- From: Jeffrey Mogul <mogul@pa.dec.com>
- Date: Fri, 20 Dec 96 17:05:51 PST
- To: Ted Hardie <hardie@thornhill.arc.nasa.gov>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
I agree that we need to resolve this, but I think that we need to
think about the proxy/cache issue as well. A server which responds
with a 1.1 version advertises its capabilities to both the
requesting user agent and any intervening proxies.
Actually, because the response version number is NOT end-to-end
(it only has meaning for a specific hop), it does advertise
the server's capabilities to the next-hop client, but not to any
subsequent client.
E.g., in this case
origin-server -----> Proxy1 ------> Proxy2 -----> end-client
If Proxy1 sends a response to Proxy2 with
HTTP/1.1 200 OK
then Proxy2 knows that Proxy1 supports HTTP/1.1, but it cannot
infer anything from this about the origin-server, and the end-client
cannot discover the version implemented by Proxy1.
This is my recollection of Roy Fielding's explanation from many
months ago, and I believe that this is the understanding under
which the HTTP/1.1 spec was written. It may need to be documented,
but I think it's too late to change.
-Jeff
P.S.: not that I disagree with your overall position, Ted.
Received on Friday, 20 December 1996 17:17:59 UTC