- From: Dave Kristol <dmk@allegra.att.com>
- Date: Tue, 20 Dec 94 14:01:20 EST
- To: ses@tipper.oit.unc.edu
- Cc: fielding@ics.uci.edu, http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
Simon E Spero <ses@tipper.oit.unc.edu> said on Tue, 20 Dec 94 13:53:33 -0500: > > It is possible to send the first request with the session establishment > packet, so you don't have to wait for the round trip. If the session method > is rejected, the wrapped request will just be discarded. I think your remarks are a response to something I wrote, but I'm not sure what. One concern I mentioned was that a server might want to initiate a session, something the client wouldn't have known about when it made its original request. Consider Basic authentication as an example (ONLY as an example, not necessarily a good one): 1) Client issues GET 2) Server says, "No, authentication needed". 3) Client issues another GET, this time with authentication 4) Server responds with object. If the server knew, at step 2, that the client would tolerate a session, it (server) wouldn't have to close the connection. The client could do the GET again on the same session. Dave Kristol
Received on Tuesday, 20 December 1994 12:02:03 UTC