Re: validator not doing application/xhtml+xml

[ BTW, mimasa, Iım looking at this squarely from the Validator ]
[ POV.  The RFC in question may well solve all problems in the ]
[ general case, itıs just that we have some odd things to take ]
[ into account where the Validator is concerned.  Lots of old  ]
[ baggage in the implementation not least of all! :-(          ]

Masayasu Ishikawa <mimasa@w3.org> wrote:

>An important information for the validator is that the body of a MIME
>entity sent as 'application/xhtml+xml' is syntactically XML.  That is,
>the validator can switch to the "XML mode" without sniffing the actual
>content.  That's a big difference with 'text/html'.

Itıs a big difference from Œtext/htmlı, but is it usefull?

Ok, so weıre now to the point where what we get is known to be generic XML.
Now what? Can I go ahead and assume SGML-type semantics for the XML
Application in question? Will it have a nice easy flattended DTD I can
expect SP to handle? Or does this particular brand of application/xhtml+xml
require XML Schema Validation? Namespaces? Do I need something that groks
M12N? What are the Character Encoding semantics? What are the higher-level
semantics so I can implement pretty-but-not-formal features (aka.
³linting²)?

Are we even at a point were I would avoid all these problems by using a
real Validating XML Processor instead of the half-baked hack that SP is in
relation to XML? Because as a practical matter, real XML Validation is
beyond us at the moment so SP-based hacks are what we can do so far. Should
we not even bother trying and instead pour all effort into getting real XML
Validation done?

³supporting application/xhtml+xml² is a little more involved then just
adding the Content-Type to the configuration file. :-(


>There is an extensive discussion about media types on the Technical
>Architecture Group [2], and there is a proposal to use a combination
>of the Content-Type and Content-Features headers.  See related thread
>on www-html [3] for details.

I saw it -- I /think/ Iım still getting all messages to www-tag -- but
havenıt had a chance to read it all yet. I also havenıt had a chance to
decide whether or not to get my hair zinged by venturing an opinion on the
topic. :-)

Thanks for the pointers. Iıll try to find time to read up on it some time
soon. With any luck itıll address my issues where application/xhtml+xml are
concerned.


>In the absense of a DOCTYPE declaration, the validator may only perform
>well-formedness check, just like it does for XML documents sent as
>'text/xml' or 'application/xml' at the moment.

Yes, we can implement that much right now, but Iım worried that
application/xhtml+xml will need to cater to the same crowd that makes
validating text/html such a joy. IOW that itıll need to be pragmatic rather
then formal and strict in some key aspects. Doing just WF checking and then
suddenly switching to full blown Validation is a sure way to get the
Besserwissers to come crawling out of the woodwork (and Iım not even sure I
blame them).

Received on Tuesday, 12 February 2002 20:53:27 UTC