Received: from localhost (daemon@localhost) by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id WAA21908; Thu, 6 Aug 1998 22:09:50 -0400 (EDT) Received: by cs.cs.utk.edu (bulk_mailer v1.10); Thu, 6 Aug 1998 22:09:18 -0400 Received: by CS.UTK.EDU (cf v2.9s-UTK) id WAA21861; Thu, 6 Aug 1998 22:09:17 -0400 (EDT) Received: from apprentice.qualcomm.com (apprentice.qualcomm.com [129.46.2.86]) by CS.UTK.EDU with ESMTP (cf v2.9s-UTK) id WAA21834; Thu, 6 Aug 1998 22:08:25 -0400 (EDT) Received: from [129.46.219.133] (randy-mac.qualcomm.com [129.46.219.133]) by apprentice.qualcomm.com (8.8.5/1.4/8.7.2/1.14) with ESMTP id TAA08052; Thu, 6 Aug 1998 19:06:46 -0700 (PDT) Mime-Version: 1.0 Message-Id: In-Reply-To: References: X-Mailer: QUALCOMM Eudora Pro v4.1 for Macintosh Date: Thu, 6 Aug 1998 18:17:11 -0700 To: "Randall S. Winchester" From: Randall Gellens Subject: Re: Call for Concensus on Recent issues (4.5.3) Cc: DRUMS WG Chair , Detailed Revision/Update of Message Standards Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Random-Signature-Tag: v1.0a10 At 3:00 PM -0700 8/6/98, Randall S. Winchester wrote: > "The general principle that relaying SMTP server MUST NOT, and delivery > SMTP servers SHOULD NOT, perform validation tests on message headers...." > >This statement is no longer accurate. With current anti-spam SMTP servers >(relaying or not), performing validation tests on message headers, is a >feature. > >As an aside, I could not find where this "principle" came from. It's in general a bad idea, and has caused interoperability problems (delivery or even relay servers take it on themselves to police messages, and decide that something or other isn't to their liking, usually incorrectly, and often corrupting the message in a misguided attempt at fixing it. >2) I sent mail about this before, but am not sure I ever recieved a >resonable responce, and since there was no wording changes... > >I currently reject any message with more then 500 (configurable) "Exposed >Recipients". This is not a temporary failure, and is not something that gets >"let in 100 at a time", like the "recipients buffer" section suggests. > >I have no limits on the total number of "RCPT TO" addresses, but do "perform >validation tests on message headers", to verify the number of "Exposed >Recipients". SMTP does allow you to reject a message for any reason you choose. You can always fall back on that general principle, even though you may be violating a rule by basing your decision on something that is likely a bad idea. In your specific case, I'm sure you have ample past experience to justify what you're doing.