Re: Proposal on removing Content Negotiation from http 1.1

Peter J Churchyard
>I would like to propose that content negotiation is removed from http 1.1
>
>Thus accept: headers are ignored. Greatly simplifies caching decisions since
>response == entity.
>
>I beleive that very little content negotiation takes place (even within
>this archive you see people explicitly refering different versions of
>papers..)
>

I sympathize with this as a pragmatic solution to the lack of resolution
on negotiating content quality, an old HTTP idea that has never quite
crystalized.

However, I would hate to see Accept: not defined in so far as client
helpers are defined. Many potential applications of helpers are crippled
currently because the client does not send the browser a list of the types
accepted by its associated helpers.

It is surely reasonable to put in the spec that browsers SHOULD send a list
of ALL the MIME types accepted by their helpers.

My recollection is that the fisrt release of Netscape did this and then
it was suddenly replaced with the few types that could be displayed in-line
plus * for the rest.

It is exasperating not to be able to get the registered helper types from
the browser.

b.

Dr Brian R Gaines               Knowledge Science Institute
                                University of Calgary
gaines@cpsc.ucalgary.ca         Calgary, Alberta, Canada T2N 1N4
403-220-5901  Fax:403-284-4707  http://ksi.cpsc.ucalgary.ca/KSI

Received on Wednesday, 24 January 1996 16:08:57 UTC