Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id IAA10858; Wed, 17 Jan 1996 08:37:55 -0500 Received: by CS.UTK.EDU (bulk_mailer v1.3); Wed, 17 Jan 1996 08:37:42 -0500 Received: from mailhost.pipex.net by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id IAA10840; Wed, 17 Jan 1996 08:37:40 -0500 Received: from pipe.pipex.net (actually mailhost.pipex.net) by pipe.pipex.net with SMTP (PP); Wed, 17 Jan 1996 13:37:34 +0000 To: drums@cs.utk.edu Subject: Re: loop detection In-Reply-To: Date: Wed, 17 Jan 1996 13:37:33 +0000 From: Tim Goodwin Message-ID: jgm+@CMU.EDU (John Gardiner Myers) wrote: > One possible approach would be to replace the function of Received: > with a different header, defined to be extensible. I've been mulling over the problem of loop detection for a while now, and it occurred to me that trace information should really be in the envelope (as it is in X.400). I am considering drafting an SMTP extension (probably called `TRACE') which would allow conforming servers to carry trace information in the envelope. The envelope trace would be converted to `Received:' headers during final delivery, or when the message was relayed to an SMTP server not supporting this extension. Is this a good idea? A bad idea? Has anyone done any work on anything similar? Since this isn't really a DRUMS issue, I suggest that replies are sent to me. If there's sufficient interest, I will set up a mailing list for further discussion. Tim.