- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Tue, 29 Aug 95 09:24:07 MDT
- To: Larry Masinter <masinter@parc.xerox.com>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
I'd be more comfortable with this and other proposals if the original
data were delivered with a 'content-ID' that was unique to the
particular content being delievered, and any follow-up request for
partial delivery were indexed not by the original URL but by the
content-ID itself. (The lack of a content-ID would be a signal that
the object wasn't suitable for partial transmission or that the server
didn't support such.)
The content-ID could also be used as a GET IF-DIFFERENT key, since the
content-ID would change when the content changed. Content-ID is
required to be globally unique. (It might even be a URN, but who
knows?)
This sounds quite like the "Cache-Validator:" unique-ID that I proposed
to replace IF-MODIFIED-SInce (sorry for the weird typing, I'm on a really
slow link). Once you have a way for the server to explicitly control
caching on an object-by-object basis, lots of problems solve themselves.
-Jeff
Received on Tuesday, 29 August 1995 09:34:45 UTC