Errors-to: owner-drums@CS.UTK.EDU
Date: Sat, 5 Aug 1995 12:14:54 +0200 (MET DST)
From: Jacob Palme <jpalme@dsv.su.se>
To: Eric Allman <eric@CS.Berkeley.EDU>
Cc: ietf-drums <drums@CS.UTK.EDU>
Subject: Re: Message-ID mandatory 
In-Reply-To: <199507291256.GAA00362@jean-baptiste.cs.berkeley.edu>
Message-Id: <Pine.SUN.3.91.950805121155.4700F-100000@ester>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Sat, 29 Jul 1995, Eric Allman wrote:

> My take on this is that creating a Message-Id that is not _guaranteed_
> to be unique is worse than having the same message with different
> Message-Ids.  I don't think it is possible to make such a guarantee
> using any algorithm of this sort.  I think you're better off having
> the gateway just create a local Message-Id that is known to be unique
> than trying to guess.

The probability of two messages getting the same Message-ID with
my suggested algorithm is very very low. I feel this is acceptable.

But in case the drums group finds this not to be acceptable,
I suggest that at least such Message-ID-s generated by gatewayed
should be marked in such a way that an intelligent recipient
system knows that this is not an original Message-ID. Certain
intelligent recipient systems might use this information,
for example to try to find duplicates using other methods.

------------------------------------------------------------------------
Jacob Palme <jpalme@dsv.su.se> (Stockholm University and KTH)
for more info see URL: http://www.dsv.su.se/~jpalme