W3C home > Mailing lists > Public > www-talk@w3.org > May to June 2001

Re: text/html for xml extensions of XHTML

From: Ian Hickson <ian@hixie.ch>
Date: Tue, 1 May 2001 16:58:44 -0700 (Pacific Daylight Time)
To: Robert Miner <RobertM@dessci.com>
cc: <hammond@csc.albany.edu>, <mozilla-mathml@mozilla.org>, <www-talk@w3.org>
Message-ID: <Pine.WNT.4.31.0105011648340.800-100000@HIXIE.netscape.com>
On Tue, 1 May 2001, Robert Miner wrote:
>
> Hi Ian,
>
>> Remind me, why would you want to send XML as text/html?
>
> This is the crucial all-important point.  No one wants to send XML as
> text/html.

If that is really the case, then I am very glad to hear it. It sure sounds
like people actually want to though.


> The problem is that due to circumstances beyond their control, large
> numbers of of authors will end up sending XML as text/html for a long
> time to come.
>
> The main reasons for this are:
>
> 1) improperly configured web servers at ISPs

Configuring files with an "xml" extension to be sent as "text/xml" takes
one line in a multi-site configuration file.

Is it really worth the effort of making the change in user agents rather
than politely asking one's ISP to fix the configuration, or taking one's
business elsewhere?

How many servers still don't come preconfigured with "text/xml" as the
MIME type of ".xml" files? Are there any?


> 2) old tools

Which "old tools"? Tools won't just "work" with XML if it is sent as HTML.


> 3) ignorance

Users don't need to know about this, so long as XML editors output files
with "xml" extensions and web servers are configured to return these files
as "text/xml".

I don't hear anyone screaming for XSL to be sent as text/html, or CSS to
be sent as text/html -- why is XHTML a special case?


> /inconvenience

I cannot believe that renaming a file to ".xml" instead of ".html" when
the file contains XML instead of HTML is any more of an effort than
changing from outputting HTML to outputting XML in the first place.


> No matter how strong the technical and theoretical arguments in favor
> of the strict use of MIME types, this fact is a major barrier to
> adoption.

It doesn't seem to have stopped the adoption of CSS.


> I understand perfectly why W3C has to take that position, and why you
> want to adhere strictly to W3C Recs.  But I still need a solution for
> our customers.

"Fix the webserver"?

Why is it easier/cheaper/safer to change the user agent and propagate it
to millions of users than it is to change a single line in a server
configuration file?


> Let's try a completely different tack.  What about a completely ad hoc
> interim fix, such as permitting an optional declaration of the form
>
> <!-- mozilla-mime-preference: text/xml -->
>
> as the absolute first thing in a file to triggers the XML parser?

Hmm.

   <!-- -moz-content-type-override: text/xml -->

I wouldn't be totally averse to that, although I still don't believe it is
easier to change every XHTML file than it is to fix the servers.  Also, I
don't see how it will help the users of other compliant web browsers.

-- 
Ian Hickson                                            )\     _. - ._.)   fL
Invited Expert, CSS Working Group                     /. `- '  (  `--'
The views expressed in this message are strictly      `- , ) -  > ) \
personal and not those of Netscape or Mozilla. ________ (.' \) (.' -' ______
Received on Tuesday, 1 May 2001 19:56:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 27 October 2010 18:14:25 GMT