13:27:29 RRSAgent has joined #prov 13:27:29 logging to http://www.w3.org/2012/11/10-prov-irc 13:27:31 RRSAgent, make logs world 13:27:31 Zakim has joined #prov 13:27:33 Zakim, this will be PROV 13:27:33 ok, trackbot; I see SW_(F2F)8:00AM scheduled to start 27 minutes ago 13:27:34 Meeting: Provenance Working Group Teleconference 13:27:34 Date: 10 November 2012 13:28:10 Agenda: http://www.w3.org/2011/prov/wiki/F2F4Schedule 13:28:25 rrsagent, make logs public 13:34:14 Dong has joined #prov 13:34:15 lebot has joined #prov 13:35:59 Luc has joined #prov 13:36:01 Zakim, do you come in on Saturdays? 13:36:01 I don't understand your question, lebot. 13:36:59 hook has joined #prov 13:37:07 Curt has joined #prov 13:37:52 @Dong, we are waiting for Ivan to bring in the speakerphone 13:41:42 SamCoppens has joined #prov 13:42:48 CraigTrim has joined #PROV 13:43:04 dong are you online? 13:43:23 Zakim, will the chairs be benevolent today? 13:43:23 I don't understand your question, lebot. 13:43:32 smiles has joined #prov 13:43:56 GK has joined #prov 13:44:06 SW_(F2F)8:00AM has now started 13:44:09 TomDN has joined #prov 13:44:13 +??P0 13:44:31 (Silence) 13:44:35 zakim, ??P0 is me 13:44:35 +smiles; got it 13:45:39 simon, dong can you get on skype 13:46:13 we don't have a polycom right now 13:46:54 Topic: Implementation Report 13:47:04 Luc: this session will be about implementation report 13:47:13 Thinks we'd like to do: 13:47:23 1. update on where we are 13:47:35 (Paul notices we're 15 minutes early) 13:48:26 +Curt_Tilmes 13:48:39 OK… we'll restart in 15 minutes… maybe we'll have a speakerphone 13:48:41 zakim, who is on the phone? 13:48:41 On the phone I see smiles, Curt_Tilmes 13:48:55 (Curt's experimenting with a mobile phone connected to Zakim) 13:48:57 yes 13:49:02 zakim, who is on the call? 13:49:02 On the phone I see smiles, Curt_Tilmes 13:49:14 +??P2 13:49:16 I hear you! 13:49:42 ??P2 is me 13:49:54 zakim, ??P2 is me 13:49:54 +Dong; got it 13:51:07 laurent has joined #prov 13:51:40 -Curt_Tilmes 13:53:29 scribe: GK 13:53:37 chair: Luc 13:56:17 ivan has joined #prov 13:56:39 + +1.617.715.aaaa 13:56:53 zakim, this is f2f 13:56:53 ivan, this was already SW_(F2F)8:00AM 13:56:54 zakim, who is here? 13:56:55 ok, ivan; that matches SW_(F2F)8:00AM 13:56:55 On the phone I see smiles, Dong, +1.617.715.aaaa 13:56:55 On IRC I see ivan, laurent, TomDN, GK, smiles, CraigTrim, SamCoppens, Curt, hook, Luc, lebot, Dong, Zakim, RRSAgent, trackbot, stain 13:57:17 pgroth has joined #prov 13:57:17 pgroth_ has joined #prov 13:57:38 pgroth has left #prov 13:58:05 pgroth has joined #prov 13:58:30 Restarting... 13:58:38 Session about implementation report 13:58:43 Would like to: 13:58:49 1. update from Paul 13:59:42 concerned about getting to end of implementation phase, then finding features are not implemented 14:00:00 would like to have advance indication of what people will implemented 14:00:36 jcheney has joined #prov 14:00:38 2. review what we'll do for constraints; in particular what we do for constraints 14:00:43 http://www.w3.org/2011/prov/wiki/ProvImplementations 14:01:16 Paul: talking about "gathering implementation evidence" 14:01:36 3 parts (see page at link above) 14:01:57 Overall happy with framework as described 14:02:16 Ivan: what are the arrows on table 2? 14:02:59 Paul: link to implementation blue arrows consumes, green arrows produces term 14:03:05 s/Paul?Dong/ 14:03:35 action: Dong to describe blue and green arrows in implementation report document 14:03:35 Created ACTION-138 - Describe blue and green arrows in implementation report document [on Trung Dong Huynh - due 2012-11-17]. 14:03:38 q? 14:03:49 Paul: more questions? 14:04:45 Ivan: minor thing… use usual W3C editorial style - do we intend to publish as note? (Looks like it might be one) Clarify that implementation report does not need to be published as TR. 14:05:20 ok, I'll change it to a note 14:05:24 action: pgroth to change the respec style for implementation report 14:05:24 Created ACTION-139 - Change the respec style for implementation report [on Paul Groth - due 2012-11-17]. 14:05:31 q? 14:05:44 http://dvcs.w3.org/hg/prov/raw-file/default/testcases/constraints/process.html 14:06:11 Paul: prov constraints process document… idea to outlines process for testing constraints 14:06:25 format for test case files (sect 2.1) 14:06:56 q+ 14:07:07 ack Luc 14:07:14 identifier… constraint identifiers are embodied in the test case identifier 14:07:34 +[IPcaller] 14:07:41 q- 14:07:42 zednik has joined #prov 14:07:43 Luc: some of the constraints will be renumbered following removal of mentionOf 14:07:49 zakim, who is here? 14:07:49 On the phone I see smiles, Dong, +1.617.715.aaaa, [IPcaller] 14:07:50 On IRC I see zednik, jcheney, pgroth, ivan, laurent, TomDN, GK, smiles, CraigTrim, SamCoppens, Curt, hook, Luc, lebot, Dong, Zakim, RRSAgent, trackbot, stain 14:08:00 action: dong check constraints are matching to the updated document 14:08:00 Created ACTION-140 - Check constraints are matching to the updated document [on Trung Dong Huynh - due 2012-11-17]. 14:08:30 GK: are constraint numbers fragile for this? 14:08:59 Paul: wanted automated reporting of test case coverage. 14:09:07 q? 14:09:12 Ivan: change respec style for this document too 14:09:35 q+ 14:10:09 Paul: hasn't really been reviewed as yet. Need some early review. 14:10:13 zakim, aaaa has SamCoppens TomDN laurent hook Curt pgroth Luc jcheney ivan GK lebot CraigTrim 14:10:13 +SamCoppens, TomDN, laurent, hook, Curt, pgroth, Luc, jcheney, ivan, GK, lebot, CraigTrim; got it 14:10:26 zakim, who is here? 14:10:26 On the phone I see smiles, Dong, +1.617.715.aaaa, [IPcaller] 14:10:27 +1.617.715.aaaa has SamCoppens, TomDN, laurent, hook, Curt, pgroth, Luc, jcheney, ivan, GK, lebot, CraigTrim 14:10:27 On IRC I see zednik, jcheney, pgroth, ivan, laurent, TomDN, GK, smiles, CraigTrim, SamCoppens, Curt, hook, Luc, lebot, Dong, Zakim, RRSAgent, trackbot, stain 14:10:54 Luc: would like to identify reviewers; preferebly developers; mostly not on call. 14:11:06 The sound on the phone line is broken, I have to rely on the scribe :( 14:11:08 this one in respec.js? : var respecConfig = { specStatus: "ED", // specification status (e.g. WD, LCWD, NOTE, etc.). 14:11:37 James: happy to look at this; biggest problem is managing data as number of test cases grows 14:12:10 zakim [IPcaller] is zednik 14:12:20 Luc: need to be clear if test case is expected to succeed; currently in table, but should be in name for automated testing? 14:12:30 Paul: I'm happy with that. 14:13:09 hook: 'unofficial' or 'base' could be used 14:13:22 Dong: prefer using identifer to directory for different outcomes (pass/fail/etc.) 14:13:24 (per http://dev.w3.org/2009/dap/ReSpec.js/documentation.html) 14:14:05 action: Dong to update naming convention to include success/failure of test 14:14:05 Created ACTION-141 - Update naming convention to include success/failure of test [on Trung Dong Huynh - due 2012-11-17]. 14:14:06 q? 14:14:09 q- 14:14:28 Paul: last thing - questionnaire 14:15:09 … idea was to ask implementers to fill out - whatthey support, and also other implementations with which they interoperate 14:15:21 Stephan: questionnaire is complete, has been reviewed 14:15:35 … want another round, get some more implementers to fill it out 14:15:42 @stephan can you add a link http://www.w3.org/2011/prov/wiki/ProvImplementations#Gathering_Implementation_Evidenence 14:15:47 q+ 14:15:57 … discussion on mailing list about external vocabs using/extending prov 14:16:10 … ask these groups to fill out questionnair 14:16:13 q+ 14:16:16 ack iv 14:16:48 Ivan: if I am an implementer, do I see what's in the Google doc? 14:16:58 Stephan: will add link to actial questionnaire 14:17:06 https://docs.google.com/spreadsheet/viewform?pli=1&formkey=dGM4cXZYMk0xaFBDT2VyRV92YkY5WkE6MQ 14:17:23 This is what implementers will see 14:18:06 q? 14:18:09 ack pgroth 14:18:15 I think we'll need a (wiki) page to explain the whole process of reporting an implementation (with links to all the relevant documents), which will be sent with the call for reports 14:18:55 Perhaps, the questionnaire can have include a link to the explanation as well 14:19:25 q+ 14:19:31 ack smiles 14:19:40 Stephan: The first page collects information that controls information displayed on subsequent pages 14:19:42 q+ 14:20:14 Smiles: are tools like prov-python, ?, a framework of an application 14:20:16 zakim, who is here? 14:20:16 On the phone I see smiles, Dong, +1.617.715.aaaa, [IPcaller] 14:20:17 +1.617.715.aaaa has SamCoppens, TomDN, laurent, hook, Curt, pgroth, Luc, jcheney, ivan, GK, lebot, CraigTrim 14:20:17 On IRC I see zednik, jcheney, pgroth, ivan, laurent, TomDN, GK, smiles, CraigTrim, SamCoppens, Curt, hook, Luc, lebot, Dong, Zakim, RRSAgent, trackbot, stain 14:20:17 Prov python is a framework 14:20:47 Stephan: they go down same path, so could combine these as single item. 14:20:54 q- 14:21:28 zakim, [IPcaller] is stain 14:21:28 +stain; got it 14:22:12 Paul: we have four divisions… is the distinction between libraries, services, applications clear? 14:22:24 q+ 14:22:46 Stephan: distinction is not large - maybe not needed? 14:23:05 I think some people might unnecessarily worry about the distinction if there are multiple options 14:23:13 q+ 14:23:18 … also no sections for pure publishers of provenance. Or is that a service? 14:23:22 q? 14:23:41 q+ 14:24:09 Hook: implementation type is single-choice, but some implementations may be more than one of these. 14:24:19 ack ho 14:24:44 Stephan: currently have to fill form multiple times; may want to change the questionnaire to clarify this. 14:25:01 … don't lnow if they can be handled in a single pass 14:25:14 Luc: MentionOf shoukd be removed from the questionnaire 14:25:22 ack lu 14:25:46 Paul: poiple would like to be able click on the questionnaire and see all the questions before filling out. 14:26:39 q? 14:26:40 … maybe have several different questionnaires for each kind of implementation. Click on link, see all questions, without having to branch within the form. 14:26:47 +1 to it's a barrier to "continue" in the survey. 14:26:47 ack pg 14:26:50 q+ 14:26:50 Stephan: I think that's reasonable 14:27:11 q? 14:27:39 q+ to ask if common questions across all questionnaiure tyopes can be auto-filled 14:27:49 action: zednik to create 3/4 questionnaires instead of a single branching one (+ remove mention) 14:27:49 Created ACTION-142 - Create 3/4 questionnaires instead of a single branching one (+ remove mention) [on Stephan Zednik - due 2012-11-17]. 14:27:51 q? 14:28:04 ack ho 14:28:19 q? 14:28:26 hook: clarofy what is meant by publisher(?) in this context 14:28:33 I added PROV-O to http://www.w3.org/2011/prov/wiki/ProvImplementations#Publishers 14:28:50 paul: anyone who creates provenance that appears somewhere on the web. (Following SKOS?) 14:29:02 q+ to ask if prov-o's prov-o is in "Publishers" like Curt 14:29:14 ack gk 14:29:14 GK, you wanted to ask if common questions across all questionnaiure tyopes can be auto-filled 14:29:22 q+ to ask about translating answers to the questionnaire to the exit criteria 14:29:54 stephan: don't know how it can be done 14:29:59 ack lebo 14:29:59 lebot, you wanted to ask if prov-o's prov-o is in "Publishers" like Curt 14:30:50 q? 14:30:53 Tim: Does the provenance in PROV-O the document count as publishing 14:30:58 q- 14:31:45 Ivan: possible add provenance statement in ReSpec … that would be an implementation, also every published spec 14:31:49 q? 14:32:24 Dong: mapping answers from questionnaire to CR exit criteria 14:32:44 q_ 14:32:46 q+ 14:32:51 ack don 14:32:51 Dong, you wanted to ask about translating answers to the questionnaire to the exit criteria 14:32:58 … need two implementations each feature. Can they be vocabs, or apps that consume/produce ? 14:33:42 Paul: we need *pairs* of impl; vocabs count toward coverage, but not really qualifying as a member of a pair 14:33:51 q? 14:34:01 ack pg 14:34:29 q+ 14:34:52 Paul: we need applications that generate/consume every construct in each serialization 14:35:05 q+ 14:35:37 q+ to say that I think consime/produce pairs for vocab terms - ensures devs agree about how the modelling works 14:36:26 Luc: hear something that bothers me - constraints don't need prodcue/conbsume pairs 14:36:32 ack lu 14:37:13 ack gk 14:37:13 GK, you wanted to say that I think consime/produce pairs for vocab terms - ensures devs agree about how the modelling works 14:37:14 q? 14:37:39 q+ 14:38:20 Paul: my biggst concern. We need to get constraint test cases in order and ready to go. Wouldlike these available before/as we go to CR, before facing the the dragon\\\\\'director 14:38:31 ack iv 14:40:05 Ivan: Director may ask: Why did we not use W3C facilities make the forms; data belongs to Google. Answer may be that form has branching structure ()but we just got rid of that). But data ownership may be a concern. 14:40:32 q+ to ask if it's enough to take a data dump and put it on W3C site 14:40:46 Ivan: some companies may have concerns about giving data to another company 14:40:57 q- 14:41:20 q+ 14:41:47 q? 14:41:59 Ivan: Once data ois stored by Google, it will stay there, can't be removed. But companies (and comany lawyers) will say "no way". 14:42:09 ack ze 14:42:11 but, won't google crawl the w3c-native results that we publish at w3.org? 14:42:46 Paolo has joined #prov 14:42:50 q+ can we have alternative of submitting a spreadsheet based on supplied template? 14:43:02 q? 14:43:22 q+ to ask craig 14:43:23 q+ 14:43:30 Ivan: Lawyers job is to be paranoid 14:43:51 ack pg 14:43:51 pgroth, you wanted to ask craig 14:43:52 +??P4 14:44:08 zakim, ??P4 is me 14:44:08 +Paolo; got it 14:44:18 q? 14:44:28 Paul: suggest consider using WBS. If it's easy, that's preferable, if it's hard we can argue the toss. 14:44:54 zakim, [IPcaller] is me 14:44:54 sorry, zednik, I do not recognize a party named '[IPcaller]' 14:45:14 q? 14:45:52 Paul: I can help with WBS 14:45:55 action: zednik to look at wbs for the implementation questionnaire 14:45:55 Created ACTION-143 - Look at wbs for the implementation questionnaire [on Stephan Zednik - due 2012-11-17]. 14:46:11 Stephan: I'll look. Questionnaire just got simpler. 14:46:15 q? 14:46:33 Hook: concern may be w.r.t. public release of intellectual property. 14:46:38 q+ to say what do sparql/xquery wgs do 14:46:42 q? 14:46:46 ack hoo 14:47:05 q+ to ask if there should be an option for confidential submission 14:47:27 @gk -- results go into public implementation report 14:47:27 q? 14:48:12 q? 14:48:12 q- 14:48:32 ack jc 14:48:32 jcheney, you wanted to say what do sparql/xquery wgs do 14:48:36 q? 14:48:50 Luc: moving on... 14:49:06 Luc: want to get a feel for which features people will implement 14:49:26 … have produced a Google doc to gather information (!) 14:49:34 https://docs.google.com/spreadsheet/ccc?key=0An15kLxkaMA3dEU1RHVFNnBvQTNrdzV1S3ZJd0ZjdFE 14:51:25 Form isn't editable yet... 14:51:34 … it should be now 14:52:38 q+ 14:52:45 POI: tracedTo is now wasInfluencedBy 14:56:50 @zednik -- take a look at the GCIS line in the spreadsheet -- edit as needed 14:57:20 q+ 14:58:16 (people are filling in the document) 14:58:25 q? 14:58:34 ack Paolo 14:58:51 q+ 14:59:07 yes 14:59:24 ack pg 15:01:48 ack zednik 15:02:01 action: Dong to remove reference of prov-json in implementation report, and allow entry for "other serialization" 15:02:01 Created ACTION-144 - Remove reference of prov-json in implementation report, and allow entry for "other serialization" [on Trung Dong Huynh - due 2012-11-17]. 15:02:04 q? 15:03:39 q+ to ask about the eligibility for PROV-JSON only implementations 15:07:07 q? 15:07:52 q? 15:08:04 q+ 15:08:16 ack dong 15:08:16 Dong, you wanted to ask about the eligibility for PROV-JSON only implementations 15:08:37 q+ 15:08:45 General discussion as people look at spreadsheet... 15:08:50 ack pg 15:08:50 ack pgroth 15:09:18 (question from phone): do we have to support one of the specific formats to be included in the report? 15:09:50 q? 15:09:53 Paul: no, we can include "other" than core serializations as evidence of use or/support for prov 15:10:13 q+ 15:10:37 How about NASA? 15:10:39 ack iv 15:10:57 Ivan: implemenations listed so far are essentially from academic sources - not so many commercial implementations. 15:11:14 q? 15:11:15 Paul: we have some 15:11:28 q+ 15:12:25 A few implementations from commercial company are currently listed here http://www.w3.org/2011/prov/wiki/ProvImplementations 15:12:30 ack pg 15:13:10 ack gk 15:13:13 q? 15:14:32 GK: would distinguish implementation for live service from for-academic-paper production 15:14:35 q+ 15:14:52 q+ 15:15:12 Ivan: this might be a useful topic for the questionnaire: is their an intention to support the provenance application beyond a current research project? 15:15:33 Paul: this could be hard to formulate appropriately. 15:15:39 even the grad students developing a prototype always hope that their product will spin off and live on in the long term 15:15:50 q+ 15:15:53 q- 15:16:43 For demonstrating interoperable implementability, intended future deployment isn;t necessarily an issue, IMO 15:16:56 q? 15:16:56 Q+ 15:16:58 Paolo: how public is the list of intended implementations? 15:17:02 Ivan: ity's public 15:17:20 s/ity's/it's/ 15:17:22 the thing that is public is this: http://www.w3.org/2011/prov/wiki/ProvImplementations 15:17:30 q? 15:17:48 Luc: we are not collecting commitments here and now - this is for WG planning, not public. 15:17:49 ack Paolo 15:17:51 ack pao 15:17:53 It's useful to include such information (e.g. future support, live service, etc.) in the report, but what is the impact it has on the exit criteria, I'm wondering 15:17:54 CraigTrim has joined #PROV 15:17:56 ack pgroth 15:18:00 Luc: see link above. 15:18:06 q? 15:18:31 q+ to say I don't think it should be included 15:18:40 ack ze 15:18:55 Stephan: we have a structure for the implementation report; are we happy putting this distinction between research/commercial in the report -m don't want to ask things that don';t go in the report 15:18:56 ack pg 15:18:57 pgroth, you wanted to say I don't think it should be included 15:19:00 q? 15:19:01 Paul: agree, shouldn't ask 15:19:08 q? 15:19:21 @ivan we can battle :-) 15:19:27 Luc: next sub-topic 15:19:41 Constraints 15:19:41 Topic: Constraints Implementation 15:19:42 pgroth: it is an information we should have if the question comes 15:19:59 +q 15:20:08 raises hand 15:20:08 will try but may not have time 15:20:25 Luc: Would be good to knwo who is planning to implement any of the constraints features. "show of hands" to IRC please 15:21:06 q? 15:21:11 ack to 15:21:13 Luc: thius could be intent to produce valid provenance, or to consume/assume/check it 15:21:14 I am planning to pursue the Datalog-based implementation which I started this year, although the extent to which that is possible using that particular framework still needs to be clarified 15:21:38 implementing constraints: perhaps. 15:21:40 q? 15:21:50 Paul: will implement, don;t know if will pass all tests, due to levels of inference needed. 15:21:50 ack pg 15:21:57 ack 15:22:15 q? 15:22:16 ack shows 15:22:17 q? 15:22:23 q+ 15:22:35 ack ze 15:23:05 q? 15:23:06 good question 15:23:18 Stephan: is there a distinction between validator or building a producer of valid prov? JHad assumed implementation must be a validator. True or false? 15:23:32 (Luc checks exit criteria) 15:23:33 so it must be a validator 15:23:38 or something similiar 15:23:39 For each of the test cases defined by the working group, at least two independent implementations pass the tests and claim to conform to the document. 15:23:44 q+ 15:24:17 ack pg 15:25:02 q? 15:25:03 q+ to say that an important part of the constraints spec is that devs can understand it well enough to produce only valid prov 15:25:14 ack gk 15:25:14 GK, you wanted to say that an important part of the constraints spec is that devs can understand it well enough to produce only valid prov 15:25:51 q+ does a implementation of the constraints require consumption + check vs. constraints 15:27:12 @GK audio is breaking up while you are talking 15:27:40 ack ze 15:28:34 q+ to say we implement some constraints 15:28:43 q? 15:28:44 q+ to say there are guidelines we don't / can't easily check 15:28:51 ac pg 15:28:58 ack pg 15:28:58 pgroth, you wanted to say we implement some constraints 15:29:01 I was saying that I think the interop report depends on good will - useful evidence may not necessarily be specific to exit criteria, or help to show uo spec defficiencies. So additional evidence that isn;t explicitly covered by the exit criteria may stil, be useful. 15:29:29 q? 15:29:38 +q 15:29:43 ^^s/or help/but still help/ 15:29:47 -q 15:30:00 +1 for what Paul just said 15:30:08 ack pgroth 15:30:11 (the "one line" thing) 15:30:11 ack jch 15:30:11