Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id FAA25555; Thu, 27 Jul 1995 05:06:26 -0400 X-Resent-To: drums@CS.UTK.EDU ; Thu, 27 Jul 1995 05:06:25 EDT Errors-to: owner-drums@CS.UTK.EDU Received: from vall.dsv.su.se by CS.UTK.EDU with ESMTP (cf v2.9s-UTK) id FAA25548; Thu, 27 Jul 1995 05:06:22 -0400 Received: from ester.dsv.su.se (ester.dsv.su.se [130.237.161.10]) by vall.dsv.su.se (8.6.10/8.6.9) with SMTP id LAA10714 for ; Thu, 27 Jul 1995 11:06:12 +0200 Received: by ester.dsv.su.se (4.1/SMI-4.1) id AA10235; Thu, 27 Jul 95 11:06:11 +0200 Date: Thu, 27 Jul 1995 11:06:10 +0200 (MET DST) From: Jacob Palme X-Sender: jpalme@ester To: ietf-drums Subject: Message-ID mandatory Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII I suggest that the following text is added to the description of this header field: A message should always be given a Message-ID by the sending mail system. If, in spite of this, a message arrives without a Message-ID, then a gateway, DL expander or other mail system may create such a Message-ID for the message according to the rules specified below. The creation of such a Message-ID can be useful to allow meaningful "In-Reply-To", "References", "Obsoletes" and "Supersedes" references to the message, to allow correlation of duplicates of the same message and stopping of loops. In certain cases, such a creation of a Message-ID is compulsory in a gateway to a network in which Message-ID-s are mandatory, such as Usenet News. Algorithm for creation of a Message-ID for a message which lacks such an ID: Take the contents of the "Date" heading field, concatenate this with the content of the "Subject" field, convert any spaces to dashes, concatenate this with the e-mail address in the "From" field. Put an asterisk between each of the values concatenated to separate the parts in the combined string. Rationale: In this way most of the advantages of a Message-ID can be available also for messages lacking such ID-s. Note that the algorithm used to generate a Message-ID is such that there is almost no risk of two different messages getting the same Message-ID, and also that different mail systems or gateways which generate a Message-ID for the same message will arrive in most cases with the same Message-ID value. (Possibly one might use a checksum instead of the whole string as defined above to reduce the length of the Message-ID generated in this way.) ------------------------------------------------------------------------ Jacob Palme (Stockholm University and KTH) for more info see URL: http://www.dsv.su.se/~jpalme