W3C home > Mailing lists > Public > ietf-http-wg@w3.org > January to March 1997

Re: PEP and new methods

From: Rohit Khare <khare@w3.org>
Date: Wed, 19 Feb 1997 13:21:53 -0500 (EST)
Message-Id: <199702191821.NAA07406@anansi.w3.org>
To: gjw@wnetc.com, http-wg@cuckoo.hpl.hp.com
X-Mailing-List: <http-wg@cuckoo.hpl.hp.com> archive/latest/2473

In this case, you don't negotiate a new method, you just use it. At worst,
you can use OPTIONS to learn about which methods a server might volunteer
to divulge.

The failure semantics of an unknown method are quite useful. It's hard to
make upgrades that fail-at-the-server-when-unknown. failure is an important
part of the versioning and compatibility story for dist. obj. systems.

So, to turn it around: if PEP really need semantics for requesting protocols:

"the server MUST encrypt the response or fail"

when talking to a server we weren't "sure" had PEP, or even if it had PEP, if
it had "encrypt", we'd need to invent a new method, say PGET, which would b
be dead on arrival at a non-pep server.

Instead, you use PEP to negotiate features which are not "showstoppers" and
use new methods for those which are.

Rohit Khare

Rohit Khare -- World Wide Web Consortium -- Technical Staff
w: 617/253-5884  --   f: 617/258-5999   --  h: 617/491-5030
NE43-344,  MIT LCS,  545 Tech Square,  Cambridge,  MA 02139
Received on Wednesday, 19 February 1997 10:38:57 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 2 February 2023 18:43:01 UTC