Re: let authors choose text/html or application/xhtml+xml (detailed review of section 1. Introduction)

Dan Connolly wrote:

> XHTML documents served as text/html result in interoperable behavior
> in typical cases, so that constraint is too strong. Please change
> it to "SHOULD be sent..." and "SHOULD NOT be served...".

Yes, but that "interoperable behaviour" is /wrong/ (I've sent
a demo. of this to Richard Ishida [1]).  An XHTML document that
contains, anywhere in the <head> region, a construct of
the form

	<meta ... />

should automatically terminate the head region at that point
if it is being parsed as HTML; the validator explains this
perfectly (if somewhat opaquely, if you don't know what
you're looking for).  Current browsers do /not/ terminate
the <head> region at this point, but they should.

Philip TAYLOR
--------
[1] http://validator.w3.org/check?uri=http://www.rhul.ac.uk/resources/Demo-documents/NET.html

> This is a demo. of the problem that I created
> some time ago, Richard : you will see that
> the validator disallows the ">" of
> "<link rel="whatever" href="#" />"
> as the "/" has already closed the element.
> 
> The ">" therefore triggers "body mode", but
> as it's not wrapped in a "<p>" or similar,
> it triggers an error.
> 
> ** Phil.
> --------
> Richard Ishida wrote:
>> Hi Philip,
>>
>> I guess I'm lacking in understanding on this particular issue.  I'm unaware
>> of the implications.
>>
>> Cheers,
>> RI 

Received on Thursday, 30 August 2007 23:41:27 UTC