Re: Interest in standard HTTP headers for specifying client

Graham Klyne:
>At 10:40 AM 9/18/97 PDT, Larry Masinter wrote:
>>But this is already work in progress:
>>      draft-ietf-http-feature-reg-02.txt describes a mechanism
>>for registration of features such as color depth and number of pixels,
>>and draft-ietf-http-negotiation-03.txt includes an "accept-features"
>This is being picky, but Koen's up to ...-04 on the 'negotiation' draft.

Yes.  By the way, I maintain a page on it at .  You can also find a scenario
draft there.

To pick another nit, I would not really call the above a `work in
progress', at least it won't progress much further in this working
group.  The working group will not propose standard HTTP headers for
specifying client display attributes.  It will finish the work above
soon by releasing the protocol in the form of `experimental RFCs'.
These experimental RFCs are not standards, but merely encourage
interested parties to try out the protocol.  If deployment of these
experimental RFCs snowballs, then there will probably be a new IETF
working group which works them into a real standard.

All this means that standard HTTP headers for specifying client
display attributes will not appear very soon, if ever.  My most
optimistic estimate puts it at 1 to 2 years.

I don't have very high hopes for there being an ad-hoc standard in the
mean time either.  I remember that some Japanese client implementers
are using a common format for encoding some atribures in the User-Agent
header field, but I can't find any reference to it.



Received on Friday, 19 September 1997 01:19:32 UTC