- From: Koen Holtman <koen@win.tue.nl>
- Date: Mon, 30 Sep 1996 23:30:25 +0100 (MET)
- To: Larry Masinter <masinter@parc.xerox.com>
- Cc: koen@win.tue.nl, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Larry Masinter: > >You are concerned about destroying the effectiveness of caches when >using Accept: <negotiable> and Vary: accept to designate variations on >some header other than Accept, e.g., Accept-Language. Yes. >Your example is where variants available in languages `en' and `nl', >user requests `dk'. > >However, this concern is primarily aimed at the situation where the >cache is unaware of transparent content negotiation. Even more, it is _exclusively_ aimed at this situation. >If the cache >_is_ aware of transparent content negotiation, it can engage in >content negotiation directly itself, can't it? Yes. And when doing do, it can disregard the constraints in the Vary header (because it is observing the constraints in the Alternates header). So for these caches, efficiency is not affected if we go for Accept: <negotiate>. Koen.
Received on Monday, 30 September 1996 15:38:26 UTC