Re: [ANN] Beta test of the W3C Markup Validator (0.8.0 beta 1)

Hi Olivier!

Am 25.04.2007 um 16:41 schrieb olivier Thereaux:

> One is the question of accept headers. On that question I agree in  
> principle that the validator could send accept headers - you'll  
> have noticed that if you read the thread and bugzilla page. Read,  
> however, David's message, who makes an apt *technical* case against  
> it.
> The other is the question of (conditionally) serving XHTML 1.1 as  
> text/html. The spec says SHOULD NOT. I say "well, if the spec says  
> SHOULD NOT, then you should not" and I add that I don't see a good  
> reason to do it. The validator issues a warning when you do. That's  
> the proper behaviour for a SHOULD NOT.
> So... where exactly is the problem?
> If you want to do something a spec says you SHOULD NOT do, because  
> you think there's a good enough reason, it's your choice. But don't  
> come complaining if a conformance checker catches you doing it, and  
> issues a warning about it.

Olivier, I have no problem with the fact, that the validator throws a  
warning message instead of an error message.
What I have a problem with, is the fact, that the validator doesn't  
send an Accept header. If it would, my pages would have been send  
correctly as "application/xhtml+xml", and the validator wouldn't be  
forced to throw this warning message, which is so far untrue and  
avoidable, because my server is configured to send the correct  
mimetype, if the accept header of the requesting client states, that  
it would accept it.

I further wonder, why the current validator 0.7.4 validates my pages  
without this warning, see
(I know about the BOM warning, and I am still unsure, wether to live  
with it or to avoid it). In contrast, validator 0.8. beta throws a  
warning about an assumed uncorrect Mimetype, although it could be  
served the right way (regulary *is* served the right way), if the  
server would *know*, that the requesting validator is accepting  
Once again, I presume the validator to be a normal requesting client  
to any webserver. Following that assumption, there seems to be not  
one argument, to refuse sending a correct Accept header.


Sierk Bornemann

Received on Wednesday, 25 April 2007 16:09:25 UTC