Re: The problem with proxy-revalidate, and a proposed solution
David W. Morris (dwm@xpasc.com)
Fri, 20 Dec 1996 12:45:19 -0800 (PST)
On Fri, 20 Dec 1996, Jeffrey Mogul wrote:
> on whether the cache involved is at an end-client system, or a
> shared proxy cache. (We can probably treat a non-shared proxy
> cache as being a sort of "distributed" implementation of an end-client
> cache.)
Good point ... actually the right wording, I think, would describe
not the implementation (end-client vs shared proxy) but more like
"on behalf of a single identified user". It should be considered legal
for a proxy to be shared but be 'end-client' under the current wording
as long as the proxy has identified the user and interprets the caching
rules on a per individual user basis just as if the proxy were not
shared. There are some interesting application and economies of scale
possibilities which I wouldn't want precluded by the letter of the 'law'.
I think the intent is that an individual 'human' receive reliable
delivery of information. I believe automated clients are a subset
of what a human might do in a global information space and such
clients can be modeled as humans. It would be OK for a single
human to use two different UAs against the same cache for example but
the automated client is probably not going to do that.
Dave Morris