W3C home > Mailing lists > Public > public-evangelist@w3.org > October 2003

Re: XHTML 1.0 spec (was Re: Call for contributions: new and improved "Web site quality" articles

From: Jim Ley <jim@jibbering.com>
Date: Thu, 2 Oct 2003 18:13:22 -0000
Message-ID: <001901c38910$de061b10$428f9bd9@Snork>
To: <public-evangelist@w3.org>

"Brian Kelly" <b.kelly@ukoln.ac.uk>
> On Thu, 2 Oct 2003, Jim Ley wrote:
> > XHTML is only "allowed" to be served as text/html if it's under Appendix
C,
> > if someone would like to correct me on that and say Appendix C isn't
> > normative and any old XHTML can be served as text/html then I'll
withdraw
> > the criticism.
>
> The document says
> "C. HTML Compatibility Guidelines
> This appendix is informative."
>
> If you think that's incorrect I suggest the editor of the spec should be
> informed (I guess there's a QA of the spec issue).

I agree Appendix C is non-normative, however the MIME-type definition 5.1 is
normative
http://www.w3.org/TR/xhtml1/#media and it says

| XHTML Documents which follow the guidelines set forth in
| Appendix C, "HTML Compatibility Guidelines" may be labeled
| with the Internet Media Type "text/html"

So whilst Appendix C isn't normative in itself (since you can use
application/xml or application/xhtml+xml for your XHTML) if you wish to
serve it as text/html 5.1 requires it to be.  This is certainly been my
interpretation and it's been discussed here and in other places often enough
that I'd've thought someone would correct me on it if it was wrong.

> C.7 says you need lang: and xml:lang attributes.  I have the lang
> attribute in the html element.  I don't need to give the xml:lang
> attribute as C.1 allows me to omit it (as I understand it).

Ah, I understood the xml declaration was the
<?xml version="1.0"?> and that was all you were being allowed to leave out,
it had no bearing on C.7, I could be wrong of course, but again, it's been
often discussed.


> So I think my approach is better than creating HTML 4 resources.

All of your reasons for serving XHTML how are you, I entirely agree with,
and are completely understandable, indeed they're probably required in the
real world.  However you've not given the reason why HTML 4.01 strict is not
appropriate for you, it's semantically identical to XHTML 1.0.  Or what it
is about it that makes you not want to author it (a point that I've not yet
seen made on this list)

> Note that an advantage with text/html is that the page will display if the
> XHTML is invalid.

I'd personally say, this is a huge disadvantage of XHTML as specified, not
an advantage of breaking the rules, it doesn't have a method of degrading,
the don't display anything in error is a key weakness, and a key strength of
HTML 4.01 as deployed.  I don't think pretending XHTML is really HTML
tag-soup and relying on that get us out of trouble if we make a mistake is
to be applauded.   I don't think we should deploy XHTML in any format,
unless we have automated publishing tools to actually ensure it's valid, the
constraint on invalid - no rendering is too severe a failure, in a document
that is mostly correct.

Jim.
Received on Thursday, 2 October 2003 14:15:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 15 July 2011 00:13:22 GMT