W3C home > Mailing lists > Public > public-html@w3.org > April 2010

RE: ISSUE-41: Facebook open graph protocol

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Fri, 23 Apr 2010 01:45:56 +0200
To: sroussey@network54.com
Cc: 'Philippe Le Hegaret' <plh@w3.org>, HTMLwg <public-html@w3.org>
Message-ID: <20100423014556586796.6ecea770@xn--mlform-iua.no>
Steve Roussey, Thu, 22 Apr 2010 13:55:09 -0700:
> - why XHTML11 should not be permitted served as text/html?
> 
> For the next version of Firebug, I've made the doctype visible in the HTML
> view, and mark it red and with an error message (on hover) for sites that do
> the above. Essentially, the browser ignores the XHTML references and obeys
> the mime-type, and web page authors can go on for a while never knowing.
> Microsoft.com, ibm.com, and developer.mozilla.org all now get flagged with
> errors.

Note that I spoke about XHTML 1.1. Some think that 'text/html' should 
be officially permitted for XHTML 1.1. When it comes to those pages you 
mentioned, then

      Developer.mozilla.org is labeled as 'XHTML 1.0'
                    Ibm.com is labeled as 'XHTML 1.0'
              Microsoft.com is labeled as 'XHTML 1.0'
http://www.w3.org/News/2010 is labeled as 'XHTML 1.0'

And all are served as 'text/html', which is permitted, according to the 
current MIME registration for 'text/html'. So no need, according to 
official standards and registrations, to give the impression that the 
very DOCTYPE makes them invalid.

> If it were not for the polyglot spec, I'd prefer to have XHTML docs root
> node be <xhtml> instead of <html> to make things clear. I was always unclear
> on the value of xhtml on the public facing internet, and found it difficult
> to find examples of real xhtml in the wild against which to test svg and
> mathml in firebug.
> 
> If the doctype switched a text/html served file from html to xhtml it would
> likely match average authors' expectations as they are today. I'm sure that
> there is a long history on why this is not so, and the HTML5 doctype no
> longer distinguishes between the two. And that is probably the most
> compatible way forward, as it will change the expectations of what the
> doctype means.

And the average Web author also think that JavaScript is related to 
Java, doesn't he? Judging from all the JavaScript tutorials which tells 
us that it does not, one should think that this was a huge problem ...

It *might* serve point to mark 'XHTML 1.1.' pages as having an error if 
served as 'text/html'. The question is *which* point it serves.
-- 
leif halvard silli
Received on Thursday, 22 April 2010 23:46:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:17:08 GMT