Using UML to Define XML DTDs

Date:     Sat, 15 Jan 2000 11:28:34 -0600
From:     "W. Eliot Kimber" <eliot@isogen.com>
To:       xml-dev@ic.ac.uk
Subject:  Draft of Paper: Using UML to Define XML DTDs

I've posted to my Web site a draft of a paper I'm developing on a technique for using UML to define XML DTDs. I'd be grateful for any review and feedback that y'all might be willing to provide. The URL is http://www.drmacro.com/hyprlink/uml-dtds.pdf (about 100K in size).

I think the technique has some very attractive features compared with, for example, using an XML-schema type approach for development and management of DTDs (although it is not necessarily a competitor with XML Schema as XML Schemas are one possible result that can be generated from the UML models).

The basic idea is quite simple: I defined a set of types that reflect the syntactic constructs of XML needed to define document types (element types, attributes, etc.) and then use those types as stereotypes in models that represent DTDs as implementation models (in the UML sense of implementation models). The implementation model is intended to be translated directly into some DTD syntax (normal declarations, XML Schema, XDR, etc.) and so is analogous to using UML to define object models that are translated directly into code. Because the models are defined in UML you get nice graphical representations for free, you can tighly bind documentation to the model, and you can formally relate the implementation model to higher-level analysis models that the DTD is an implementation of, and of course you can use existing UML-base CASE tools to do the development work. You can also use XMI as the normative representation format (which allows, for example, quite sophisticated documentation structures to be included in the model). You also get true modularization and name mapping pretty much for free using UML's package facilities.

I'm pretty happy with the design at this point, although I haven't applied it to enough examples yet to be confident it is complete. The paper does include a tutorial example DTD defined using the approach.

Thanks,

Eliot

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ or CD-ROM/ISBN 981-02-3594-1

Prepared by Robin Cover for the The SGML/XML Web Page archive.

SEARCH Advanced Search ABOUT Site Map CP RSS Channel Contact Us Sponsoring CP About Our Sponsors NEWS Cover Stories Articles & Papers Press Releases CORE STANDARDS XML SGML Schemas XSL/XSLT/XPath XLink XML Query CSS SVG TECHNOLOGY REPORTS XML Applications General Apps Government Apps Academic Apps EVENTS LIBRARY Introductions FAQs Bibliography Technology and Society Semantics Tech Topics Software Related Standards Historic	Using UML to Define XML DTDs Date: Sat, 15 Jan 2000 11:28:34 -0600 From: "W. Eliot Kimber" <eliot@isogen.com> To: xml-dev@ic.ac.uk Subject: Draft of Paper: Using UML to Define XML DTDs I've posted to my Web site a draft of a paper I'm developing on a technique for using UML to define XML DTDs. I'd be grateful for any review and feedback that y'all might be willing to provide. The URL is http://www.drmacro.com/hyprlink/uml-dtds.pdf (about 100K in size). I think the technique has some very attractive features compared with, for example, using an XML-schema type approach for development and management of DTDs (although it is not necessarily a competitor with XML Schema as XML Schemas are one possible result that can be generated from the UML models). The basic idea is quite simple: I defined a set of types that reflect the syntactic constructs of XML needed to define document types (element types, attributes, etc.) and then use those types as stereotypes in models that represent DTDs as implementation models (in the UML sense of implementation models). The implementation model is intended to be translated directly into some DTD syntax (normal declarations, XML Schema, XDR, etc.) and so is analogous to using UML to define object models that are translated directly into code. Because the models are defined in UML you get nice graphical representations for free, you can tighly bind documentation to the model, and you can formally relate the implementation model to higher-level analysis models that the DTD is an implementation of, and of course you can use existing UML-base CASE tools to do the development work. You can also use XMI as the normative representation format (which allows, for example, quite sophisticated documentation structures to be included in the model). You also get true modularization and name mapping pretty much for free using UML's package facilities. I'm pretty happy with the design at this point, although I haven't applied it to enough examples yet to be confident it is complete. The paper does include a tutorial example DTD defined using the approach. Thanks, Eliot xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ or CD-ROM/ISBN 981-02-3594-1 Prepared by Robin Cover for the The SGML/XML Web Page archive.