November 8, 1999
Chair: Glenn Parsons
Scribe: John W. Noerenberg
Charter & Chair
The IESG is happy with draft charter, except the WG needs a chair who isn't also principal author or significant contributor to the documents produced by the Working Group. The concern is based on experience with other industry groups that have tried to create 'rubber-stamping' WGs in the IETF. The ADs have suggested John Noerenberg as a co-chair for group. John and Glenn have agreed to work together. With this development, then IESG will charter the WG. This should be completed within the next 2 weeks.
VPIM v2 to DRAFT
The are several requirements to advance VPIM v2 (RFC 2421) from Proposed to Draft Standard status: operational experience; multiple, independent implementations; and dependence only on protocols that have advanced to at least Draft status. There is some additional minor document editing as well.
RFC 2421 depends on several protocols that are currently at Proposed Standard status. The WG needs to decide if the portions of VPIM v2 which depend on these protocols must remain in the definition or can be removed without sacrificing necessary functionality. Two that were discussed at length were protocols for Delivery Status Notification (DSN) and Message Delivery Notification (MDN). Both of these protocols are necessary, but their advance is reasonably close at hand. The vCard protocol's status wasn't known, but it may be possible to eliminate it from the v2 definition if it cannot advance. Other documents mentioned but not discussed since they are actually part of VPIM v2 were image/tiff (RFC 2302), Audio/32kadpcm (RFC 2422), Content Duration (RFC 2424) and multipart/voice-message (RFC 2423).
A new draft has been released <draft-ema-vpim-vpimv2r2- 02.txt>. Changes that have been made are to more clearly separate send and receive semantics. Support for vCard has been reverted to attached spoken name to ensure that this is not an impediment to move forward. New language also makes it clear that Sensitivity MUST prohibit a message from being forwarded. A fax part MAY be discarded. V2 specifies that if any addressee is unknown, the message MUST be refused. For multipart/mixed messages containing multipart/voice-message, the protocol recommends the voice- message part is the first part of the message.
There were some changes to clarify differences between Clients and Servers. These affected Notification primarily. DSNs are specifically to announce delivery to the recipients mailbox, and MDNs reflect when the entire message is retrieved from the mailbox. There was some discussion about ORCPT. The only time it may be used is when the original recipient is returning delivery status.
To document interoperability and to report on experience, a website has been available to implementors for the past couple of years. The website contains the vendor completed conformance appendix of RFC2421 as well as results based on the test suite. The site is at the EMA hosted VPIM website: http://www.ema.org/vpim/conformance/conformance.html
There have been several "bake-offs" conducted between implementors, and some results have been posted. Glenn encouraged implementors to post their independent test results.
Anne Brown spent a few minutes describing directory work being conducted in other WGs, and elsewhere in the industry, which affect VPIM. She identified where additional work is necessary to support VPIM. First of all, the ability to return spoken name as part of the address resolution needs to be addressed. Industry experience indicates this must be accomplished in less than 2 seconds. Secondly, directory services must also return voice mail capabilities information which is not likely to be covered in other schema. Previously she had authored a draft (now reissued as draft-ema-vpimdir-schema-01.txt) to address directory issues. It is useful principally to identify these issues. Though this draft assumes an LDAP directory service, it may be more appropriate not to assume an underlying protocol at this stage. She described the ENUM WG work, and encouraged the group to investigate this further. She also mentioned the RESCAP WG. Outside the IETF, there are two trade groups working on voice mail directory problems, a group in Europe, the Voice Mail Association, and another, similar group in the US, the Tele-Messaging Industry Association.
VPIM Desktop (VPIM v3)
Several people are working on documents for VPIM v3. Laile Di Silvestro, Glenn Parsons, and Stuart McRae described their accomplishments.
Laile Di Silvestro indicated she has collected a number of changes for the goals document, but has not yet produced a new draft. The changes are considered to be major. Glenn and Stuart described how the first attempt at v3 has been pared down into what is now known as simple V3.
At a recent meeting sponsored by EMA, it was recognised that accomplishing the goals for v3 couldn't be accomplished by simply revising the v2 document. Since v2 focuses on communication between voicemail servers, instead of between the endpoints, it isn't really a good basis for v3. As a result, instead of referring to it as VPIM v3, Stuart recommended we describe this new document <draft- ema-vpim-voicemsg-00.txt> as Internet Voice Messaging.
There are several notable changes from VPIM v2. Internet Voice Messaging (IVM) reduces the number of codecs to a single mandatory one and an optional codec. Some of the criteria for choosing the codec include wide availability on general purpose user computers, supports streaming audio, and has no IPR use restrictions. MS-GSM carried in "audio/wav" would be an acceptable choice. The format is described in draft-ema-vpim-wav-00.txt and draft-ema-vpim- msgsm-00.txt, but Microsoft has not yet released its IPR. IVM adds the concept of a Primary Content indicator, and seeks support for MDN and DSN reports for only parts of a message. Internet Fax is also very interested in the latter two features.
It is expected that IVM implementations will be able to process v2 messages, but there is no expectation that a processor only capable of v2 will be able to process IVM messages. No compatibility language is in the current IVM draft, Stuart confimred that it should appear in a separate document (there is currently an early draft by Lance Mansfield). VPIM v2 describes support for TIFF-F fax, but IVM will likely support extended mode. It is unclear whether IVM will require support for vCard.
There was extended discussion on the purpose of the Primary Content indicator. Most likely it will be interpreted as an indication from Sender to Receiver which part of a multipart message the Sender considers the primary part of the message. Emily Candell has volunteered to take a first cut at drafting a revised primary content proposal. In addition, Eric Burger volunteered earlier to create a draft of partial non-delivery.
A series of IMAP extensions that may be important to VPIM was enumerated, but there was little discussion.
The group will meet as a WG at IETF 47. Glenn suggested that an additional meeting early in 2000 would be helpful to advance the work.
Glenn summarized the meeting. To advance VPIM v2 up the standards track, Glenn encouraged interoperability test results be sent to him. The people working on the IVM documents must continue to incorporate the feedback from the group and create new drafts in the next few months. The (soon-to-be-chartered) WG should plan on meeting at the 47th IETF to continue development of the IVM documents.
Internet Voice Message