SCIM 2012-11-08 =============== Author: Milan Sova Date: 2012-11-08 Thu 1 Note well ------------ 2 Participation ---------------- 3 Goals for the Design Team ---------------------------- - [LJ]: DT not expected to be around for the lifetime of the WG - no discussion 4 A word from the tracker general ([MS]) ----------------------------------------- - 29 tickets, 18 from Google tracker - tracker open to anyone with a tool account - [LJ]: are those issues reasonable? volunteers? - Kelly Grizzle, Eric Wahlstrǒm 5 vCard investigation ([CD]) ----------------------------- - Adopt the vCard schema? - should decide sooner than later - [TN].. are we talking about a default vs the only schema? - [CD] looking at it from the replacement POV: are there other alternatives? - [TN] what is in the spec would be mandatory, we need extensibility - [CD] vCard as mandatory - [PHo] vCard format vs schema? format is "old" - [CD] looking at vCard4 - esp the JSON representation vs looking at the semantics - [CD] goals for the schema? - [??] looking at the functionality of the format (vCard) not that much the semantics - [CD] is JSON for vCard viable for SCIM? - Yes. Are there enough advantages in vCard to justify a change for SCIM? - No - [PSA] Doing the thing from scratch => would make sense to use vCard. Nowadays there's not much use in switching to vCard. - [??] Did you check the tracker for requirements? - [CD] presents a list of pros and cons for vCard - [??] using vCard might cause loosing traction for SCIM from the current community - [PSA] vCard folks did some work we would do ourselves anyway (take lessons from their work) - [LJ] are you supporting the vCard? - [PSA] No - just take the lesson from vCard work - [PHo] to [PSA] is it OK for 2 IETF groups to have 2 different extensibility mechanisms, 2 registries... The switch might solve that for the community - [MS] If switching is a big win it would show up. the namespace issue is not that big - [CD] there are lessons from vCard 4. there might be a way how to reuse the registries - [MA] How is the extensibility of vCard? - [CD] there's a slight advantage on SCIM schema on the extensibility side - [PHo] using vCard registries by referring eg. all calendar related stuff to vCard registries - [??] some existing vCard implementation might benefit from "provisioning" features of SCIM - [CD] Are those in the same category of apps? And would they be compatible with our using of JSON? - [EL] (Jabber) aren't we at the mapping discussion now? - [LJ] I don't think we are ready to decide, however what is the opinion of the room? - [PSA] I'm not sure I agree with all the statements on the slide. I think there's more we need to do.. - [EL] (Jabber) I'd prefer asking for the room's opinion after the mapping discussion - [JB] Changing SCIM to a JSON format - how much work is that be for the folks vs changing to vCard. Advantages to stay with SCIM 1.1 is more political than technical. 6 vCard mapping ([BG]) ----------------------- - draft-greevenbosch-vcard-mapping-00 - mapping SCIM -> vCard - mapping vCard -> SCIM - [PHo] how is UID mapping from SCIM to vCard done when the IdP name is not known - [BG] you have to make up a name for it - [PHo] OK - [CD] have you looked in cardinality incompatibilities between SCIM and vCard (multiple vs single occurrence) - [BG] we have to look at it - [JR] in view of the previous presentation you just described backwards compatibility translation - another reason to make the new SCIM schema to reuse as much from vCard as possible - [PSA] how much has the SCIM 1.1 been extended (is extensible)? - [KL] even if we don't use vCard, is it still useful to have the mapping document. - [PSA] mapping is good as the input for decision. If WG goes for non-vCard schema keeping the two in sync would not bring any value - [TN] I need a data model no matter what we choose. Not sure we're in the stage to decide the format unless we have the data model - [JB] Tony Nadalin is right: we need the data model for the format decision - [EW] there are fields in vCard that are too loosely defined (eg. KEY) - interop? - [BG] agree - [LJ] not talking about thew WG adoption now but keep on working on ti (keeping it under your control) and ... - [CD] what about the discussion about the data model? - [LJ] could you provide a kick-off document/email? - [LJ] hum: - vCard only - vCard not at all - undecided - keep investigating- HUM WINNER 7 Use-case doc ([ZZ]) ---------------------- - draft-zeltsan-scim-use-cases - [BL] are there objections to adopting the ID as WG doc? 2012-11-08 - [KG] the use-cases are pretty specific, would make more sense to have more generic use case. - [KL] the scenarios doc is much better - [MA] scenario doc outdated - do we have a volunteer to carry on? - [LJ] we seem to have 2 candidates: whet do you ([KL]) bring the alternative to the use case doc - [KL] in a month - [TN] the docs serve different purposes: scenarios (generic), use-cases (specific usage patterns). We might use both - [LJ] IF we have two, we will compare them 8 Open mic ----------- - [BL] as AD the charter looks OK for the current work - [TN] multiple tenants cases - the doc does not deal with it => potential interop problems. For me it is one of the larger issues. We need some very strongly recommendations on how do we do tenancy. - [JB] some people may argue that sales force is multi-tenant. we should have as little way to handle MT as possible - [PHu] uniqueness of attributes across all tenancies (one email for all tenancies) - [MA] let DT to identifies a couple of ways how to deal with MT - [TN] see if we could throw our thoughts bout our use cases on the list - [KG] issue 27 (remove XML from the spec) - seems to be supported by a reasonable number of people - [LJ] hum - drop XML - WINNER (both room and jabber) - keep support for XML (silence) - unsure (silence) - [TN] interop? - [LJ] branding? does the group want to have a domain run under IETF.org? - [PHo] has been declined by IESG in the past - [BL] it goes both ways, I'm willing to take it to IESG - [LJ] use simplecloud.info vs something with SCIM in it? group: keep current site - [EL] by vCard we mean JSON representation of vCard?if yes who;s working on it? - [PSA] we have a 00 id on JSON representation of vCard. - [CD] (Cyrus) working on vCalendar in JSON - - [PR] - weirds WG considering using JSON/vCard. Who's capable/willing to work on JSON/vCard WG? - [TN] Need to look at authorization aspect of SCIM. - [LJ] Charter: "authorization out of scope" - [BL] You can document using existing authorization scheme - [TN] We could work on how to profile OAuth2 - [EL] We need to decide quick (about the format) - [PSA] To Pete's point: if we're going to the JSON/vCard WG we need input from the WG - [LJ] - issue 25 - [LJ] : propose appropriate text to deal with it and send it to the list - [LJ] is the group interested in iterop session? - [PSA] virtual or real no consensus 9 Personae ----------- [BG] Bert Greevenbosch [BL] Barry Leiba [CD] Cyrus Daboo [EL] Eliot Lear [EW] Erik Wahlström [JB] John Bradley [JR] Justin Richer [KG] Kelly Grizzle [KL] Kepeng Li [LJ] Leif Johansson [MA] Morteza Ansari [MS] Melinda Shore [PHo] Paul Hoffman [PHu] Phil Hunt [PR] Pete Resnick [PSA] Peter Saint-Andre [TN] Tony Nadalin [ZZ] Zachary Zeltsan