggered by match patterns Similarly a good solution for XQuery rewrites starts with recurs ively applying pattern matching on expressions 212 Declarative XQuery Rewrites for Profit or Pleasure 2 2 A Simple Rewrite Rule The simplest rewrite rules identify an expression and unconditionally transform it into another expression fn CountEqZero count se eq 0 gt empty e The rewrite rule above consists of three components 1 Its name fn CountEgZero a QName followed by a colon The name is purely descriptive and can be useful when debugging the rewrite process 2 An expression pattern count e eq 0 This looks like an XQuery expression but allows the use of the tilde character followed by a QName to represent a named wildcard expression 3 The rewrite operator gt followed by a result expression empty e The result expression references the expression matched by the named wildcard in the expression pattern Wildcard expressions in expression patterns match any XOuery expression subtree and in the process assign them a name and make them available for use in other parts of the rewrite rule The expression pattern given above matches an expression eguivalent to an eguality comparison of
270. syntax to Relax NG compact syntax A useful automatic conversion will be from Relax NG to DTDs This will enable authoring DITA specializations in Relax NG and then still be able to use tools that do not provide support for Relax NG This is a main processing problem because that can be easily solved with a pre processing step that adds in the default attributes and then the documents with all the defaults added can be normally processed by any processing flow Some more work on the schemas Define URIs and an XML Catalog to map those URIs to the actual schemas and use the URIs when importing or referring to the schemas Identify more cases when Relax NG constructs can improve the design of the schemas Integration in DITA OT Ideally the Relax NG implementation will became part of the DITA OT distribution 6 References The DITA NG is an open source project hosted on Google code with an Apache 2 0 license The license was chosen to be the same to the DITA OT license to facilitate an eventual merge with that project The project main page is https code google com p dita ng Jing is available also from Google code http code google com p jing trang The Xerces project is available from Apache Software Foundation http xerces apache org xerces2 j The Relax NG DTD compatibility specification is available from http www oasis open org committees relax ng compatibility html 175 176 XQuery Injection Easy to exploi
271. t easy to prevent Eric van der Vlist Dyomedea lt vdv dyomedea com gt Abstract We all know and worry about SQL injection should we also worry about XQuery injection With the power of extension functions and the implementation of XQuery update features the answer is clearly yes and we will see how an attacker can send information to an external site or erase a collection through XQuery in jection on a naive and unprotected application using the eXist REST API This was the bad news The good news is that it s quite easy to protect your application to XQuery injection and after this word of warning We ll discuss a number of simple techniques literal string escaping wrapping values into elements or moving them out of queries in HTTP parameters to do so and see how to implement them in different environments covering traditional programming languages XSLT XForms and pipeline languages Keywords XQuery XQuery injection security Note I am not a security expert and as far as I know the domain covered by this paper is very new The list of attacks and counter attacks mentioned hereafter is nothing more than the list of attacks and counter attacks I can think of This list is certainly not exhaustive and following its advises is by now mean a guarantee that you Il be safe If you see or think of other attacks or solutions drop me an email so that I can improve the next versions of this document Many thanks to A
272. t stream stdout by default The latter returns a string pl xml write lt python spam eggs gt What do you mean bleh lt python gt You can manipulate the information content of XML nodes as well Some manipulation pl xml_attributes None u spam u greeneggs pl xml children 0 xml value u Close to the edit pl xml write 2 1 1 Writing XML and HTML from nodes As demonstrated above the xml write methods can be used to re serialize a node to XML to as stream sys stdout by default Use the xml encode method to re serialize to XML returning string These work with entity as well as element nodes node xml write Write an XML document to stdout node xml encode Return a UTF 8 XML string There are special methods to look up a writer class from strings such as xml and html from amara writers import lookup XML W lookup xml HTML W lookup html node xml write XML W Write out an XML document node xml encode HTML W Return an HTML string The default writer is the XML writer i e amara writers lookup xml The pretty printing or indenting writers are also useful node xml write lookup xml indent Write to stdout a pretty printed XML gt document node xml encode lookup html indent Return a pretty printed HTML string Note you can also use the lookup strings directly node xml write xml Write out an XML document node xml encode html Return
273. t with the Object Oriented Programming when you can have a derived class but an application can still handle that if it knows how to process a base class from the derivation hierarchy The class value contains exactly the whole derivation hierarchy and the processing tools work with those values to determine how an element should be processed There is nothing that prevents the use of Relax NG up to here You start to un derstand the problem when you look at a DITA document and you see no class at tribute All the class attributes are defaulted in the corresponding DTD or XML Schema It is not feasible to request from all the authors to enter consistent class values for all elements so having them defaulted in the DTD or Schema that the document refers to makes this totally painless and also does not leave any possibility for error As Relax NG does not have a mechanism for default values then in order to use DITA with Relax NG one has to specify all the class attributes and values explicitly in the XML documents As mentioned earlier that is not a feasible approach people will quickly give it up and return to DTDs instead The good news is that there is a Relax NG DTD compatibility specification that solves the problem of default values by specifying an annotation mechanism that can be used to specify a default value for an attribute However there is was no implementation for this The DITA specification mentions the DTD and XML Schema implementat
274. ta nodes leaf container leaf list list and anyxml1 as well as the choice statement map to RELAX NG patterns in a relatively straightforward way Grammatical constraints that are not supported by RELAX NG for example the minimum and maximum number of list entries end up as se mantic annotations in the hybrid schema and then are transformed to Schematron rules However YANG and RELAX NG differ in several important aspects In the following subsections we will mention two of them that presented the most inter esting challenges to the mapping procedure 4 2 1 Handling of Names YANG groupings and type definitions are often mapped to RELAX NG named pattern definitions but there are a few caveats First of all names of groupings and typedefs imported from another module keep the namespace of the module in which they are defined In contrast names of RELAX NG named pattern definitions share the same flat namespace even if they come from an included grammar Therefore the mapping procedure must disam http dublincore org 109 Configuring Network Devices with NETCONF and YANG biguate the names to avoid clashes it prepends the originating module name sep arated by two underscores Further RELAX NG named pattern definitions are always global inside a grammar whereas YANG allows for lexically scoped definitions appearing anywhere in the schema tree Again the mapping procedure has to disambiguate their names this t
275. tem is the specification of a package format for XML techno logies It uses the ZIP format to gather in one single file all components and resources needed by a package that is a library or an application The package contains also a package descriptor which associates a public URI to each public component of the package This URI can by used by user code to import those components exposed by the package The Packaging System defines also an on disk repository structure so different processors and different implementations can share the same repository of packages When compiling the user stylesheet pipeline query the processors simply delegate the resolution of imported components to the repository 35 CXAN a case study for Servlex an XML web framework resolve references Saxon eXist Calabash delegate compile user stylesheet This package format makes it possible to distribute XML libraries and applications in a standard way using a format supported by several processors All the library author needs to do is to provide such a package created using standard tools The user just downloads the package and gives it to his her repository manager or dir ectly to his her processor in order to install it automaticaly 1 1 CXAN The Packaging System makes it possible for a library author to put the package on his her website in order for its user to download it and install it automatically But still a user h
276. ter You set up an envelope of the XML structure and a marker to which you can send inner elements as you prepare them The following simple example from amara writers struct import structwriter E NS ROOT RAW E CURSOR class event handler object def init self feed self feed feed def execute self n self feed send E u event unicode n output structwriter indent u yes feed output cofeed ROOT E u log E CURSOR u events u type gt u numberfeed h event_handler feed for n in xrange 10 h execute n feed close Generates the following XML lt xml version 1 0 encoding utf 8 gt lt log gt lt events type numberfeed gt lt event gt 0 lt event gt lt event gt 1 lt event gt lt event gt 2 lt event gt lt event gt 3 lt event gt lt event gt 4 lt event gt lt event gt 5 lt event gt lt event gt 6 lt event gt lt event gt 7 lt event gt lt event gt 8 lt event gt lt event gt 9 lt event gt lt events gt lt log gt 4 Modeling XML XML is eminently flexible but this flexibility can be a bit of a pain for developers Amara is all about making XML less of a pain for developers and in Amara 2 0 you have a powerful new tool You can control the content model of parsed XML docu 63 Akara Spicy Bean Fritters and XML Data Services ments and you can use such information to simplify things with just a little up front work
277. tern combinations and duplicate elimination among them are explicit and implicit join reordering incorporating XQuery 3 0 features like native group by outer joins and error handling as well as index usage Further more we plan to investigate the upcoming SPARQL 1 1 language which provides updates grouping and aggregation All of these features should be easy to express in XQuery Bibliography 1 W Akhtar et al XSPAROL Traveling between the XML and RDF worlds 5th European Semantic Web Conference ESWC 2008 2 ARQ a SPARQL processor for Jena http jena sourceforge net ARQ 3 D Beckett et al RDF XML Syntax Specification 4 N Bikakis et al The SPAROL2XQuery Framework Technical Report National Technical University of Athens Greece 2008 5 C Bizer A Schultz Benchmarking the Performance of Storage Systems that expose SPAROL Endpoints SSWS 2008 6 S Boag et al XQuery 1 0 An XML Query Language http www w3 org R xguery 7 P Fischer XOBench A XOuery Benchmarking Service XML Prague 2010 poster http xgbench org 8 S Groppe et al Embedding SPAROL into XQuery XSLT SAC 2008 9 T Grust J Rittinger and J Teubner Pathfinder XQuery Off the Relational Shelf ICDE Bulletin Special Issue on XOuery Vol 31 No 4 December 2008 97 Translating SPARQL and SQL to XQuery 10 Martin Kaufmann Mapping SPARQL and SQL to XQuery Master s Thesis ETH 2010 11 Martin Kaufmann Web interf
278. that allowed chunking of events to reduce the number of callbacks by grouping a set of events into a single event A numberOfEvents attribute was added to the XMLItemEvent interface along with a getEvent method that returns an event based on an index This avoids the need to create an array data structure When the chunking was enabled and the chunk size was set to 32 the parsing time decreased to a 2 times penalty While reduced the delivery of events by invok 142 Efficient XML Processing in Browsers Memory Usage m 1MB m 10MB m 100MB Memory MB 2000 3000 4000 5000 6000 7000 1000 l M 0 xhr events no handler Figure 1 Memory Performance in Comparison to XMLHttpRequest ing Javascript based event handlers incurs a significant penalty This penalty cannot be avoided but may be reduced by improvements to the scripting environment For example the number of events for the 1MB test document is 526306 and so that translates into the same number of function invocations within Javascript When the events are chunked into groups of 32 that translates into 16448 function invoc ations The chunking reduces the overall processing time by around 40 but still leaves parsing at about twice as slow The sheer number of events to deliver and as a result functions to invoke limits the speed of parsing 143 Efficient XML Processing in Browsers Parsing Time m 1MB S E 10MB IE 100MB o O m oO
279. that is a domain is included the corresponding domain value also appears The domains attribute in Relax NG is defined like this lt define name domains atts combine interleave gt lt optional gt lt attribute name domains a defaultValue topic concept topic hi d topic ut d topic indexing d topic hazard d topic abbrev d topic ui d topic pr d topic sw d gt lt optional gt lt define gt and each module contains a marker domain pattern like lt define name domains atts value combine choice gt lt value gt topic concept lt value gt lt define gt which allows to automatically check with a Schematron schema that if a module is included then the corresponding value is added to the domains attribute 3 3 2 Automatic domain extensions In DTD each shell needs to specify what elements extend for example the pre element lt ENTITY pre pre Spr d pre Ssw d pre Sui d pre while in the case of Relax NG each domain adds automatically the pre element to the corresponding pattern and the main Relax NG schema only includes the domain lt include href programmingDomain mod rng gt and the included schema contains lt define name pr d pre gt lt ref name codeblock element gt lt define gt lt define name pre combine choice gt lt ref name pr d pre gt lt define gt Similarly the other included schemas define their contributions to the pre pattern 172
280. the HTTP response a count of elements therein The simple case working involves with a Python function that takes a few parameters and returns a result wrapping this whole thing as a Web service For this case you can use the simple_ service decorator import amara from akara services import simple service respons ECOUNTER SERVICE ID http purl org akara services demo element counter simple service GET ECOUNTER SERVICE ID ecounter text plain def ecounter uri Respond with the count of the number of elements in the specified XML gt document e g curl http localhost 8880 ecounter uri http hg akara info b testdoc xml doc amara parse uri 0 ecount doc xml select u count return str ecount All Akara services have an ID a URI ECOUNTER_SERVICE_ID in the above which represents the essence of that service i e its inputs outputs and behavior You and I might take the same Akara code and you host it on your server and I host it on mine The service ID will be the same in both cases but the access endpoint i e what URL users invoke to use the services will be different Use the simple service decorator to indicate that a function is a service and specify what HTTP methods it handles the ID for the service and the default mount point which is the trailing bit of the access endpoint URL If you mount this service on an Akara instance running at http localhost 8880
281. the same origin policy which should prevent cross site scripting Therefore we had to integrate all data into the the tests and could not load anything from outside 5 1 1 1 Test performance optimizations For each test the whole browser was restarted what took several seconds per test class So to optimize it instead of running a lot of test classes we combined them using the TestSuite class provided by JUnit This way we were able to run the whole XQTS in under an hour Because we wanted to have it still faster we saw that parsing the XML from strings was a bottleneck Therefore we cached the token stream This reduced the combined build test time to under 20 minutes This is a good result considering that also the native Java version of MXQuery uses 3 to 4 minutes for the test suite 5 1 2 Testing results Our goal was of course to be 100 compatible In this section we will evaluate how far we came to achieving this goal and the limitations of the platform or our design choices Among the minimal compliance test of XOTS 1 0 2 we currently pass 14433 giving us 98 6 percent conformance In addition we pass all Use Case and Full Axis tests Of the 468 total cases which we do not pass we skip the following e The unicode character assemble and disassemble tests 89 do not compile because the GWT compiler has trouble finding the correct encoding e We have implemented fn doc to load documents from the url given using XMLHttp
282. the start of lines except when it did but we didn t find those few cases including 542 chapters in a row of italic text in one translation until late in the process Other inline markup included markers for glossary terms and for footnote refer ences Glossary markers are codes usually g except when it s not at the start of the word from the glossary except glossary terms can be multiple words the multiple words can have a line break between them the glossary marker can be at the end of the term instead of the beginning or the glossary term can have markup at both its start and end Footnote references have their footnote text on a following line usually after the end of the scope of the current V SFM code except when it s not Footnotes conceptually aren t part of the flow of the text so they don t take part in grouping lines into paragraphs or line groups and some of the SFM even had inline highlighting markup that started in one q line before a footnote text and closed in another q line after the footnote text An early stage in the processing of SFM into OSIS is the moving of footnotes into the place of their footnote marker where they still interfere with the processing of inline highlighting markup In other words SFM as practised by the Bible translating community is a textbook demonstration of why Schema are a good idea However Bible translators are not early adopters of new technology
283. ting out unwanted functionality such as XQuery updates serialization support for JAXP Java ex tension functions and unnecessary options like the choice between TinyTree and Linked Tree or the choice never in practice exercised of different sorting algorithms The next 20K was a little more difficult and required more delicate surgery For example it involved changes to the XPath parser to use a hybrid precedence parsing approach in place of the pure recursive descent parser used previously offloading the data tables used by the normalize unicode function into an XML data file to be loaded from the server on the rare occasions that this function is actually used rewriting the regex handling to use the regex library provided by GWT which is merely a wrapper around the Javascript regex library sometimes combining two types of expression implemented separately to use a single generic implementation A further reduction from 60K to 40K will be more challenging still and may turn out not to be necessary It may involve removing some optimizations and some of the performance oriented machinery such as the NamePool A useful measure may be to implement parts of the standard function library in XSLT 128 XSLT in the Browser itself to be compiled on demand It may involve being a little bit less dogmatic about conformance in edge cases such as matching of non BMP characters in regular expressions when there s a minor difference bet
284. tion This is the subject of the next section 3 The Future I believe that to make the case for using XSLT on the browser compelling it has to do everything that can currently be done with JavaScript on the browser and it has to do it better In the first place XSLT has to be able to handle user initiated events We have a head start here because XSLT s natural processing model is event driven Suppose for example we want to sort a table when the user clicks on the column heading We should be able to write lt xsl template match th mode onclick gt lt xsl apply templates select ancestor table 1 mode sort gt lt xsl with param name colNr select count preceding sibling th 1 gt lt xsl apply templates gt lt xsl template gt lt xsl template match table mode sort gt lt xsl param name colNr as xs integer required yes gt lt xsl result document href select method replace content gt lt xsl copy of select thead gt 131 XSLT in the Browser lt tbody gt lt xsl perform sort select tbody tr gt lt xsl sort select td ScolNr gt lt xsl perform sort gt lt tbody gt lt xsl result document gt lt xsl template gt Note that like CSS styling rules this is a generic rule that applies to every element that matches the match pattern In the same way as CSS has moved the description of styles from individual element instances to whole classes of elements
285. tion of our translation tests ona range of sample queries covering all language features turned out to be correct The translated XQuery tests were run on the XQBench Benchmarking Service 7 putting native SPAROL execution on ARQ Jena 2 8 7 2 and SQL on MySQL 5 1 41 against a number of open source XQuery processors and databases capturing exe cution results and timings Saxon HE 9 3 0 2 Qizx Open 4 1 XQilla 2 2 4 Zorba 1 5 pre release processors BaseX 6 5 MonetDB Pathfinder October 2010 Sedna 3 4 BerkeleyDB XML 2 5 16 and eXist 1 4 0 databases To the best of our knowledge this is a fairly complete collection of freely available XQuery implementations For an initial performance study we chose the Berlin SPAROL Benchmark 5 which is one of the few existing benchmarks suites for SPAROL It provides queries and a data generator for both SPARQL and SQL on the same conceptual schema The tests have been executed for Berlin scaling factors from 10 5K triples to 5000 1 8M triples on an Intel Xeon X3360 quad core 2 8 ghz with 8 GiB RAM and 2x 750 GB S ATA disks In our measurements we excluded start up warm up time by deducing the time of a dummy query and repeating the measurements until they became stable Furthermore we kept the setup of all engines to the out of the box settings and did not any additional optimizations like custom indexes The translation from both SPARQL and SQL to XQuery took around 10 25 msec while the executi
286. tion xquery gt module namespace mod http www example com module declare sequential function module messagebox Sevt loc b alert div 1 be lt script gt lt script type application xquery gt import module namespace my http www example com module b addEventListener b dom input onclick xs QName my messagebox lt script gt lt head gt lt body gt lt div gt some text lt div gt lt input type button value Button gt lt body gt lt html gt This will display a page containing a button When pressing the button it will display a message box with some text 3 2 Program structure XQuery code can be included in HTML script tags with a MIME type application xguery A page may contain several of these tags and each of these may contain either a main module or a library module The previous example is a page with a main module and a library module The main module imports the library module The semantics of a library module is that it is made available in the context Any other module on the page may import it We also allow importing modules from an external location instead of declaring in a script tag In this case a location hint needs to be specified and the same origin policy will be obeyed import module namespace m http www xqib org module at module xquery The semantics of a main module is that it will be evaluated as soon as the page has fini
287. to be converted in one go as in a traditional typesetting scenario where text flows between pages Selecting particular sections of content on the basis of milestones is much harder because milestones can be separated by arbitrary amounts of content and may not be siblings which is the whole reason for using milestones in the first place For example from the Bible en Francais Courant lt verse osisID Gen 12 5 sID Gen 12 5 n 5 gt Abram prit donc avec lui sa femme Sarai et son neveu Loth ils emportaient toutes leurs richesses et emmenaient les esclaves achet s a Haran Ils se dirig rent vers le pays de Canaan lt p gt lt title type x section gt Abram au pays de Canaan puis en Egypte lt title gt lt p gt Lorsqu ils arriv rent au pays de Canaan lt verse eID Gen 12 5 gt where a verse begins in the middle of a paragraph and ends in the middle of another paragraph separated from the first by a title osisbyxsl needs to split books into discrete chapters because users expect per chapter EPUB navigation and because several biblical books are too large for a single EPUB manifest item The first approach used a complex XPath to select on a combination of node order and sibling relationship However the search space was huge because no assumptions could be made about the location of the milestones When we found that splitting Psalms in one translation took two hours we were forced to look for an alternative s
288. to the user the output the other the communica tion from the user to the system the input where the structure of the dialog can be represented by the schemata for these two documents and the processing logic as a streamed transformation of the input document to the output document This gives a structured way to think about the states represented by URLs appearing in the browser window to which the user can return by hitting the back button Hopefully this will pave the way to a more declarative way of programming such dialogues in which the maintenance of state across user interactions is the respons ibility of the system not the individual web developer I hope also to examine how XForms fits into this picture Clearly if the aim is to eliminate low level JavaScript programming and replace it with a higher level more declarative style of user interface definition XForms must play a significant part 4 Conclusions In this paper I have outlined why I believe XSLT 1 0 on the browser has failed to capture the public imagination I ve described the mechanics of how the Saxon XSLT processor has been adapted to run natively in the browser by cross compiling it to Javascript And I ve described some ideas for how we can take advantage of the declarative event driven nature of XSLT to enable the language to be used not just for rendering XML for display on the screen but for handling interactions with the user and providing a much higher le
289. trategy The current approach uses six interlocking recursive functions to tree walk between the start and end milestones building a new document as it goes This has consequences for other aspects of the processing because for example generate id can no longer be used to link between chapters 163 EPUB Chapter and Verse However the performance benefit outweighs all other considerations in one case the time required to produce an epub has been reduced from 5 hours to 18 seconds 3 3 E Reader rendering For testing purposes we used e Calibre 4 a popular open source e reader for desktop machines e Adobe Digital Editions 1 for desktop machines e Sony Reader PRS350 e Apple iPhone 4 Those are some of the more capable e readers we tried some Android e readers did little more than split on paragraphs We expected some e readers to offer more functionality than others But we didn t expect e readers which excelled in some areas to fail in others and we did expect the general quality of XHTML rendering to approach that of a modern browser For example percentage values for vertical align worked perfectly in Calibre but seemed to be interpreted upside down by the Adobe reader which also failed to render drop caps that worked in the other two e readers But Calibre proved quirky at splitting lines of text and with spacing in general We failed to find a way to use superscript without disrupting line spacing but were
290. traversing the DOM but is no longer connected each token links explicitly back to its orginating DOM node and also draws as much information as possible from there instead of copying it as shown in Section 4 1 1 1 By doing so the tokens stay in sync with the 201 XQuery in the Browser reloaded DOM and we only create them when there is a specific access to the DOM e g by navigating a specific axis We changed to Token implementation of MXQuery to support this new kind of Token alongside to the existing eager standalone tokens Hello World Dom Token Stream Figure 2 Mapping HTML DOM to a lazy Token Stream 4 2 3 Node IDs XDM mandates that every node has a unique identifier which may also encode the document order or provide structural informations e g parent child relationships Node IDs may not change during the execution of a query In the standalone applic ation MXQuery generates IDs for the nodes while generating XDM from input or constructing new elements typically using ORDPath Dewey IDs We also considered this approach for the DOM5 HTML wrapping but then decided to utilize existing means from the DOM in order to avoid the overhead of generating new identifiers and the effort of keeping the DOM and the XDM identi fiers synchronized DOM nodes in JavaScript are uniquely identifiable using the Node interface We therefore can compare two nodes for equality using the Java operator It will correctly
291. type u poet name address info return u lt p gt Poet s of s lt p gt info elif entitytype u work name poet info return u lt p gt Work s lt a href s gt click for poet info lt a gt lt p gt info Focusing in on some key lines from akara import reguest entitytype shift path info reguest environ The reguest object which becomes available to your module through the import is the main way to access information from the HTTP request using WSGI conven tions such as the environ mapping The Python stdlib function wsgiref shift path info allows you to extract one hierarchical path component from the URL used to access the service So going back to the sample URL for a poet http localhost 8880 poetdb poet ep Akara itself is mounted at http localhost 8880 and the service defined above is mounted at http localhost 8880 poetdb The first wsgiref shift path info extracts the poet component There is a second one that extracts the ep component simple service GET POETDB SERVICE ID poetdb text html Notice the additional argument which declares the return content type The output of this service is HTML return u lt p gt Poet s of s lt p gt info The return value is a Unicode object You can return from an Akara service handler string or Unicode or even parsed Amara XML objects Deploy this module and restart Akara and now if you go to e g http localhost
292. uld almost certainly result in an error according to the XQuery specification 3 1 Normalizing Rewrites When considering correct rewrite rules there are appear to be two broad categories that beneficial rules fall into normalizing rewrites and optimizing rewrites XQuery is a broad declarative language and there is often more than one way to write any given operation Given that there is often a need to normalize the types of expres sions used into a smaller set of expressions so that the surface area of the language is smaller and it is easier to find more generally applicable optimizing rewrites The XQuery Formal Semantics 2 defines one set of normalizing rewrites al though it should by no means be considered to be the only set possible The choice of post normalization language often makes a difference to how easy it is to identify different optimizations Consider for instance the rewrite rule discussed in Section 2 2 which optimizes the expression count e eg 0 In XQuery there are two types of comparison oper ator eq and The former is a straightforward transitive equality operator oper ating on singleton items whereas the latter is existentially quantified and can operate over sequence operands This means that the rule matching the eq operator will not automatically match an expression using the operator However a normalizing rewrite can solve this problem NormalizeToEQ a b where rw subtype a
293. vel approach to web applications design than is customary today with JavaScript References 1 XSL Transformations XSLT Version 1 0 W3C Recommendation 16 November 1999 James Clark W3C http www w3 org TR xslt 2 XQuery in the Browser The same as JavaScript just with less code 23 January 2007 Ghislain Fourny et al XML Prague March 2010 Prague Czech Republic http www xmlprague cz 2010 files XMLPrague 2010 Proceedings pdf 3 XML Pipeline Processing in the Browser Toman Vojt ch Balisage Aug 3 10 2010 Montr al Canada 10 4242 BalisageVol5 Toman01 http www balisage net Proceedings vol5 html Toman01 BalisageVol5 Toman01 html 133 XSLT in the Browser 4 XQIB XQuery in the Browser JavaScript Edition ETH Zurich Systems Group and FLWOR Foundation http www xqib org p 15 5 Google Web Toolkit GWT Google http code google com webtoolkit 6 XSLTForms Alain Couthures agenceXML http www agencexml com xsltforms 7 The Saxon XSLT and XQuery Processor Michael Kay Saxonica http www saxonica com 8 Hash URIs Jeni Tennison 6 March 2011 http www jenitennison com blog node 154 134 Efficient XML Processing in Browsers R Alexander Milowski ILCC School of Informatics University of Edinburgh lt alex milowski com gt 1 Overview of the Problem The XMLHttpRequest XHR 1 interface has been available in various browsers since 1999 While the name is prefixed with
294. ves the original number or letter or number range or because there s too much variation in the SFM to be able to reliably reconstruct the number from osisID for presentation in the EPUB SFM codes beginning with f are grouped into a lt note gt Since NBS does not use footnote references in its verses the note is moved to the start of the verse where in the EPUB it is replaced by a that is a link to the footnote text Usually SFM codes beginning with denote the start or end of a highlight In most translations seen so far a x closes all open highlights the translation that has separate closing delimiters for nested highlights simply uses a different stylesheet at this stage g before or in NBS after a word indicates a glossary term Since some glossary terms are multiple words and since some translations have glossary markers for words not in the glossary the glossary stage also reads the glossary XML and matches on the words or phrases that appear in the glossary Verse ends are not marked in SFM OSIS best practice recommends that mile stones should not cross unnecessary boundaries so the verse end milestone is inserted at the end of the text rather than say after the lt I gt or even after the lt lg gt or lt div gt Chapter ends are not marked in SFM In the printed Bible c codes just print chapter numbers but the XML markup represents containment and the EPUB presents one chapter at a time so the titl
295. vided by Java s Calendar This class fixed a lot of tests but included information for all time zones which increased the file size by 300 KB and increased the loading time by more than a second Fortunately the XPath Data Model XDM does not require daylight savings time awareness So the exact time zone does not matter to an XQuery program only 204 XQuery in the Browser reloaded the time offset to the UTC We therefore kept the API of gwt calendar class but re implemented the relevant functionality significantly reducing the code footprint and the initialization delay The current implementation is very close to the XDM specification and has good test conformance see Section 5 1 2 This will need to be further improved to support conversion to the end user s time zone Also the implementation is leap second unaware and will therefore have small errors when computing durations 4 2 6 2 Deviating from the standard In some cases fully conforming to the standard would increase the download size a lot and the performance would decrease drastically One of these cases are regular expressions GWT does not provide the java util regex package because it would be difficult to implement it correctly As JavaScript already provides regular expres sion functionality we just used that one It may not offer all functions e g the dot all option is missing but its performance is much faster than an implementation in JavaScript because
296. vlet technology for the link to HTTP and it uses Saxon and Calabash as its XSLT XQuery and XProc processors To install Servlex you first need to get a Servlet container The easiest is to install a container like Tomcat or Jetty Then follow the instructions to deploy the Servlex WAR file in the container go to your container admin console select the WAR file on the disk and press the deploy button As simple as that The only option to con figure is the location of the webapp repository For instance in Tomcat you can add the following line in conf catalina properties org expath servlex repo dir usr share servlex repo This must be a standard EXPath package repository At startup Servlex looks into that repository Every package with a web descriptor aka the file expath web xml is considered a webapp The descriptor is read and the Servlex application list is initialized Each webapp has an abbreviation used to plug it in the container URI space For instance let us assume Servlex has been deployed on a local Tomcat instance at http localhost 8080 servlex When Servlex receives a request at http localhost 8080 servlex myapp some thing it uses myapp as an ID of the application Once it knows the application it can retrieve its map It then uses the path here some thing to find a component in the map by trying to match the path against components URL regex An interesting particularity of Servlex is its ability to have a read onl
297. w strftime SY m SdTSH M SZ E u category u term t for t in tags E u content u type u xhtml E XHTML NAMESPACE u div E u p u Happy to be here This generates an Atom feed and Atom is a pretty good torture test for any XML generator library The output lt xml version 1 0 encoding UTF 8 gt lt feed xmlns http www w3 org 2005 Atom xml lang en gt lt id gt urn bogus myfeed lt id gt lt title gt MyFeed lt title gt lt updated gt 2008 09 12T15 09 16 321630 lt updated gt lt name gt lt title gt Uche Ogbuji lt title gt 61 Akara Spicy Bean Fritters and XML Data Services lt uri gt http uche ogbuji net lt uri gt lt email gt uche ogbuji net lt email gt lt name gt lt link href blog gt lt link rel self href blog atoml 0 gt lt entry gt lt id gt urn bogus myfeed entryl lt id gt lt title gt Hello world lt title gt lt updated gt 2008 09 12T15 09 16 322755 lt updated gt lt category term xml gt lt category term python gt lt category term atom gt lt entry gt lt content type xhtml gt lt div xmlns http www w3 org 1999 xhtml gt lt p gt Happy to be here lt p gt lt div gt lt content gt lt feed gt A few interesting points e Structwriter tries to help the lazy hand a bit If you create an element with a namespace any child element without a namespace will inherit the
298. w XMLHttpRequest the last argument determines wheter the request asynchronously req open GET http www example org file txt false req send null if req status 200 HTTP OK window alert req responseText While the combination of DOM manipulation event handling and background downloads provides a powerful basis for rich Web applications these APIs are at a quite low level of abstraction and not fully standardized across browsers Therefore libraries which hide many of the compatibility issues and provide higher level APIs as well as UI components have gained popularity examples are jQuery or Dojo 2 3 Cross Compilation Google Web Toolkit The Google Web Toolkit GWT provides a full framework for creating web applic ations which allows a developer to write most of the code in Java It offers widgets and RPC mechanisms but the main innovation is a Java to JavaScript Compiler GWT implements a subset of the Java standard library in JavaScript thus allowing reuse of code on both the client and server side Means for dealing with multiple browser versions are also integrated in GWT As our approach makes use of GWT to compile our code to JavaScript we will describe GWT in more details in Section 4 1 2 2 4 XML based approaches XSLT and XQuery in the Browser 2 4 1 XSLT XSLT eXtensible Stylesheet Language Transformations is a declarative language to transform XML documents While there is supp
299. w to read configuration information The following listing is an Akara module for accepting reservations of business resources such as conference rooms and the like Note This example is designed to illustrate the mechanics of POST handling but is not a good example of REST style presented as it is for simplicity Akara does support strong REST principles including hypermedia and proper use of HTTP verbs import shelve from amara import bindery http xml3k org CherryPy 77 Akara Spicy Bean Fritters and XML Data Services from amara lib import U import akara from akara services import simple service DBFNAME akara module config dbfile NEWPOET SERVICE ID http example org my services new poet simple service POST NEWPOET SERVICE ID newpoet plain text def newpoet body ctype Add a poet to the database Sample POST body lt newpoet id co gt lt name gt Christopher Okigbo lt name gt lt address gt Christopher Okigbo lt address gt lt newpoet gt dbfile shelve open DBFNAME Warning no validation of incoming XML doc bindery parse body dbfile U doc newpoet id U doc newpoet name U doc newpoet address dbfile close return Poet added OK This module requires a configuration variable dbfile which you can provide by adding the following or similar to akara conf class tutorial post dbfile tmp poet Once the service is running
300. we are moving the specification of behaviour from the instance level to the class level The class of elements to which the behaviour applies can be specialized of course by using a more specific match pattern but if written in this form it produces consistent behaviour for every table in the document This is already a significant improvement on the JavaScript way of doing things There s some magic of course buried in the href and method attributes of the lt xsl result document gt instruction My tentative syntax here uses a query URL that addresses the current document that is the contents of the browser window select means the output should be located at the same point in the tree and method replace content means that the existing content of the current element but not the element itself should be written to this location Is this kind of in situ update consistent with the XSLT processing model Is it safe in a declarative language where order of execution is not defined I believe it should be possible to define the semantics rigorously but I don t claim to have done so yet I think the formal processing model will turn out to be rather like the model for XQuery Updates all the result documents generated during a transformation phase are accumulated in a pending update list they are not allowed to conflict with each other and they are then applied together at the end of this processing phase the processing phase being all th
301. ween the requirements of the XPath specification and the behaviour of the available Javascript library such as the exact rules for rounding in double arithmetic the choice is between using the library and hoping users don t mind too much or writing thousands of lines of code that do things slightly differently and in practice users are no more likely to be affected by such departures from the spec than they are to be affected by the many areas where the specification leaves matters completely implementation defined Second it was of course necessary to make the code compile GWT doesn t support every single interface in the JDK Coverage is good but not 100 Some of the omissions are for obvious reasons for example there is no file I O Java s standard XML interfaces such as SAX and DOM are absent replaced by DOM like APIs that wrap the XML DOM and HTML DOM Javascript APIs The standard Java regex handling is absent replaced by a different API that wraps the Javascript regex library Some of the omissions seem to have no particular rationale there s no StringTokenizer no java net URI and no java util BitSet In all cases it was easy enough to fill the gap by writing new classes that do what Saxon requires in some cases such as BitSet and URI this was a tiny fraction of what the JDK class offers Finally itis necessary to invoke the XSLT processor within the browser environ ment and for the XSLT processor to be able to get i
302. with a strong preference for Web based integ ration The umbrella project is a Web framework of which a key component is 54 Akara Spicy Bean Fritters and XML Data Services Amara 2 a port of the well known 4Suite and Amara XML processing components for Python Amara 2 provides optimized XML processing using common XML standards as well as fresh ideas for expressing XML pattern processing based on long experience in standards based XML applications Some of these features include e XPath e XSLT e lightweight dynamic data binding e XML modeling and processing constraints by example using Examplotron e Schematron assertions e XPath driven streamable processing e low level support for lazy iterator processing and thus the map reduce style Amara 2 x is designed from the ground up for the architectural benefits discussed above It treats data as much as possible in the data domain rather than in the code domain In practice one still needs good code interfaces but the key balance to strike is in the nature of the resulting code basic planks of the design principles are e syncretism combining the practical expressiveness of Python with the declar ative purity of XML e itis very difficult to balancing such divergent technologies e less code support compact code so there s less to maintain e grace making it easy to do the right thing in terms of standards and practices encourage sound design and modeling using
303. x means there is no prefix value This will only happen for attribute names that do not have a namespace or when representing a default namespace declaration e An empty string for a localName only occurs when the interface is used for namespace declarations e An empty string for a namespaceURI only occurs when the name has no URI as sociated with it This may occur for attributes that do not have a namespace or for setting the default namespace to the empty string 3 Usage Examples The use of the XMLReader interface is much like the usage of XMLHttpRequest A script instantiates the object sets any request properties and then calls send and open in sequence When the response entity body is read XMLItemEvent instances are sent to any event handler registered for XML events For example here s a simple request collects all links from an XHTML document as a filter var reader new XMLReader var links reader onxml function e if e itemType XMLItemEv nt START ELEMENT amp amp e localName a var href e getAttributeValue href if href links push href 140 Efficient XML Processing in Browsers l reader open GET http www milowski com reader send From a completeness perspective the events can be serialized to an XML document using this event handler var xml reader onxml function e switch e itemType case XMLItemEvent START ELEMENT
304. y Insert Into 1 the candidates are street and flat Then we apply insert after algorithm for each child From the previous example street and flat candidates could follow the street element 5 Implementation and Analysis We verified algorithms by implementing the editing functionality as part of this project The implementation follows the theoretical algorithm and is easy under standable We could verify that our methodology is complete enough to cover most of W3C XML Schema structures and provides sufficient performance rates The Figure 4 illustrates an example from the previous section We ran three sets of experiments in the first we used the above simple example to evaluate the validation process for the second set we used documents that cover popular W3C XML Schema constructions to evaluate the editing process All tests were run on Gecko 1 9 Mozilla Firefox 3 6 WebKit 534 Google Chrome 7 and Trident VI Internet Explorer 8 layout engines using an Intel Core 2 Duo 2 0GHz machine with 4G of RAM and Microsoft Windows 7 OS The Linux environment was omitted because the workability depends on engine Web browser but not on OS Both the validation and editing tasks for all tests were performed without 10 Client side XML Schema validation XML Information XML URL http xml waw lv example xml Schema URL http xml waw lv example xsd Parameters Reflected XML Document XML Schema Result BP Expand Collapse in
305. y repository that does not use the filesystem Thanks to the open source implementation in Java of the Packaging System and its repository layout Servlex can look instead in the classpath Instead of using a directory on the filesystem it uses the name of a Java package For instance let us say we have a JAR file with a Java package org example repo and within this Java package sub packages and resources follow the same structure as an on disk repository but instead in the classpath We can then use the name of this Java package to configure the repository of Servlex instead of using a directory on the disk This is particularly interesting to deploy Servlex in disk less environments like Google Appengine and Amazon Cloud EC2 Of course a repository in the classpath is read only so you cannot install and remove webapps on the move this is fixed at Servlex deployement 43 CXAN a case study for Servlex an XML web framework Now that we have a Servlex instance up and running let us have a look at a real sample of webapp Servlex distribution comes with a simple example called hello world The source and the compiled package are both included in the distribution All you need to do in order to install the sample webapp is to go the Servlex Manager at http localhost 8080 servlex manager select the file hello world 0 4 0 xaw from the distribution this is the webapp package and press the deploy button The package is read by Servlex
306. zero and the fn count function applied to an arbitrary expression The rewrite engine automatically handles simple expression eguivalence matters like commutative operators with inverted operands Having found an expression subtree that matches the pattern the rule replaces it with a copy of the result expression Named wildcard expressions in the result expression are called expression references and are replaced with those of the same name matched by the pattern In this way expressions that do not take part in the rewrite can be preserved The special expression reference this is replaced by the entire matched expression The net effect of this example is to replace all occurrences of a fn count com pared to zero with a more efficient function call to fn empty 2 3 Rule Predicates It is often necessary to check conditions beyond a simple expression pattern before deciding that a particular rewrite is valid For this reason rewrite rules allow an arbitrary predicate after the pattern denoted by the where keyword rw BooleanIfElseTrue if condition then then else true where rw subtype then xs boolean gt not condition or then 213 Declarative XQuery Rewrites for Profit or Pleasure This example finds simple conditional expressions and reduces them to boolean logic A check is made on the inferred type of the then expression to ensure that it produces a boolean result and is therefore eligible to be c

XML Prague 2011 Conference Proceedings

Contents

Download Pdf Manuals

Related Search

Related Contents