Re: Content Negociation: State of the art ?

Larry Masinter:
>
>The work on content negotiation and feature negotiation is active in
>the HTTP working group (http-wg@cuckoo.hpl.hp.com). There's a draft by
>Andy Mutz and Koen Holtmann, and comments on their draft would be
>appreciated.

I maintain a homepage with links to the content negotiation drafts and
other relevant documents: <URL:http://gewis.win.tue.nl/~koen/conneg>.

As for my assessment of the state of the art:

># For us, content negociation deals mainly with terminal capabilities
># (screen size, color vs B&W, memory, ... and client capabilities)

If you want to negotiate on these things *now*, you will have to use
some `click here for the X version' scheme, and you could use cookies
or bookmarking to make such choices persistent.

Near future: The HTTP-WG has been working on a standard for `feature
negotiation' since early this year.  I expect the feature negotiation
standard to be finished within a few months.  All research problems
have been solved.  

Browsers supporting feature negotiation can deal with the terminal
capabilities you mention.  I don't think negotiation on memory is a
big problem; basically you can negotiate on everything the browser can
measure.  Note however that it is impossible to predict how long it
will take until most web browsers support feature negotiation.  Broad
deployment has the usual chicken-and-egg problems.

> but
># also with connection capabilities (bandwidth, reliability, delay).

The state of the art for this part: measuring these things on the
internet in the general case is still very much a research problem,
and the HTTP-WG is currently not trying to solve it.  Proxy caches
between the browser and the server make the problem even more
complicated in the HTTP case.

Of course, negotiating on very coarse-grained differences like GSM
bandwidth vs. Ethernet bandwidth is no big problem.

Koen.

Received on Saturday, 9 November 1996 11:41:57 UTC