Re: text/html for xml extensions of XHTML


Our recent dialog has run:

ian> Would you also like PNGs incorrectly identified with the MIME
ian> type text/html due to circumstances beyond the authors control to
ian> be sent to PNG decoders?

robert> Is this a trick question, perhaps?  I think I would, wouldn't
robert> I?  At least it seems like I would be happier just having the
robert> image appear properly, than having it interpreted as horribly
robert> garbled HTML.  What's the catch?

ian> Should we throw away the whole basis of MIME types and the HTTP
ian> Content-Type header, and just use content sniffing instead?

This leaves me wondering if you are debating in good faith.  If you
are just marking time defending a decision that is already carved in
stone, just say so and let's quit wasting time.  I have a bunch of new
documentation to write explaining to our customers all the extra work
they will have to do if they want to try to accomodate readers using

If not, you need to give me some sign that you actually understand the
issues at stake.  From what you write, you give the clear impression
that you don't think either of the following issues are important:

1) For some time to come, most web authors will be preparing content
   that will be read predominantly with older user agents, and
   therefore need to send documents as text/html.

2) For some time to come, many web authors will end up sending XHTML
   as text/html due to circumstances beyond their control, even if 
   they are willing to send it as text/xml.

If you don't acknowledge those points, there is nothing to talk about.
Good luck popularizing your software.  You've got your work cut out
for you.

If you do acknowledge those points, then you don't need me to point
out why your analogy with PNGs is not very relevant.


Robert Miner                          
MathML 2.0 Specification Co-editor                    651-223-2883
Design Science, Inc.   "How Science Communicates"

Received on Thursday, 3 May 2001 10:39:22 UTC