20:14:56 RRSAgent has joined #ws-async 20:14:56 logging to http://www.w3.org/2005/06/08-ws-async-irc 20:15:11 Meeting: WS Async Task Force 20:15:15 Chair: Glen Daniels 20:15:20 Scribe: GlenD 20:15:34 Topic: What are we doing? 20:15:46 Glen describes making a decision tree populated with concrete proposals 20:16:57 Topic: San Francisco F2F 20:17:22 Umit, Steve, maybe DaveO, Glen, are all going to be at JavaOne 20:17:27 might be a good opportunity for some discussion 20:17:48 ACTION: Inquire about who's going and getting space with a speakerphone 20:18:17 ACTION 1: Glen to inquire about who's going and getting space with a speakerphone 20:18:47 Topic: What are we doing? 20:19:04 Dave: Can we consider the specs on a document-by-document basis and investigate changes 20:19:19 Dave: Can we say for sure that WSA core + SOAP docs are OK as they stand? 20:20:04 Glen: Some of the stuff we've discussed seems a little sticky... 20:20:34 Dave: A lot of this is about MEP bindings at various levels. If we need to have a MEP identifier in the message, for instance, that might affect SOAP binding. If not, we're ok. 20:20:58 MSEder has joined #ws-async 20:22:29 Glen: There are actually potentially other situations (challenge-response security) which can cause a "rebinding" of WSDL MEPs to SOAP MEPs... 20:22:51 Dave: We should make sure to say that's explicitly out of scope (we can't cover it all) 20:23:15 Glen: Or leaving it open enough to be general and not specify everything, but allow it 20:24:10 Glen: So we seem to agree on the wire stuff - and we're deciding on spec-ese, right? 20:25:10 Dave: There is a question at the wire level - if you may-or-may-not get a response, how do we bind it? As an HTTP 202 (empty) or a well-specified SOAP message? If we do the latter, (even if that's optimized to 2XX) that makes things clearer... 20:26:30 Glen: OK, so this is a decision point - model the "ack response" in the midst of an async request-response as a SOAP message or not 20:27:45 Mike: What about errors? 20:28:08 Tony: What we're talking about is the ability to indicate that the HTTP connection is "done"... so you know when to close the channel 20:28:59 Tony: The multi-connection people won't be happy with the ack coming back (even virtually) as a SOAP message. I like it a lot, but there'll be backlash. 20:29:28 Glen: I think this makes a good deal of sense as well. 20:31:58 Glen: So what does this actually mean? Is this all described in the WSA WSDL binding document? 20:32:38 Tony: Can we mark in the WSDL that we're using this approach? 20:32:55 Glen: We're defining the meaning of , for instance, so yes 20:33:54 Dave: (describes generic knobs on bindings which talk about backchannels, MEP-complete markers, etc) 20:36:09 (discussion of BEEP and binding to protocols that have backchannels but not correlation) 20:36:26 http://lists.w3.org/Archives/Public/public-ws-async-tf/2005May/0037.html 20:38:16 Glen: We need to start pulling this together as concrete proposals 20:39:39 ACTION: Dave to take a stab at drafting his earlier thoughts (http://lists.w3.org/Archives/Public/public-ws-async-tf/2005May/0037.html) into a more concrete proposal 20:39:40 http://lists.w3.org/Archives/Public/public-ws-async-tf/2005May/0038.html (starting point) 20:42:57 (discussion of this approach) 20:43:30 Glen: How do we bind this stuff up to the next (wsdl) level 20:46:14 Dave: Example - req/resp. My client-side looks to see if there's a non-anonymous - if so sending is fine no matter the binding. 20:46:41 ... if it's anonymous, check to see that the binding (or something else) supports the back-channel feature, and if so OK. 20:47:00 ... if not, we can fail before sending. 20:47:53 (discussion of same thing for ) 20:49:26 Dave: Key point is that the infrastructure can look at a) message contents, b) features provided by bindings, and c) the WSDL MEP in effect, and know what to do 20:49:40 Glen: WSDL MEP generalizes to "context" 20:50:34 Glen: What about specifying the kinds of things you can specify (bindings) in headers? 20:50:57 Dave: Orthogonal, but needs to be on the tree as well. Need a means of advertising this. 20:54:03 Glen: I'll try to pull together the decision tree frameworks before next week, and we can try to get things solidified before then / during that call 20:58:15 The soul-searching long post that triggered the shorter posts after it: http://lists.w3.org/Archives/Public/public-ws-async-tf/2005May/0027.html 20:59:25 -GlenD 20:59:27 -Dave_Hull 20:59:27 -MSEder 20:59:28 -TonyR 20:59:29 WS_TF(async)4:00PM has ended 20:59:30 Attendees were GlenD, MSEder, TonyR, Dave_Hull 20:59:47 Present: GlenD, MSEder, TonyR, Dave_Hull 20:59:58 RRSAgent, generate minutes 20:59:58 I have made the request to generate http://www.w3.org/2005/06/08-ws-async-minutes.html GlenD 21:00:06 RRSAgent, bye 21:00:06 I see 3 open action items: 21:00:06 ACTION: Inquire about who's going and getting space with a speakerphone [1] 21:00:06 recorded in http://www.w3.org/2005/06/08-ws-async-irc#T20-17-48 21:00:06 ACTION: 1 to Glen to inquire about who's going and getting space with a speakerphone [2] 21:00:06 recorded in http://www.w3.org/2005/06/08-ws-async-irc#T20-18-17 21:00:06 ACTION: Dave to take a stab at drafting his earlier thoughts (http://lists.w3.org/Archives/Public/public-ws-async-tf/2005May/0037.html) into a more concrete proposal [3] 21:00:06 recorded in http://www.w3.org/2005/06/08-ws-async-irc#T20-39-39