Lists Home |
Date Index |
>De : Jonathan Borden [mailto:firstname.lastname@example.org]
>Envoyé : vendredi 18 janvier 2002 01:43
>À : Nicolas LEHUEN; Paul T; Bullard, Claude L (Len); Leigh Dodds;
>Objet : Re: [xml-dev] RDDL (was RE: [xml-dev] Negotiate Out The Noise)
>> ... If I'm supposed to use RDDL to find the proper schemas or
>> stylesheets or whatever, where can I find a RDDL document
>for this XHTML
>> document that contains tags from the MathML namespace ? At the XHTML
>> namespace URL, or at the MathML namespace URL ? It's funny
>to notice that
>> the same problem exist for RDDL document themselves, since
>they mix the
>> (X)HTML, RDDL and XLink namespaces... The problem is that with
>> namespace-centric view of RDDL just prevents document types
>that can mix
>> tags from different namespace from having an associated RDDL
>> doesn't, show me the URL where I can find it), resulting in
>a big hole in
>> the practical usability of RDDL.
>To expand on this. There are at least two ways to mix
>namespaces in schemas.
>The first is along the lines of the RDDL DTD itself which uses XHTML
>Modularization to explicitly include modules for each of the
>the schema/DTD. XML Schema and RELAXNG similarly provide for explicit
>inclusion of namespace qualified elements from different namespaces.
>The second way to mix namespaces is something only coming under
>investigation -- but I think it _is_ something which should be
>where RDDL might help. This is when something like XHTML might act as a
>container for elements in other namespaces, particularly for
>are NOT explicitly described in the container document_. In
>such cases, one
>might reference the RDDL for the _contained_ element namespace
>to discover a
>schema useful for locally validating the namespace qualified
>element and its
>contents. This requires RDDL aware schema validators (such as XSV).
What is understand from this is that RDDL should not be named RDDL (Resource
Directory Description Language) but NDL (Namespace Description Language).
But anyway, I think there are two problems that RDDL try to solve in the
same document and that should be separated : resource linking and
Resource linking is very, very important, but using this feature to justify
the need for RDDL is like telling 'hey, my solution is the best because it
solves a problem'. While I do agree with the requirement for a resource
linking scheme featuring links with different roles (or arcroles or
purpose), I'd like to remind you that RDDL is not the only way to do this,
far from it, RDF being another serious candidate. RDF is not perfect either
for this purpose, I think there should be some distribution of the system
and I suspect that's what Paul is thinking about for the 'next version' to
be built like a DNS.
Human-readable documentation is another important problem, with lots of
specifics consequences, amongst which the need to deal with
internationalisation. Come on, you can't be serious about using xml:lang.
Browsers don't support this, and the whole objective of RDDL is that by
pointing my browser to a RDDL URL, I can see some documentation appear in my
browser, right ? So either you tackle this problem with browser friendly
solutions, or you drop the idea of stuffing ALL resources and documentation
in the same document.