Re: Comments on WD-html-in-xml-19990304

Shane P. McCarron <ahby@themacs.com> writes:
> 
> I just have a few comments on the draft during this last call period:
> 
> SUBSTANTIVE COMMENTS:
> 
> Internet Media Types:
> 
> I think it is very important that the test/xhtml Internet Media Type be
> defined by the W3C and associated with the document types defined in
> this draft.  I think it equally important that the document make it
> clear that this media type is ONLY for use by documents that adhere to
> those DTDs. text/xhtml is a solution for a certain class of documents
> for a few years.  Ultimately we will be able to move to pure XML
> documents that do not rely upon arcane knowledge in browsers in order to
> render documents. For now that is unrealistic.  text/xhtml gives us the
> best of the XML world while briding us to our HTML past.

For the record (without rehashing the many past comments on this issue),
Sun Microsystems opposes creation of a new text/xhtml Internet Media Type
as a mistake.  The web doesn't need a new media type for each markup 
language, the W3C needs to create a generalized solution for XML. The 
text/shoeml argument, reductio ad absurdum.

I believe we need to further coordinate with existing content negotiation
efforts in a suitable solution. I thought we'd begun progress on this
issue, but I've seen nothing posted to the WG mailing list on the topic.
Has the XML CG not provided any comments? Has there not been any discussion?

> Value of the xmlns attribute:
> 
> I believe that it makes sense to have a separate namespace definition
> for each of the defined DTDs. After all, they have different sets of
> legal elements and attributes from one another, and that is what a
> namespace isolates.

And yet any occurrence of an element of blockquote, table, etc. is the
same regardless of DTD. You are conflating namespaces with DTDs, which
are different animals. We consider all three DTDs to be "XHTML", and
semantically, that is the namespace.
 
> FPIs for XHTML:
> 
> I think that the system ID in the defined FPI should be a fully
> qualified URI that points to the definition of the DTD on the W3C
> system.  This can be easily overridden by a local catalog if a developer
> wishes, but for most people the DTD will not be available locally. 
> Also, the W3C should be the definitive repository for these DTDs.

I am not opposed to this (and have stated so both publicly and privately
to Paul Grosso, who provided the spur of this argument), so long as the
W3C makes a policy decision to actually provide such a service, makes
this decision publicly, and provides this WG with three suitable URIs
for the DTD. Until such time, we'd be guessing the URIs, and we'd
also be assuming that they were canonical. This would be a mistaken
assumption without the W3C making this official.

> On a related note, Appendix A should contain an example SGML Open
> Catalog for these DTDs and the Entity Definitions.

I agree with this, and certainly of the rest of Shane's editorial remarks.
I think removing the spaces from " />" in the examples (as suggested by
Shane) would satisfy the last call comment by Bill Smith regarding the
space not being normative.

Murray

...........................................................................
Murray Altheim, SGML Grease Monkey         <mailto:altheim&#64;eng.sun.com>
Member of Technical Staff, Tools Development & Support
Sun Microsystems, 901 San Antonio Rd., UMPK17-102, Palo Alto, CA 94303-4900

              An SGML declaration does not an i18n make.

Received on Wednesday, 24 March 1999 12:45:07 UTC