Discussion of Alternative Schema Languages and Type System Support in WSDL 2.0 http://www.w3.org/TR/2005/NOTE-wsdl20-altschemalangs-20050817 W3C Working Group Note 17 August 2005 http://www.w3.org/TR/2005/NOTE-wsdl20-altschemalangs-20050817 PDF PostScript XML plain text http://www.w3.org/TR/wsdl20-altschemalangs Amelia A. Lewis TIBCO Software, Inc. Bijan Parsia Maryland Information and Network Dynamics Laboratory at the University of Maryland at College Park

This document captures the result of discussions by the Web Services Description Working Group regarding WSDL 2.0 type system extensibilty at the time of its publication. The Working Group normatively defines the use of XML Schema 1.0 as a type system in the WSDL 2.0 Core specification. This document sketches out the basics of extensions for Document Type Definitions (DTDs) and Relax NG.

This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.

This document is a W3C Working Group Note. It has been produced by the Web Services Description Working Group, which is part of the Web Services Activity.

The material in this note was previously published as an Appendix of the Web Services Description Language (WSDL) 2.0: Core Language Last Call specification. In response to Last Call comments, the Working Group agreed to remove this material from that specification and publish it separately as a Working Group Note. Current versions of WSDL 2.0 Core no longer contain this material. This publication differs from the previous material in that it also includes some expanded discussion of issues that should be given consideration by type system extension designers.

No further work on this topic is planned at this point. Errors in this document can be reported to the public public-ws-desc-comments@w3.org mailing list (public archive).

Publication as a Working Group Note does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.

 English

Last Modified copy $Date: 2005/08/17 08:40:53 $
Introduction

WSDL 2.0: Core Language describes Web Service interaction in terms of exchanges of typed messages. WSDL only provides general support for type systems based on the XML Infoset and specific support for the W3C XML Schema Description Language . Describing messages with WSDL using schema languages other than XML Schema or non-XML Infoset type systems requires extending the WSDL component model. While the Web Services Description Working Group has not defined any such extensions, there were discussions in the Working Group about how those extensions might be defined and used. This document is the result of those disucussions and captures part of the Working Group's thinking about schema language and type system extensibilty at the time of its publication.

 Issues facing multiple schema languages/type systems

Without the use of an extension, a WSDL document can only use a single type system, XML Schema. If extensions are defined to support alternative schema languages or non-XML type systems, then issues regarding the mixing of type systems in a single document arise. Part 1 does not define the behavior of mixed type system documents, so it is incumbent on extension authors to do so.

For example, suppose a WSDL author used a extension supporting Relax NG alongside the native support for XML Schema. Further suppose that there is an element component which has a definition in both the referenced XML Schema and Relax NG schema. There are several possibilities for interpreting such a document:

Multiple definitions in distinct type systems is always an error

Multiple definitions must be in some sense equivalent, for example, if XML Schema type and an Relax NG production validate exactly the same set of Infoset fragments, otherwise, an error

Multiple definitions are legal, and are interpreted as a union type constraint

The last interpretation suggests a further general possibility: being able to define a union type (or other compound type) that spans distinct type systems (and, to further generalize, where the unioned types had distinct QName identifiers). The Data Access Working Group had a use case wherein their return message could either be in RDF/XML, which cannot have an interesting XML Schema but does have an interesting Relax NG schema, and their other results format, which they would prefer to specify with an XML Schema. This example is little artificial, as the Data Access Working Group could easily describe the entire results format in Relax NG.

The first interpretation is most in the spirit of WSDL and was strongly preferred by the Working Group. Since WSDL extensibility points are generally quite unrestricted, the Working Group did not try to enforce the first option, but the general belief of the Working Group was that the other options were confusing and unwise.

 Examples of Specifications of Extension Elements for Alternative Schema Language Support

This section contains two examples of specifications of extension elements for alternative schema language support. Please note that those examples did not receive any implementation testing.

DTD

A Document Type Definition (DTD) as defined in may be used as the schema language for WSDL. It may not be embedded; it must be imported. A namespace must be assigned. DTD types appear in the element declarations property of the Description component and may be referenced from the wsdl:input, wsdl:output and wsdl:fault elements using the element attribute information item.

The prefix, dtd, used throughout the following is mapped to the namespace URI http://www.w3.org/2005/08/wsdl/dtd-import.

The dtd:import element information item references an external Document Type Definition, and has the following Infoset properties:

A [local name] of import.

A [namespace name] of "http://www.w3.org/2005/08/wsdl/dtd-import".

One or two attribute information items, as follows:

A REQUIRED namespace attribute information item as described below.

An OPTIONAL location attribute information item as described below.

 namespace attribute information item

The namespace attribute information item sets the namespace to be used with all imported element definitions described in the DTD. It has the following Infoset properties:

A [local name] of namespace.

A [namespace name] which has no value.

The type of the namespace attribute information item is xs:anyURI.

The WSDL author should ensure that a prefix is associated with the namespace at the proper scope (probably document scope).

 location attribute information item

The location attribute information item, if present, provides a hint to the processor as to where the DTD may be located. Caching and cataloging technologies may provide better information than this hint. The location attribute information item has the following Infoset properties:

A [local name] of location.

A [namespace name] which has no value.

The type of the location attribute information item is xs:anyURI.

 References to Element Definitions

The element attribute information item MUST be used when referring to an element definition (<!ELEMENT>) from a Interface Message Reference component; referring to an element definition from a Interface Fault component is similar. The value of the element definition MUST correspond to the content of the namespace attribute information item of the dtd:import element information item. The local name part must correspond to an element defined in the DTD.

Note that this pattern does not attempt to make DTDs namespace-aware. It applies namespaces externally, in the import phase.

 RELAX NG

A RELAX NG schema may be used as the schema language for WSDL. It may be embedded or imported; import is preferred. A namespace must be specified; if an imported schema specifies one, then the [actual value] of the namespace attribute information item in the import element information item must match the specified namespace. RELAX NG provides both type definitions and element declarations, the latter appears in the element declarations property of Description component respectively. The following discussion supplies the prefix rng which is mapped to the URI "http://www.relaxng.org/ns/structure/1.0".

Importing RELAX NG

Importing a RELAX NG schema uses the rng:include mechanism defined by RNG, with restrictions on its syntax and semantics. A child element information item of the types element information item is defined with the Infoset properties as follows:

A [local name] of include.

A [namespace name] of "http://www.relaxng.org/ns/structure/1.0".

Two attribute information items as follows:

A REQUIRED ns attribute information item as described below.

An OPTIONAL href attribute information item as described below.

Additional attribute information items as defined by the RNG specification.

Note that WSDL restricts the rng:include element information item to be empty. That is, it cannot redefine rng:start and rng:define element information items; it may be used solely to import a schema.

ns attribute information item

The ns attribute information item defines the namespace of the type and element definitions imported from the referenced schema. If the referenced schema contains an ns attribute information item on its grammar element information item, then the values of these two attribute information items must be identical. If the imported grammar does not have an ns attribute information item then the namespace specified here is applied to all components of the schema as if it did contain such an attribute information item. The ns attribute information item contains the following Infoset properties:

A [local name] of ns.

A [namespace name] which has no value.

The type of the ns attribute information item is xs:anyURI.

 href attribute information item

The href attribute information item must be present, according to the rules of the RNG specification. However, WSDL allows it to be empty, and considers it only a hint. Caching and cataloging technologies may provide better information that this hint. The href attribute information item has the following Infoset properties:

A [local name] of href.

A [namespace name] which has no value.

The type of the href attribute information item is xs:anyURI.

 Embedding RELAX NG

Embedding an RNG schema uses the existing top-level rng:grammar element information item. It may be viewed as simply cutting and pasting an existing, stand-alone schema to a location inside the wsdl:types element information item. The rng:grammar element information item has the following Infoset properties:

A [local name] of grammar.

A [namespace name] of "http://www.relaxng.org/ns/structure/1.0".

A REQUIRED ns attribute information items as described below.

Additional attribute information items as specified for the rng:grammar element information item in the RNG specification.

Child element information items as specified for the rng:grammar element information item in the RNG specification.

 ns attribute information item

The ns attribute information item defines the namespace of the type and element definitions embedded in this schema. WSDL modifies the RNG definition of the rng:grammar element information item to make this attribute information item required. The ns attribute information item has the following Infoset properties:

A [local name] of ns.

A [namespace name] which has no value.

The type of the ns attribute information item is xs:anyURI.

 References to Element Declarations

Whether embedded or imported, the element definitions present in a schema may be referenced from a Interface Message Reference or Interface Fault component.

A named rng:define definition MUST NOT be referenced from the Interface Message Reference or Interface Fault components.

A named Relax NG element declaration MAY be referenced from a Interface Message Reference or Interface Fault component. The QName is constructed from the namespace (ns attribute information item) of the schema and the content of the name attribute information item of the element element information item An element attribute information item MUST NOT be used to refer to an rng:define element information item.

References RELAX NG Specification, J. Clark, MURATA Makoto, editors. OASIS Committee Specification, 3 December 2001. Web Services Description Language (WSDL) Version 2.0 Part 1: Core Language, R. Chinnici, J-J. Moreau, A. Ryman, S. Weerawarana, Editors. World Wide Web Consortium, 3 August 2005. This version of the Web Services Description Language Version 2.0 Core specification is http://www.w3.org/TR/2005/WD-wsdl20-20050803. The latest version of "Web Services Description Language (WSDL) Version 2.0 Part 1: Core Language" is available at http://www.w3.org/TR/wsdl20. Extensible Markup Language (XML) 1.0 (Third Edition), T. Bray, J. Paoli, C. M. Sperberg-McQueen, E. Maler, and F. Yergeau, Editors. World Wide Web Consortium, 4 February 2004. This version of the XML 1.0 Recommendation is http://www.w3.org/TR/2004/REC-xml-20040204/. The latest version of "Extensible Markup Language (XML) 1.0" is available at http://www.w3.org/TR/REC-xml. XML Information Set (Second Edition), J. Cowan and R. Tobin, Editors. World Wide Web Consortium, 4 February 2004. This version of the XML Information Set Recommendation is http://www.w3.org/TR/2004/REC-xml-infoset-20040204. The latest version of "XML Information Set" is available at http://www.w3.org/TR/xml-infoset. XML Schema Part 1: Structures (Second Edition), H. Thompson, D. Beech, M. Maloney, and N. Mendelsohn, Editors. World Wide Web Consortium, 28 October 2004. This version of the XML Schema Part 1 Recommendation is http://www.w3.org/TR/2004/REC-xmlschema-1-20041028. The latest version of "XML Schema Part 1: Structures" is available at http://www.w3.org/TR/xmlschema-1. Acknowledgements

This document is the work of the W3C Web Service Description Working Group.

Members of the Working Group are (at the time of writing, and by alphabetical order): Allen Brookes (Rogue Wave Softwave), Dave Chappell (Sonic Software), Helen Chen (Agfa-Gevaert N. V.), Roberto Chinnici (Sun Microsystems), Kendall Clark (University of Maryland), Ugo Corda (SeeBeyond), Glen Daniels (Sonic Software), Paul Downey (British Telecommunications), Youenn Fablet (Canon), Hugo Haas (W3C), Tom Jordahl (Macromedia), Anish Karmarkar (Oracle Corporation), Jacek Kopecky (DERI Innsbruck at the Leopold-Franzens-Universität Innsbruck, Austria), Amelia Lewis (TIBCO Software, Inc.), Michael Liddy (Education.au Ltd.), Kevin Canyang Liu (SAP AG), Jonathan Marsh (Microsoft Corporation), Josephine Micallef (SAIC - Telcordia Technologies), Jeff Mischkinsky (Oracle Corporation), Dale Moberg (Cyclone Commerce), Jean-Jacques Moreau (Canon), Mark Nottingham (BEA Systems, Inc.), David Orchard (BEA Systems, Inc.), Bijan Parsia (University of Maryland), Tony Rogers (Computer Associates), Arthur Ryman (IBM), Adi Sakala (IONA Technologies), Asir Vedamuthu (Microsoft Corporation), Sanjiva Weerawarana (Independent), Ümit Yalçınalp (SAP AG).

Previous members were: Lily Liu (webMethods, Inc.), Don Wright (Lexmark), Joyce Yang (Oracle Corporation), Daniel Schutzer (Citigroup), Dave Solo (Citigroup), Stefano Pogliani (Sun Microsystems), William Stumbo (Xerox), Stephen White (SeeBeyond), Barbara Zengler (DaimlerChrysler Research and Technology), Tim Finin (University of Maryland), Laurent De Teneuille (L'Echangeur), Johan Pauhlsson (L'Echangeur), Mark Jones (AT&T), Steve Lind (AT&T), Sandra Swearingen (U.S. Department of Defense, U.S. Air Force), Philippe Le Hégaret (W3C), Jim Hendler (University of Maryland), Dietmar Gaertner (Software AG), Michael Champion (Software AG), Don Mullen (TIBCO Software, Inc.), Steve Graham (Global Grid Forum), Steve Tuecke (Global Grid Forum), Michael Mahan (Nokia), Bryan Thompson (Hicks & Associates), Ingo Melzer (DaimlerChrysler Research and Technology), Sandeep Kumar (Cisco Systems), Alan Davies (SeeBeyond), Jacek Kopecky (Systinet), Mike Ballantyne (Electronic Data Systems), Mike Davoren (W. W. Grainger), Dan Kulp (IONA Technologies), Mike McHugh (W. W. Grainger), Michael Mealling (Verisign), Waqar Sadiq (Electronic Data Systems), Yaron Goland (BEA Systems, Inc.), Ümit Yalçınalp (Oracle Corporation), Peter Madziak (Agfa-Gevaert N. V.), Jeffrey Schlimmer (Microsoft Corporation), Hao He (The Thomson Corporation), Erik Ackerman (Lexmark), Jerry Thrasher (Lexmark), Prasad Yendluri (webMethods, Inc.), William Vambenepe (Hewlett-Packard Company), David Booth (W3C), Sanjiva Weerawarana (IBM), Charlton Barreto (webMethods, Inc.), Asir Vedamuthu (webMethods, Inc.), Igor Sedukhin (Computer Associates), Martin Gudgin (Microsoft Corporation).

The people who have contributed to discussions on www-ws-desc@w3.org are also gratefully acknowledged.