SIMPLE WG IETF 57 Monday July 14, 2003, 1430-1630 Hall G Chairs: Jon Peterson, Robert Sparks Minutes taken by Dean Willis. --------------------------------------------------- Meeting called to order by Robert Sparks Agenda review ------------- No changes. Simple Arch Draft, Avshalom Houri draft-houri-simple-arch-01.txt --------------------------------- Slides presented. Rationale, mission, and status of draft reviewed. Need to prune scope of draft, as if completed with current outline it would be longer than RFC3261. WG is asked to review and comment on the list. Volunteers for sections are requested to step forward. Comment: Aki Niemi suggests breaking it up into three different documents: Example call flows, document summary, and descriptions of deployments. Message Sessions, Ben Campbell draft-ietf-simple-message-sessions-01.txt ------------------------------ Slides presented. Reviewed discussion from interim meeting in Ottawa. Draft is currently in WGLC. Discussion of accept-wrapped-types attribute to clarify text of slide. Draft is correct, but example on slide is misleading. We need a better estimation of the session inactivity timer values. Question: How do you take a relay out of service for administrative reasons. We once had a way to inform users that this could happen. We did this before by not allowing a lease renewal. Since this has been replaced by auto-renewing timers, this doesn't work anymore. This needs to be discussed on list to see if we still need the administrative function. Adopting a well-known service port. Note that negotiation will still provide a port, but it should generally be the wks number. The text includes IANA boilerplate for a port assignment. Is this all we need to do for a port number, or do we need to use the online form for port number assignment? AD discussed process -- since we can test without the IANA action, we can proceed with IANA considerations section in document. Open Issue, Clean Shutdown: Do we need to make sure there are no outstanding SENDS before ending a session? Consensus that the possibility of pending messages needs to be discussed in the draft. Open Issue, 426 Response Code Extensibility: Does this need to be extensible to other protection types? Consensus, no. Open Issue, Session Inactivity Timer: Does this need to be something different than 1 minute? Open Issue, StartTLS: Discussion indicates that a 3-byte lookahead is enough to detect a start-TLS. Should we add one? Consensus: yes Poll: Are we happy enough with the doc that we can finish the nits and send it to IESG? Answer: Yes. Data Manipulation, Jonathan Rosenberg draft-ietf-simple-data-req-03.txt draft-ietf-simple-xcap-00.txt draft-ietf-simple-xcap-auth-usage-00.txt draft-ietf-simple-xcap-list-usage-00.txt draft-ietf-simple-xcap-package-00.txt ------------------------------------- Slides presented. XCAP Issue #1, actual filesystem hierarchy for buddy lists: Discussion indicates that what we're really talking about is atomicity of resources, not filesystems. Batching operations: Solution here is not to do batching now. Proposal on HTTP usage: accepted by consensus. Open issue: HTTP limitations. Discussion that we can use separate sub-URIs instead of XPATH. Data Manipulation Requirements: Proposal accepted. Issue, Union vs Most Specific: Proposal to union offered, discussion reuqested. Many open issues not discussed. Publication, Aki Niemi draft-ietf-simple-publish-01.txt ---------------------- Slides presented. Changes since last version reviewed Open issue, Atomicity of publication: 3 options presented. This is essentially the partial publication problem as Jon Peterson undertands it. Is this in scope at this time? Comment: this applies to any event package with segmented information, such as watcher-info. Much discussion with no conclusions. Event Filtering Requirements, Hisham Khartabil draft-ietf-simple-pres-filter-reqs-01.txt draft-ietf-simple-winfo-filter-reqs-00.txt draft-ietf-simple-presinfo-deliv-reg-00.txt ---------------------------------------------- Open issue: watcher "learning new items of the presence information"? Do we need this capability as a requirement? Proposed that we reject this requirement at this time. Poll for volunteers to formally review: Ben Campbell, Avashalom Houri, Jonathan Rosenberg SIMPLE Filter Format, Hisham Khartabil draft-khartabil-simple-filter-format-00.txt draft-khartabil-simple-filter-funct-00.txt -------------------------------------- Issue, XPath Usage: used to declare the elements that should be included in the NOTIFY body. Comment that XPath is a very heavy tool. Another issue is that the expressivity of a raw element name is insufficient for dealing with complexities of derived and transformed data. Issue, Current draft applies filters to URIs: Is something more like authorization list (allows domains) more appropriate? Issue, Full or partial state?: Not discussed. Issue, Server in a foreign network: Can we enforce filtering in intermediate nodes? Deferred to list. Issue: Subscribe refresh with selected filter indications. Discussed, no conclusion, deferred to list. Partial Notification, Robert Sparks draft-lonnofors-simple-partial-notify-02.txt ----------------------------------- (Mikko did not get to discuss this draft during the meeting due to time contstraints.) Poll: accept this as a working group item if it can be added to charter? Hum, as evaluated by chair, indicated consensus to accept. RPIDS, Henning Schulzrinne draft-ietf-simple-rpids-01.txt -------------------------- Slides presented. Material reviewed in presentation. Issue, Document structure options: Proposed that time element be moved into predictive presence category. Discussion will follow. Open Issue, DND: Used priority indication in previous revs. Discussed, no apparent conclusion. Open issue, XML declarations of attribute values. Proposal accepted. Presence Capabilities, Lonnfors draft-lonnfors-simple-prescaps-ext-01.txt ------------------------------- Discussion as to whether we are ready to see this as a WG item. Chairs decided that another rev is required first.