simple@conference.ietf.jabber.com - 2003/03/17

[14:40] %% kibozer has arrived.
[14:54] %% hildjj has arrived.
[14:54] %% hildjj has left.
[15:05] %% kibozer has left.
[15:07] %% kibozer has arrived.
[15:13] %% kibozer has left.
[15:13] %% kibozer has arrived.
[16:02] %% ulieberman has arrived.
[16:03] %% ulieberman has left.
[16:20] %% louie has arrived.
[16:31] %% jasson has arrived.
[16:38] %% mrose has arrived.
[16:39] %% kibozer has left.
[16:43] %% warlord has arrived.
[16:44] %% gwachob has arrived.
[16:46] %% pgmillard has arrived.
[16:52] %% kibozer has arrived.
[16:52] %% hildjj has arrived.
[16:53] * hildjj has changed the subject to: SIMPLE WG - IETF 56
[16:53] <hildjj> MSRP == Message Session Relay Protocol
[16:54] <hildjj> needed to dispense with dependency on commedia
[16:54] <hildjj> depends on TCP, arbitrary MIME payloads
[16:54] <hildjj> but, no COMEDIA dependency
[16:55] <hildjj> Supports 1 or 2 NATs
[16:55] <hildjj> rosenberg: or N
[16:56] <hildjj> trouble with comedia: how to tie to message sessions? port?
[16:56] <hildjj> roseberg: comedia uses port to demux. that means dynamic ports.
[16:57] <hildjj> r: single, well-known port
[16:57] <hildjj> primitives: BIND/RELEASE, VISIT/LEAVE, SEND
[16:59] <hildjj> flow diagram...
[17:00] <hildjj> 1. Offer (i-am:A, u-b: B)
[17:00] <hildjj> <- 2. Visit (B)
[17:00] <hildjj> lol. exodus rendered that as beer.
[17:00] <hildjj> <- 2. Visit ( B )
[17:01] <hildjj> 3. SEND from gateway to both
[17:02] %% resnick has arrived.
[17:03] <hildjj> open issues
[17:03] <hildjj> ACK related bug
[17:03] <hildjj> (may be able to handle with update)
[17:03] <hildjj> do we need a refresh mechanism for BIND state?
[17:03] <hildjj> Race condition when tearing down a session.
[17:08] <hildjj> digest auth on BIND not fully specified
[17:08] <hildjj> do we need an msrps: URI scheme?
[17:08] <hildjj> e2e security:
[17:08] <hildjj> - MIKEY
[17:08] <hildjj> - S/MIME
[17:08] %% pgmillard has left.
[17:08] <hildjj> - Protection across CPIM gateways
[17:09] %% jishac has arrived.
[17:10] %% Cyrus has arrived.
[17:14] %% pgmillard has arrived.
[17:23] %% resnick has left.
[17:23] %% resnick has arrived.
[17:26] %% gwachob has left.
[17:26] %% lisaDusseault has arrived.
[17:33] %% pgmillard has left.
[17:35] <lisaDusseault> I'm going to transcribe for a while apparently, once I figure out where we are.
[17:36] <lisaDusseault> They're arguing about what tuples are.
[17:36] <warlord> *sighs*
[17:36] %% takamiya has arrived.
[17:37] %% carlalf has arrived.
[17:39] <lisaDusseault> OK the discussion on rpids is beginning now. I thought for the last five minutes they were talking about "armpits".
[17:39] <lisaDusseault> Henning is presenting.
[17:39] %% randy has arrived.
[17:39] <lisaDusseault> Richer presence info than basic PIDF would be cool.
[17:39] <lisaDusseault> Much of it is derivable mechanically
[17:40] <lisaDusseault> This relates to policy somehow...
[17:41] <lisaDusseault> Some of the information that could be included is location type, category, privacy... this is all in the draft.
[17:43] <lisaDusseault> Some of this is controversial and I wanted to get feedback on it
[17:43] <lisaDusseault> Is timed status useful?
[17:43] <hildjj> lisa: was that a question from him, or you?
[17:44] <lisaDusseault> Device capabilities is related to caller preferences
[17:45] <lisaDusseault> Can represent groups, not just individuals, each with their own status.
[17:45] <lisaDusseault> This intercepts with the SIMPLE concept of lists.
[17:46] <hildjj> XML vs. MIME
[17:46] <hildjj> recommendation: leave groups out of RPIDS for now.
[17:47] <lisaDusseault> [I generally transcribe straight and interpose my own comments in square brackets
[17:47] <lisaDusseault> ]
[17:47] <hildjj> [cool. i like that approach. we need a 'ms. manners guide to IM.']
[17:48] <lisaDusseault> The label issue, an important philosophical debate.
[17:48] <hildjj> PIDF defines syntax, but almost no semantics
[17:48] <lisaDusseault> The proposal is to change some things without changing the entire document.
[17:49] %% pgmillard has arrived.
[17:49] <hildjj> who does it? the PUA? the compositor?
[17:49] <hildjj> rosenberg: it's that way in PIDF since PA is the only actor that sets these things.
[17:50] <hildjj> speaker: but when would i use that? i don't have to label them.
[17:50] <lisaDusseault> [Speaker is Henning Schulzrinne, btw]
[17:51] <hildjj> [thx]
[17:51] * hildjj has a problem with (name, face) tuples
[17:51] <hildjj> (back to henning)
[17:52] <hildjj> separate label tag proposed. similar semantics, but set by presentity and left alone by PA.
[17:52] <lisaDusseault> [It almost sounds like rather than CSS Henning wants stylesheet transformations -- XST]
[17:53] <hildjj> for policy filtering, ala CSS: id is unique, class = type of element
[17:56] %% venaas has arrived.
[17:57] <lisaDusseault> Oh no, "what is a tuple" again.
[18:00] <lisaDusseault> Jonathan R thinks the wg should have spent the entire time talking about what is a tuple.
[18:01] %% perry has arrived.
[18:03] <lisaDusseault> Henning is now showing a state transition diagram for sending composing and idling.
[18:03] <lisaDusseault> Rohan doesn't like Henning's tutorial.
[18:03] * pgmillard presses the FF key
[18:04] <lisaDusseault> Henning asks what to do with dead composers.
[18:04] <lisaDusseault> Jon Peterson talks about whether this is a wg deliverable.
[18:06] <lisaDusseault> Ted Hardie asks about the extensibility model, perhaps having the wg do all extensions, or perhaps something more... extensible.
[18:07] <lisaDusseault> Bob Wyman points out the patent that may be a problem in this area.
[18:09] <lisaDusseault> Rosenberg: We need to figure out how to manage buddy lists, white lists, etc.
[18:10] <lisaDusseault> Rosenberg: We should look at this more generally than just a protocol for doing buddy lists.
[18:10] <lisaDusseault> R: ACAP is what I had in mind -- it's all about server storage and client access to config type info.
[18:11] <lisaDusseault> R: The server is agnostic about data.
[18:11] <lisaDusseault> R: Multiple client access is the problem being solved.
[18:11] <hildjj> [note: BuddyList is a TM of AOL]
[18:12] %% venaas has left.
[18:12] %% venaas has arrived.
[18:12] %% venaas has left.
[18:12] <lisaDusseault> Patrik: There is a difference between moving fast and talking fast. Slow down your speech but speed up the content.
[18:12] <lisaDusseault> R is skipping over auth stuff.
[18:13] <lisaDusseault> R: I've written a draft that analyzes usage of ACAP for simple.
[18:15] <randy> What is the name of that draft?
[18:16] <lisaDusseault> R: This is a role-based system as well as direct identity access control.
[18:16] <pgmillard> randy: I think: draft-rosenberg-simple-acap-data-00.txt
[18:16] <lisaDusseault> Randy, I don't know -- Rosenberg just asserted that he wrote a draft.
[18:16] <lisaDusseault> R: I love the data model in ACAP. Inheritance and access control stuff is nice, it's a really good fit.
[18:17] <lisaDusseault> R: The protocol itself is troublesome because it's a long-lived client-server connection.
[18:17] <randy> I found it: draft-rosenberg-simple-acap-data-00
[18:17] <lisaDusseault> R: It supports change notification but only if both support long-lived connections. Resynching data is not easy.
[18:18] <randy> Thanks pgmillard
[18:18] <lisaDusseault> Rohan asked some question about intermediary requirements.
[18:19] <lisaDusseault> Rosenberg: It might be one of these 3GPP things.
[18:19] <pgmillard> http://www.ietf.org/internet-drafts/draft-rosenberg-simple-acap-data-00.txt
[18:19] <lisaDusseault> Henning: I wasn't clear on the data model, is the auth model inherently ACAP or is it replacable.
[18:19] <hildjj> [i think the issue is your roster should come from a local server, so why do you need intermediary support]
[18:20] <lisaDusseault> Joe: How can it come from a "local" server when your "local" can change every day?
[18:20] %% jm has arrived.
[18:20] <pgmillard> heh, true.
[18:21] <lisaDusseault> BTW, I'm experimenting with different ways of showing what I'm transcribing as opposed to what I'm saying... I'm not quite happy with what I've used so far.
[18:21] <hildjj> data has to be stored in a single, place, though, right?
[18:21] <lisaDusseault> Christian at the mike > Why do you want this old thingy that nobody has used for five years?
[18:22] <lisaDusseault> Rosenberg> Everybody except for MS.
[18:22] <lisaDusseault> R> It's true, very little commercial deployment.
[18:23] <hildjj> r: we're going to suggest a new thing that takes from different places
[18:23] <hildjj> lisa at mike: what about Web-DAV?
[18:23] <hildjj> rosenberg: no.
[18:24] <hildjj> [DHCP!]
[18:24] <lisaDusseault> Rohan> The user profile stuff that is going to support ACAP or some other stuff as long as it uses a URI.
[18:25] <hildjj> jon: hum. should rosenberg's draft go forward?
[18:25] <lisaDusseault> Peterson> Asking for hum on whether Rosenberg's draft should become a WG item.
[18:26] <hildjj> lisa: I like > for quotes, : for addressing a user in the room.
[18:26] %% Cyrus has left.
[18:26] <lisaDusseault> Joe: I agree.
[18:26] <hildjj> and nothing (like this) for just talking as yourself.
[18:26] <lisaDusseault> Rosenberg> Closing point here: this stuff is pressing.
[18:26] <randy> Hum
[18:27] <lisaDusseault> R> I get blasted for simple not having a solution to this.
[18:27] <lisaDusseault> R> I don't want this to turn into a three-ring circus.
[18:28] <lisaDusseault> Cullen> I read the ACAP documents with a bad attitude.
[18:28] <hildjj> peterson> literally takes away mike.
[18:28] %% carlalf has left.
[18:28] <lisaDusseault> New topic, adam roach - ?
[18:29] <lisaDusseault> Roach put some stuff on the screen that I can't possibly read.
[18:29] <lisaDusseault> R> Open issue is how we treat filtering.
[18:29] %% randy has left.
[18:30] <lisaDusseault> R> The filters themselves should be part of the resource (i.e. the list). But that's not flexible. Alternate proposals?
[18:30] <hildjj> lisa> BIG sneeze
[18:31] <lisaDusseault> R> I need help with XML. THat said, here goes.
[18:31] <lisaDusseault> I still can't read the text on his screen.
[18:31] %% jasson has left.
[18:32] %% randy has arrived.
[18:32] <lisaDusseault> R> Need discussion about filter handling on list.
[18:32] <lisaDusseault> Peterson> We will victimize to get that discussion. Rohan is assigned.
[18:32] <lisaDusseault> Rohan> We need to get approval from Allison to do this here because of the option tag.
[18:33] <lisaDusseault> Arcane discussion of stuff ensues between Peterson and Rohan about mechansim.
[18:35] <hildjj> Hisham Khartabil at the screen
[18:35] %% resnick has left.
[18:35] <lisaDusseault> Khar> The subscribeer can select what information they want
[18:35] <lisaDusseault> Khar> Not just the status, but only certain values of status.
[18:36] <lisaDusseault> Khar> This reduces the number of notifies sent, reduces cost.
[18:36] <lisaDusseault> Khar>XPath standard was borrowed as the syntax for this.
[18:37] <hildjj> i'm interested to know about namespace issues w.r.t. XPath.
[18:37] <lisaDusseault> Khar> We'd like these drafts to be wg items.
[18:37] <hildjj> does the thing I'm sending this to have the same ns prefixes?
[18:37] <lisaDusseault> Joe: Namespcaes and XPath together SUCK.
[18:37] <hildjj> nod. my point.
[18:38] <lisaDusseault> Rosenberg> Both drafts describe filter management -- setting the filter, lifetime, replacement.
[18:38] <lisaDusseault> Rosenberg> That's common to both and should be abstracted out.
[18:39] <lisaDusseault> Rosenberg> I'm not convinced XPath is the right thing.
[18:39] <lisaDusseault> R> It's a generic filter but it isn't a trigger. It doesn't make sense to do that separately.
[18:39] <lisaDusseault> Henning> Generally XPath not expressive at the right level, may not be on the right element
[18:40] <lisaDusseault> H> It's not intuitive and may not work for all the things we want to do.
[18:40] %% takamiya has left.
[18:41] <lisaDusseault> Peterson> Can I get a hum of people who think we should take on the requirements and framework for both these things -- throttling and filtering notifications.
[18:42] %% jm has left.
[18:43] %% kibozer has left.
[18:43] %% kibozer has arrived.
[18:43] %% hildjj has left.
[18:43] %% hildjj has arrived.
[18:43] %% pgmillard has left.
[18:44] %% lisaDusseault has left.
[18:44] <hildjj> p> the hum was withdrawn
[18:44] <hildjj> (we lost network connectivity, and subsequently we missed something)
[18:46] %% pgmillard has arrived.
[18:46] %% pgmillard has left.
[18:46] %% lisaDusseault has arrived.
[18:47] <lisaDusseault> Peterson> Is it worth going forward with this req'ts document? Hum?
[18:48] <lisaDusseault> Peterson>We'll try to get reviewers to look hard at this.
[18:48] <hildjj> lukewarm hum, but no real opposition.
[18:48] %% mrose has left.
[18:48] <lisaDusseault> Peterson> In this area I'm just a working group chair; Ted Hardie is new area director for this area.
[18:50] <lisaDusseault> Meeting adjourned.
[18:50] <perry> cool.
[18:50] <lisaDusseault> "Proceed to dinner"
[18:50] %% hildjj has left.
[18:50] <warlord> later
[18:50] %% perry has left.
[18:50] %% warlord has left.
[18:50] %% perry has arrived.
[18:50] %% lisaDusseault has left.
[18:50] <perry> there is no "dinner" chat room!
[18:50] %% perry has left.
[18:59] %% kibozer has left.
[18:59] %% jishac has left.
[23:04] %% randy has left.