CURRENT_MEETING_REPORT_ Reported by C. Allan Cargille/MCI Minutes of the Mail Extensions Working Group (MAILEXT) Agenda o Introduction, participants, approval of minutes, revise charter, revise agenda. o Individual work items: - 521 status codes proposal (John Myers). - Language tag proposal (Harald Alvestrand). - File transfer body part MIME mapping (Ned Freed). - Pipelining SMTP extension (Ned Freed). - Checkpointing SMTP extension (Dave Crocker and Ned Freed). - Binary and chunking SMTP extensions (Greg Vaudreuil). - A-BoMBS and C-BoMBS (Jerome Houttein). - Text/html proposal from the HTML working group (Eric Huizer presenting). Jacob Palme asked if this group is an appropriate forum for discussion of work being done in the X.400/OSI community on incorporation of Internet addressing. John Klensin responded that this is not an appropriate forum for such discussion, but that it can be discussed if time permits. 521 Status Codes Proposal John Myers presented an overview of the 521 status code proposal. In brief, this proposal is something that network entities that do not support e-mail would support. It causes mail to immediately bounce when an attempt is made to send it to such an entity. Currently such mail tends to sit in some queue somewhere until it times out and is returned. Keith Moore raised the issue of how MX records pointing at such entities should work. The current specification says that the message should bounce. Keith argued that it would be preferable to try other MX records instead. John Myers then pointed out that he had recently realized that this proposal is not necessary. It is almost as easy to implement a simple SMTP parser that returns a 5xx code to any RCPT TO command. This approach has the advantage it works with existing broken implementations that effectively ignore the initial banner status. Harald Alvestrand and others countered by saying that standarding the meaning of 521 in this context is useful in its own right. In addition, John's proposal is problematic in that it could be construed that the postmaster address should always work. Harald also suggested that this proposal needs to be tried to be assessed. If it is deployed and is useful it could be moved to standards track. If not, it should be dropped. This in turn argues that the document should be issued as Experimental. Allan concluded by suggesting that the document authors should attempt to reach consensus on the list for the documents moving forward as Experimental. Language Tag Proposal The group considered the language header proposal. Harald summarized this proposal as one that allows labelling the content of a message or part of a message as being in a particular language. Jacob asked if the language concept extends to computer languages. Harald said it does not, but it could be extended to cover this. Nathaniel Borenstein asked about how multi-lingual text is handled -- it is done by specifying multiple language values. Ned Freed noted that this work may be part of future efforts in extending MIME to uses in non-messaging environments. This should not preclude moving the document along the standards track now, however. John Klensin pointed out that the syntactic use of dash differs from previous usage in RFC 822. Harald responded that this is by design, in that while the dash is a syntactic element he wants to capitalize on the effect it has of making it look like a single token. Dave Crocker added that this needed to be made explicit in the text. Dave Crocker also raised the issue of why this does not make sense as a content-type parameter. Ned countered by pointing out that global content type parameters are not allowed in MIME, and that this information spans multiple content types. The group concluded that this document should be submitted for consideration by the IESG as a Proposed Standard. File Transfer Body Part MIME Mapping Ned presented his file transfer body part, SMTP pipelining, and SMTP checkpointing extension. Ned prefaced his remarks by stating that only the pipelining extension is a candidate for advancement at this time. The file transfer body part work is a result of work undertaken by the EMA Message Attachment Working Group (MAWG). The specification seeks to map X.400 file transfer body parts into appropriate MIME objects and vice versa. Discussion focused on the need for a general file transfer mechanism in MIME and whether or not this mechanism should be extended to provide this facility. Ned pointed out that since the EMA's use of file transfer body part is to provide a MIME-like mechanism for object encapsulation, not to use the mechanism for the (obvious) purpose to transfer files per se, and as such the mapping specification seeks to provide an object mapping rather than a file transfer mechanism. Eric Fair noted that in the best of all worlds it would be possible to specify generic formats for generic object types. Dave Crocker and others noted that while this would be desirable it is not present-day reality, and is not likely to become reality in the foreseeable future. The discussion concluded with Ned stating that he wants to consider the various issues brought up by the group and see how the document needs to be changed (or not changed) to accommodate them. Pipelining SMTP Extension Ned reported that, as tasked by the working group, he had communicated with Eric Allman about the issues of implementing the pipelining extension in sendmail. Eric had responded that the problems with implementing pipelining in sendmail (e.g., the use of a fork system calls in the middle of the SMTP dialogue) would probably be addressed in the long term by other planned modifications. As such, Ned recommended moving the document forward for consideration as a Proposed Standard without any sendmail-specific changes, and the working group agreed. Checkpointing SMTP Extension The checkpointing proposal was discussed. One of the coauthors (Ned) had thought of this proposal as something of an academic exercise. However, the other author (Dave Crocker) disagreed and so did a considerable fraction of the group. It was accordingly decided that implementation of this proposal is needed, and if it proves to be useful the working group should consider moving this work forward on a standards track. Binary and Chunking SMTP Extensions Greg Vaudreuil presented his chunking and binary SMTP extension specification. Discussion centered on interoperability testing and issues surrounding binary MIME. The latter were ruled out of scope for the present discussion, but the former are of such concern that the working group decided to ask for some indication of interoperability before advancing the document to Proposed Standard. A-BoMBS and C-BoMBS The BoMBS documents were reviewed. A large number of substantive comments were made about both documents. Dave Crocker stated that this work is extremely important and long overdue, and that the best approach would be to charter a separate working group to deal with these documents in detail. This suggestion was favorably received by the working group. Text/html Proposal from the HTML Working Group Eric Huizer presented a proposal for the handling of text/html, which was worked out by the HTML Working Group (which met in parallel with Mail Extensions Working Group). In brief, the HTML group has decided on a canonical form that agrees with MIME text/plain (i.e CRLF line terminator sequence). Clients are supposed to support a wider variety of terminators, but text/html is supposed to be in this form. This proposal was very favorably received by the Mail Extensions Working Group. Eric Huizer also presented a proposal for a MIME testing service. A document will be posted to the list describing this service. Conclusion Allan Cargille concluded the meeting by thanking the various document authors for the work they have done on these proposals.