- From: Chris Lilley <chris@w3.org>
- Date: Tue, 18 Mar 2003 21:06:23 +0100
- To: "Williams, Stuart" <skw@hplb.hpl.hp.com>
- CC: "'www-tag@w3.org'" <www-tag@w3.org>
On Tuesday, March 18, 2003, 1:22:52 PM, Stuart wrote: WS> Just wondering if the topic of mixedUIXMLNamespace-33[1] is WS> related to XHTML modularisation? It is related in being one, but not the only, example of where a finding would help. XHTML Modularisation is just a way to write complex, opaque, DTDs spread across multiple files that take significant time to parse ;) as is SVG Modularisation which has the same drawbacks. However, it does allow a careful human to write a driver DTD that can do DTD validation on a mixed-namespace SVG and XHTML document, given appropriate namespace prefix declarations as parameter entities in the internal DTD subset. The actual 'what does it mean' and 'how do I render this' questions are not addressed explicitly in this modularisation approach but are addressed implicitly in the design choices of the HTML and SVG WGs when creating their modules. Its entirely possible for someone to use the modularisdation approach to modularise or re-modularise a specification such that it is entirely conformant to the modularisation specification, instances can be made to validate against it, and the result has zero meaning or contradicts chuncks of human readable prose in the relevant spec. Its also possible to take the existing modules and compose them in meaningless ways. In other words, modularisation is an abstract tool for specification writers. The actual modularised XHTML 1.1 and SVG 1.1 specifications are concrete tools, but still for specification writers. The issue for content creators, of 'how do I make up a little XML grammar that suits my needs and have it render' is not addressed by modularisation nor is the issue of 'how do I get the links to work, and include graphics and equations and ruby and ....'. So I see this mixedUIXMLNamespaces-33 issue as being more aimed at (serious) content creators and at creators of other, UI-oriented (in the sense of directly renderable, document-like, namespaces) as to how to combine the concrete tools that we provide with their own vertical markup and have the whole thing do something more interesting than emit 'yes' or 'no'. -- Chris mailto:chris@w3.org
Received on Tuesday, 18 March 2003 15:06:28 UTC