http protocol for caching, not http caching

The primary goal of the http-caching subgroup is to review those
elements of Roy's HTTP/1.1 proposal that affect caching proxy servers.
Primarily you want to make sure that the kind of caching proxy servers
in use today (e.g., proxy caches for companies, firewalls,
hierarchical national caches) are supported. Secondarily, you want to
not make some of the more experimental caching schemes (using
multicast, netnews, prospective retrieval) impossible.

On the other hand, the HTTP caching group need not actually design or
promote a particular kind of caching and proxying, or to find ways of
performing caching or proxying that is more effective than what we
have currently deployed.

Give us the language, don't write the novels.

I think in particular that you must leave open the possibility that
servers will want to do content negotiation using factors that cannot
be easily downloaded to the proxy or client (since that's what we have
now).

I don't think you need to or should design a system for explicit
revocation.

Roy made some explicit proposals in his HTTP/1.1 specification. I hope
that the subgroup starts out reviewing them, with regard to their
adequacy or lack thereof for caching.

Received on Friday, 29 December 1995 21:48:02 UTC