Re: content negotiation

Regarding user agents telling servers that they cannot (or will not) view
frames, and servers responding intelligently, Jeff de la Beaujardiere wrote:
>> I believe that transparent content negotiation in HTTP (see
>> http://gewis.win.tue.nl/~koen/conneg/) will provide that functionality.
>> [...]

David Perrell responded:
> This is a very interesting proposal, but it implies a separate file for
> each possibility, doesn't it?  This would certainly be a boon in the case
> of muli-language documents, where the bulk of the content is
> different. However, there are so many non-related features that may or may
> not be enabled on a UA, a separate file for each possible combination
> isn't practical. With an active server page, content with applicable
> features can be composed from one file.

Content negotiation will be exceedingly useful when generating HTML
dynamically and will perhaps be used most often in that context.  It's
certainly much better than the horrible practice of checking the user agent
field.  However, I don't know whether the negotiation scheme would allow
different elements of the same static document to be handled well.  I think
the server must respond with the entire document, at which point it's up to
the UA to display what it can.

--Jeff dLB

  = J-F Pitot de La Beaujardiere
  = delabeau@iniki.gsfc.nasa.gov
  = http://globe2.gsfc.nasa.gov/

Received on Thursday, 1 May 1997 17:38:20 UTC