Minutes of the Internet Relay Chat Update (IRCUP) BOF
Reported by: Ian King <firstname.lastname@example.org>
John Noerenberg II, Chair, commenced the meeting with a review of the agenda. He announced that two persons would be presenting drafts, and asked for other drafts; no other drafts were suggested.
The Chair then presented a draft charter for the group. One point made is that there is currently no draft on the standards track addressing chat. Jake Khuon, an attendee from Merit Network, Inc., stated that there is a group in the existing IRC community which has discussed chat technology issues for approximately five years. This group is composed of users from several nets, and there is not necessarily a great deal of consensus within the group on any given issue. Khuon agreed to discuss participation in IETF's effort with the IRC-resident group.
Discussion of the draft charter continued. Area Director Harald Alvestrand asked whether this working group would review the fundamental model of chat, or simply assume the IRC model. One presenter, Kevin Hay, agreed that is an issue; the AD then observed that the issue of overall model belongs in the charter. The Chair indicated he would discuss such issues regarding the charter with the AD outside this meeting, and post a revised draft charter to the mailing list.
Kevin Hay then presented his draft on a Multicast Chat (MCC) protocol (draft-hay-mcc-00.txt). The AD asked about reliable multicast in the context of MCC; an attendee from Sun stated that this protocol is at least a year from standards track. Although unreliable multicast is available now, it was pointed out that it is unsuitable for text transfer, where losses are quite obvious. The AD stated that this, too, is an issue for a possible Working Group.
The AD asked about scalability, with a target figure of one million users on 100,000 channels. The presenter opined that it might be necessary to establish additional Control Channels to meet this. There were more questions, but the presenter's allotted time had expired.
Lisa Dusseault presented the next draft (draft-pfenning-irc-extensions-01.txt). This draft assumed existing IRC technology, but proposed modifications and extensions to client-server protocol to allow various features. The draft specifically does not address client-client or server-server issues.
Contention concentrated on two areas: the increased server overhead and statefulness which might be required by these extensions, and whether the extensions would be useful in the two distinct configuration models (small nets, typically privately owned, or large nets, typically public). It was also questioned whether some features proposed could be addressed with the existing CTCP.
The presenter also discussed private email she received from Mr. Ierymenko, who has worked on the NIRCD [?] project. The presenter stated that his email set forth several interesting ideas, although most were very specific to the NIRCD implementation. The presenter will continue discussions with Ierymenko. She also invited client implementers to contact her for further discussion.
The Chair informally presented a proposed agenda, which projects adoption of a standard by the end of 1998. He will publish this agenda on the mailing list for comment.
Discussion returned to the charter and the question of whether to create a Working Group. Issues identified were scaling of a multicast solution; reduction in protocol overhead; and distinction of client-server and server-server considerations. Problems identified were reduction of traffic; authentication; namespace; ratings and restrictions; i18n; persistence of channels; and discovery of channels and users.
Upon the question being raised, there was consensus that the issues warrant formation of the Working Group; however, the AD pointed out that the group must first investigate the existing IRC community and encourage them to be part of the process. (One person pointed out that the existing community is usually contemptuous of standards and process.) The Chair will seek to invite participation from the existing IRC community in the work of this potential Working Group. The group also agreed that regardless of whether there is participation by the existing IRC community, the work proposed is significant to the IETF and should be forwarded by this group.
A new draft charter will be posted to the mailing list for review.