Re: XHTML 1.1 as text/html

On Fri, Apr 27, 2007 at 02:07:04PM +0200, Sierk Bornemann wrote:
> >3. Accept Header application/xhtml+xml is at work for the validator.
 
> Does this mean, Validator will send an Accept Header in the near future?

It means it probably will be. The issue is under discussion in
Bugzilla.

> >   (Note btw, that once the validator sends accept header  
> >application, the
> >    configuration mistake of Sierk would not be detected, which is  
> >bad too)
> 
> What kind of Accept Header precisely will Validator send?

That is still to be determined.

> Why wouldn't it be detected by my configuration either?

The error in your configuration would not be detected (i.e. that it
sometimes serves XHTML 1.1 documents as text/html).
 
> Would you point each user out there to configure his browser/OS so,  
> that Firefox will open, if he wants to surf a XHTML webpage, which is  
> served with "appliocation/xhtml+xml"?

That was the default behaviour of Internet Explorer 6 on my Windows
2000 system after I installed Firefox. No configuring required.

> I think, you wouldn't. Because that's far from being practical for  
> public.

Exactly, XHTML 1.1 is not a practical choice of markup language for
use client side for the typical webpage[1].

> If that is your intention, then XHTML 1.1 and all successors  
> have no chance to be used public and they would never have.

Not while a very significant number of users use a client that doesn't
support it.

It is unfortunate that Microsoft abandoned development of their
browser for a long time, and didn't encourage users to switch to
something more suitable when it came available. Happily, development
on Internet Explorer is underway again, and XHTML support appears to
be on the planned feature list.

> >The validator did the right thing in this case and at the time being:
> >    an XHTML 1.1 document is sent with text/html.  Wrong.
> 
> I would agree, if there wouldn't be any mechanism to detect the  
> clients' capabilities. But there is a mechanism. Which doesn't work  
> properly, because Validator doesn't behave like a talkative client  
> and doesn't tell details about its capabilities.

As has been mentioned:

* the Accept header doesn't claim _support_ for a given media type
* the Accept header IS optional, and not providing one claims acceptance
  of everything

Also, the point of content negotiation is to find a version of a
resource that is acceptable to a client, not to provide the same
resource while claiming it is different things.

> >I wonder if Sierk is using ruby markup in his pages. If not, your  
> >document could be XHTML 1.0.
> 
> Surely, my document could be XHTML 1.0. It surely could be HTML 4.01.  
> Both served as "text/html". But the markup of the most of my webpages  
> fulfills the needs to use a XHTML 1.1 doctype.

I skimmed your homepage and couldn't see any sign of data that was
marked up with features added in XHTML 1.0 or 1.1. Could you provide
an example of why you need XHTML 1.1?

> So why not using it?

Because (a) You aren't using it according to spec and (b) Client
support for it is poor

>  I intentionally want to use that Doctype,
> which fits the content most, and vice versa, I try to write my
> content to fit the needs not only of XHTML 1.0 but even to fit the
> needs of XHTML 1.1, I even try to write my content in a manner, that
> it is relative easy to make a transition to XHTML 2

XHTML 1.1 is very unlikely to provide a better point to switch to
XHTML 2 then XHTML 1.0 is.

> or (X)HTML 5 (or
> whatever name it would be in the future according to the HTML WG).

HTML 4.01 could very well be a better starting point for a move to
whatever the HTML Working Group provides then any form of XHTML.

Using a markup language that is poorly supported because it MIGHT make
it easier to switch to a markup language that is still being designed
IF browser support ever appears is, frankly, silly.

Content should be written for the clients people use today, not
speculative future clients (which are likely to have a high level of
backwards compatibility anyway).

> I intentionally have on my agenda to write some pages with
> XHTML/SVG/ MathML. Therefore, I *have to* use
> "application/xhtml+"-Mimetype.

So write THOSE pages in XHTML 1.1 and serve THOSE pages as
application/xhtml+xml (and only as application/xhtml+xml).

> I intentionally want to use the appropriate and recommended XML-
> Mimetype for XHTML (whatever version, 1.0 or 1.1). I intentionally
> want to avoid to serve my webpages with an SGML-Mimetype "text/html"

Why? It is well supported, it is very likely to continue to be well
supported. 

XHTML support is browsers which do support it is relatively poor ...

   "there is no need to serve application/xhtml+xml to Mozilla. In
   fact, in versions prior to Gecko 1.9/Firefox 3, doing so would
   deprive the Mozilla users of incremental display, because
   incremental loading of XML documents has not been implemented in
   those versions. Serving valid HTML 4.01 as text/html ensures the
   widest browser and search engine support."
     -- http://www.mozilla.org/docs/web-developer/faq.html#accept

... and Firefox 3 is not out yet.


[1] It is useful in other situations, my website is written in XHTML,
but webpages don't remain in that form when they reache a browser
(although the same data stays in XHTML form when embedded inside
ATOM). The practical advantages are primarily on the server.

-- 
David Dorward                                      http://dorward.me.uk

Received on Friday, 27 April 2007 13:49:21 UTC