- From: Masayasu Ishikawa <mimasa@w3.org>
- Date: Fri, 16 Aug 2002 00:41:33 +0900 (JST)
- To: www-html@w3.org
"Manos Batsis" <m.batsis@bsnet.gr> wrote: > The first question to ask is, what forms of validation will the WG > offer? Good question. Our position is that the Abstract Definition is the normative definition, which is schema-language independent. then it may be implemented in any schema language, and we will provide some implementations. > I am sure that an XML Schema based on [1] will be included > as normative. Yes. > IMHO a Relax NG should also be included; the > language is very popular and this will certainly be a demonstration > of good will from W3C. This is not clear at the moment. We haven't discussed whether we should (normatively or informatively) provide a RELAX NG implementation as an official product yet. I would say it depends on how much demand we would receive. Technically I don't see a problem, given that if we do so most probably I would work on it, and I (personally) already have a prototype of it. If RDF would go for RELAX NG, that would be a good motivation. > I also suppose a DTD will probably be offered as well (based on [2]). Well, officially yes. Personally not so sure. I understand many people still want it and we could offer something, but most probably it cannot fully describe the language and even worse, it would introduce some unreasonable restrictions. Personally I'm not tempted to require DTD validity as part of conformance. > The above three may not provide the same level of validation though... > Has the WG considered this as a problem? My experience with various XHTML modularization implementations is that nothing is perfect to fully describe the language, unfortunately. So far we've required DTD validity for "strictly conforming" XHTML documents, as those are relatively simple and DTD is more mature than other schema languages. But for XHTML2, I think we should seriously reconsider this issue. That's what the "DTD Bias" issue implies. http://www.w3.org/TR/2002/WD-xhtml2-20020805/conformance.html#s_conform_issue_0 > > namespace="##other" with processContents="skip", you mean? > > Yes. Here's something based on [3]: <snip/> > Um... forgive my ignorance but my age puts me in the "new generation" > of authors; I have never used DTDs. Can we offer the above functionality > > using DTDs? Unfortunately, no. DTD is damn too silly on this front. > The above may seem kind of loose; one may argue that I can put any > number of attributes and still be valid. That is the intention, since > the > document is always the subject and each attribute-value pair > corresponds to predicate-object, so we do have correct triples. Meanwhile, if you are interested in it, you may take a look at the proposed new Metainformation Module, currently implemented in RELAX NG at: http://www.w3.org/People/mimasa/test/schemas/rng/xhtml-metadata-2.rng As I mentioned in other message, it's a small change and doesn't go very far, but at least it allows you to specify subject-property- object triple rather than key-value pair. > Finally, will the WG consider to document the use of the link element > for metadata purposes? > > <link rel="meta" href="someName.rdf" /> Since it's already documented in RDF Model and Syntax REC, nothing prevents you to use it, just like P3P's usage of the link element. Anyway, I see no problem to mention it if people find it useful. > [1] http://www.w3.org/TR/xhtml-m12n-schema/ > [2] http://www.w3.org/TR/xhtml-modularization/ > [3] > http://www.w3.org/TR/xhtml-m12n-schema/schema_module_defs.html#a_smodule > _Metainformation In another message, you wrote: > I also like Patrick Stickler's approach at [1] (actually that's more > like my initial thought about introducing a section in the head element; Unfortunately, as I said in other message, it just doesn't work in DTD. ANY in DTD is not as smart as ##any in XML Schema. Regards, -- Masayasu Ishikawa / mimasa@w3.org W3C - World Wide Web Consortium
Received on Thursday, 15 August 2002 11:41:35 UTC