Re: distinguishing browser types

Brian Gaines (gaines@fsd.cpsc.ucalgary.ca)
Sat, 22 Apr 1995 17:19:00 +0500


>Mike Meyer remarks
> Yup. Personally, I think that the current mechanism is fine for
> distinct objects; most browsers have a "save-to-file" facility that
> gets invoked on any type they don't otherwise know how to deal with,
> so you can just send "*/*" (or nothing) if you don't want to deal with
> the quality issue.
>

Hi Mike -- the problem is format negotiation. I have services that can
give better interaction if an appropriate helper is installed, but can
still give useful functionality if not. So I need to know at the server
whether the helper application is registered or not. That was the purpose
of the Accept field -- not to say "I accept *.* because I can dump it in
a file regardless of whether that file is usable or not."

It's like a lot of HTML/HTTP, what was intended has never been defined,
and I guess Roy Fielding's work on defining HTTP is where we should look for
clarification. The debate on www-talk at least demonstrates that there
are 2 schools of thought (if not more!).

b.

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