Date: Thu, 13 Aug 1998 10:48:09 -0700 (PDT)
From: DRUMS WG Chair <Chris.Newman@innosoft.com>
Subject: Re: serious VRFY ambiguity
In-reply-to: <19980812124153.12832.qmail@cr.yp.to>
To: Detailed Revision/Update of Message Standards <drums@cs.utk.edu>
Message-id: <Pine.SOL.3.95.980813102746.19761I-100000@elwood.innosoft.com>
MIME-version: 1.0
Content-type: TEXT/PLAIN; charset=US-ASCII
Originator-Info: login-id=chris; server=THOR.INNOSOFT.COM

On Wed, 12 Aug 1998, D. J. Bernstein wrote:
> Why doesn't Klensin just make a clear statement that an implementation
> is permitted to always return 252 to VRFY?

At the June 1996 DRUMS meeting, there was a clear concensus to leave VRFY
as a MUST implement feature.  One justification behind this concensus was
the observation that verify is important for debugging network problems,
so it's important that all implementations include real support for VRFY. 
So I'd interpret the "MUST implement" as "MAY generate 252 for all VRFY
requests by default, but MUST be configurable to provide real VRFY
support (so interoperability/functionality problems can be debugged)." 

VRFY&EXPN were discussed at the following DRUMS meetings:
December 95
March 96
June 96
August 97
and perhaps a couple others I don't have the minutes handy.

It will not be discussed at this meeting unless someone proposes specific
clarifying language which matches previous rough concensus.

In scanning the minutes I noticed that there was desire to clarify the 
distinction between VRFY & EXPN functionality.  I'd like to request the
people expressing that concern make sure the current SMTP draft is
sufficient in this area.

		- DRUMS WG Chair