W3C home > Mailing lists > Public > www-html-editor@w3.org > April to June 2008

Re: please add Accept header to http request containing application/xhtml+xml

From: Olivier Thereaux <ot@w3.org>
Date: Sun, 20 Apr 2008 19:54:48 -0400
To: Noa Resare <noa@resare.com>
Cc: www-validator@w3.org, www-html-editor@w3.org
Message-ID: <20080420235448.GB11032@w3.mag.keio.ac.jp>

[copying www-html-editor for the bug report about XHTML 1.0, section 5 ]

On Fri, Apr 18, 2008, Noa Resare wrote:
> The recommendation to use mime type application/xhtml+xml has the  
> problem of not being compatible with for example Internet Explorer 6  
> (which has quite a market share)
> There are various recommendations about how to work around that at the  
> httpd level by adding special rewrite rules handling internet explorer  
> as a one off case. However, a better solution IMHO would be to use the  
> http request Accept header, looking for the application/xhtml+xml  
> string.
> This strategy works beautifully with firefox, safarai and Internet  
> Explorer, but the w3 validator unfortunately doesn't send any Accept  
> headers, which makes a apache httpd server with the rewrite rules  
> specified below outputting a warning that the server does send text/ 
> html instead of application/xhtml+xml

At this point in time the XHTML working group is considering allowing
sending any of the XHTML family as text/html, so the validator (in dev
for now) will not output that warning.

It's important to understand that when served text/html, browsers will
treat content as HTML. In other words, this validator is among the very
rare user agents which bothers to look at whether the doctype is for
XHTML. Serving XHTML as text/html is really akin to serving tag soup.

Now here comes the dilemma:

- most people seem to hate getting warnings from the validators. I find
  that terribly disturbing, since warnings are meant to say "please
double check this as it might be an issue, might not" and the reactions
to these warnings show a very wrong "religious" attitude towards the
validation results. It certainly makes it difficult for us to add to the
validator features which would help improve web quality, even if they
are not normative requirements.

- Steven Pemberton, staff contact for the XHTML WG, recently told me in
a discussion that the HTML compatibility guidelines were meant to be
informative. I noted that, although informative, that section was
refered to in normative parts of the XHTML 1.0 specification 
-> see http://www.w3.org/TR/xhtml1/#issues
(this is a bug in the spec, IMHO. Normative and informative should not
be mixed like this.)
Anyway,  if the guidelines are indeed meant to be informative, it will be
harder to add them to the validator, for the reason detailed above.

Food for thought, I believe. Until then I'm sure a lot of people will be
happy not to receive the warning for XHTML 1.1 as text/html...

Received on Sunday, 20 April 2008 23:55:27 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:39:50 UTC