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

On Tue, Apr 24, 2007 at 12:33:27PM +0200, Sierk Bornemann wrote:
> I want to use XHTML 1.1

Why? It doesn't bring any benefits over HTML 4.01 if you serve it as
text/html.

> , and I want to serve the apropriate MIME type to all browsers,
> which do suffice these standards requirements and who support this
> MIME type. So the only web browser, who is served with "text/html"
> seems to be the Internet Explorer (which has got a lower priority in
> my concerns, but that is another debate).

Last time I checked GoogleBot and Lynx didn't support
application/xhtml+xml either.

>  per default, I serve the .html-Suffix as "text/html",

So, by default, you serve XHTML 1.1 as a mime type which has SHOULD
NOT slapped all over it.

> My question is: why doesn't the validator catch that MIME type, that  
> is served as "text/html" but re-written to "application/xhtml+xml"? I  
> must assume, that the current validator 0.8 beta doesn't send an  
> Accept-Header, so that the Rewrite-Rule has no chance to work.

http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.1

"If no Accept header field is present, then it is assumed that the
client accepts all media types."

Since the validator sends no Accept header, it should be assumed that
the client accepts all media types. If your rule assumes that it
doesn't support XHTML because of that, then isn't that a flaw in the
rule?

> If I am right, why doesn't validator 0.8 beta send an Accept-Header,  
> and would'nt it be better to do so?

I'm going to waffle here for a bit. I haven't looked at the on going
discussions on Accept headers and the validator for a while, but here
are some thoughts on the subject.

As a general markup validator, it should be able to cope with any XML
or SGML thrown at it. There is a convention that XML media types take
the form of application/SOMETHING+xml, and I don't think there is a
way to express that it accepts that in the Accept header.

Of course it could claim to accept text/*, application/*, but that
will match a bunch of things that it can't deal with
(application/javascript for example), although it would deal with
image/* and so on.

Received on Tuesday, 24 April 2007 11:03:48 UTC