xml-dev - Re: XHTML 1.0 returned to HTML WG

["Thomas B. Passin" <tpassin@mitretek.org>]

From: Paul Prescod


>David Megginson wrote:
>>
>>   The XML Namespace for HTML is http://www.w3.org/1999/11/HTML/ and
>>   all names are lower case.
>>
>> (or whatever suits).  Really -- just give us a Namespace, and we'll
>> do stuff you barely dare to dream about.  That way, they can work on
>> XHTML itself, modules, and all that at leisure.
>
>Sure. Without a conformance specification, they could add in features
>like BLINK, LOOP and SHIFT. I don't see how having dozens of groups
>using HTML incompatibly is "progress". After all, we've been doing it
>for five years now WITHOUT an HTML namespace.
>
>Let me be concrete. Let's say you're developing a whizzy app to display
>HTMLish documents. Now half-way through a document you run into:
>
><html:table><html:title/></html:table>
>
>And thanks to the one line spec, this is now legal. What the hell do you
>do? How does one write whizzy apps to work with the HTML namespace? And
>if it isn't possible, then WHAT IS THE POINT OF THE NAMPESPACE?? If you
>want to use HTML elements in your specs, DO IT...name elements in your
>domain the same as HTML's. You can tell them that you have my full
>blessing to do so. If you want to define a useful HTML namespace with
>conformance requirements, I encourage you to do that also.
>
> Paul Prescod
>


It seems to me that using one of the (3 - or is it to be 1?) xhtml namespaces
could basically have three reasons.  The other posts are of course talking about
them, but I thought I'd just condense them together in one place.

1) To inform the receiving html processor which variety of html is in the
document. We already have a declaration for that.  As other posts have
mentioned, it is often ignored by existing browsers.  No namespace is required,
we just output the right declaration as Rev. Bob is doing.

2) We want to tell our own processor, the one that will produce the document,
which style of html to output.  Of course, there are zillions of ways to do
this.  The hard-core xml-ish way would presumably be to use an xsl or xslt
stylesheet to create the xhtml output.  Here, it would make good sense to use
different namespaces to inform the processor what to output.  But observe - we
would use the namespace in the stylesheet, not in the source!  If we want to
produce another flavor of xhtml output, we use a stylesheet that declares a
different namespace.  This is a classic use for a namespace.  But the namespace
does not need to appear in the output, only in the stylesheet!  The output could
include the html flavor declaration if desired.

3) To inform some (non-browser) receiving processor which flavor of html is in
the document.  In other words, if I want to do some xml-like activity - like
extracting data from the xhtml document - by parsing the xhtml, constructing a
tree, using XPATH or DOM, etc., I might need to know this information.  Now here
is the main reason we might want to use xhtml, I think.  It lets us use xml
tools to get at the parts of the html document for other proposes than browsing.
It seems like the existing html declaration would be adequate for this too, just
as it is for case 1).

So neither for purpose 1), 2), or 3)  is there any reason to include the
namespace in the output document.  The only time you would need a namespace in
the result document would be if you wanted to mix and match xhtml styles within
the same document.  The namespace would let you handle that.  I suggest that
this is bogus, there is no practical need to allow for this mixing and matching.

On the other hand, it seems like there could be valid reasons for wanting to
declare appropriate DOCTYPEs in the xhtml document.

Conclusions:  there is no value in including an xhtml namespace - as opposed to
a DOCTYPE declaration - in the actual xhtml document. There is potential value
in defining these namespaces for use by an xslt processor, just as xslt already
has some escape mechanisms that let it output html rather than only strict xml.
More than one DOCTYPEs could be useful for xhtml.

How am I doing here? is there something significant that I missed?

Tom Passin


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev@ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo@ic.ac.uk the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo@ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa@ic.ac.uk)