W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1999

Re: WG meeting in Washington?

From: Roy T. Fielding <fielding@kiwi.ICS.UCI.EDU>
Date: Wed, 06 Oct 1999 15:56:08 -0700
To: Mark Nottingham <mnot@mnot.net>
cc: HTTP Working Group <http-wg@hplb.hpl.hp.com>
Message-ID: <199910061556.aa00872@gremlin-relay.ics.uci.edu>
>I didn't intend to 'make over' the advice to look like protocol requirements, 

Sorry, that wasn't intended as a criticism of you -- just the current
format of your draft.  If it were actually a protocol spec, I'd be
congratulating you on following the proper form and making it easier on
the reader.  What I am worried about is misinterpreting implementation
guidance as protocol requirements.

>but I freely admit that they can be interpreted that way; I 
>chose to do it that way because it seemed the most natural and clear way to do
> so. My logic was that the requirements set forth 
>in it were only in the scope of the document, so that a product could be
>referred to as compliant with it, over and above protocol 
>compliance. This will hopefully be more clear in a future draft of the
>document (if there is indeed any point in further revision).

I think it is a worthwhile document, provided that it emphasizes rationale
and not requirements.  For example, there is very good rationale for having
the server constrain the i/o interface to conform to the protocol, and there
are many examples you might use to explain the rationale, but the protocol
requirements do not change as a result.  However, phrasing that rationale
as "a server MUST do ..." is stating a protocol requirement even if the
scope of the document excludes protocol requirements.

>I apologize if it's not politic to submit this to the IETF, but it seemed the 
>best place to start (as I don't have the cash to join the 
>W3C). I am (or at least I was during the writing of this draft) very much an
>outsider. If someone has a suggestion about where 
>this document would best reside (whether that place is in the http-wg or not, 
>as there does seem to be some dissention), I'll be 
>more than happy to talk about it (as I've stated before).

I'm not a W3C member, so I wouldn't encourage that path either.  What I
meant is that W3C does theoretically include implementation conformance 
among its members as part of its spec writing scope.

Larry is the one who decides what is or is not appropriate for this forum,
and it seems that he wants to discuss the document here regardless of its
eventual publication status.  I agree with his reasons, though again I would
prefer that the format be changed so that it won't be misinterpreted.

....Roy
Received on Wednesday, 6 October 1999 23:57:19 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:33:34 EDT