14:28:54 RRSAgent has joined #dawg 14:29:08 DanC has changed the topic to: RDF Data Access 20 July 14:29:16 +Kendall_Clark 14:29:18 rob has joined #dawg 14:29:39 +Kevin 14:29:51 +??P27 14:30:13 agenda + Convene, take roll, review agenda http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0114.html 14:30:22 agenda + Review Carlsbad meeting record 14:30:28 +RobS 14:30:29 TomAdams has joined #dawg 14:30:32 agenda + Web Services Directory use cases 14:30:41 agenda + Toward updated Use Cases and Requirements publication 14:30:45 zakim, who is here? 14:30:45 On the phone I see EricP, Yoshio, SteveH, Kendall_Clark, Kevin, ??P27, RobS 14:30:47 On IRC I see TomAdams, rob, RRSAgent, kendall, SimonR, Zakim, SteveH, KevinW, DanC, Yoshio, AndyS_, ericP 14:30:48 agenda + Test case maintenance 14:30:52 + +1.714.539.aaaa 14:31:06 +DanC 14:31:29 zakim, ??P27 is HP 14:31:29 +HP; got it 14:31:49 zakim, HP has AndyS, DaveB 14:31:49 +AndyS, DaveB; got it 14:31:55 Zakim, next agendum 14:31:55 agendum 1. "Convene, take roll, review agenda http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0114.html" taken up [from DanC] 14:32:00 Zakim, who's on the phone? 14:32:00 On the phone I see EricP, Yoshio, SteveH, Kendall_Clark, Kevin, HP, RobS, +1.714.539.aaaa, DanC 14:32:02 HP has AndyS, DaveB 14:32:03 zakim, aaaa is SimonR 14:32:03 +SimonR; got it 14:32:07 Zakim, who's on the phone? 14:32:07 On the phone I see EricP, Yoshio, SteveH, Kendall_Clark, Kevin, HP, RobS, SimonR, DanC 14:32:09 HP has AndyS, DaveB 14:32:38 scribe: rob 14:33:20 agenda + objective 4.2 provenance/data management 14:33:41 agenda + XQuery requirement/objective 14:33:48 +Tom_Adams 14:34:51 next meeting: 27 July. DaveB to scribe. 14:35:03 regrets: Howard Katz, Alberto Reggiori, Enrico Franconi, Jim Hendler, Jos De_Roo 14:35:12 (from dawg list) 14:35:33 ... and Hiroyuki Sato 14:35:36 all actions listed in agenda to be continued without discussion... 14:36:03 Zakim, next agendum 14:36:03 agendum 2. "Review Carlsbad meeting record" taken up [from DanC] 14:36:32 Dan wants to add links to NI objections to meeting record. 14:37:15 PROPOSED: to accept http://www.w3.org/2001/sw/DataAccess/ftf2 1.33 plus minor edits by DanC 14:37:18 support for approving the record as is, with Dan making minor edits over the next week 14:37:21 so RESOLVED. 14:37:59 Zakim, next agendum 14:37:59 agendum 3. "Web Services Directory use cases" taken up [from DanC] 14:38:08 Dan notified semantic web group about XQuery issue, but action ongoing 14:38:17 I can barely hear DanC over the loud typing. 14:38:24 Zakim, mute me 14:38:24 RobS should now be muted 14:38:34 Zakim, unmute me 14:38:34 RobS should no longer be muted 14:39:06 ACTION: RobS write email to Farrukh about traversing a taxonomy, continues. 14:39:12 Zakim, mute me 14:39:12 RobS should now be muted 14:39:15 Zakim, next agendum 14:39:15 agendum 4. "Toward updated Use Cases and Requirements publication" taken up [from DanC] 14:39:22 (I think I'm the loud typist this morning.) 14:40:01 Kendall: 6-8 outstanding changes to make to doc 14:40:36 Kendall: plan to finish by end of tomorrow (Wed) 14:40:53 Kendall: should be reviewed by other members by next Tuesday 14:41:17 changes include all results from f2f 14:41:43 AndyS: date didn't change when rev number changes 14:42:02 Kendall: try to address that; tool problem 14:42:47 would be good to link to meeting minutes suggesting changes in the doc's changelog 14:43:02 kendall: good idea 14:43:21 Zakim, unmute me 14:43:21 RobS should no longer be muted 14:43:29 Zakim, mute me 14:43:29 RobS should now be muted 14:43:58 simon meant to write about CONSTRUCT 14:44:05 kendall: not crucial in short term 14:44:25 sorry, i meant: not crucial for the impending UC&R release 14:44:35 eric to continue... 14:44:52 are we re-wording his action? 14:45:10 eric's work unlikely to be in next publication 14:45:27 ACTION ericP: send federation use case motivating premises (in Algae) 14:45:41 kendall: a few actions could affect UC&R 14:45:56 kendall: should we push back publication? 14:46:15 kendall: Aug 1 was pulled out of his "hat" 14:46:41 AndyS would be fine if provenance isn't in next publication 14:46:49 it will be, fwiw 14:47:32 andys: if waiting 2 more weeks for publication means a couple more use cases, it's worth holding up publication 14:48:04 kendall: ebXML use case is almost ready; disjunction use case needs to be added 14:48:04 kendall, Is Jim H in town this week? 14:48:22 i'm having around 5% loss from occasional 30 second cut-outs 14:48:24 Tom: I think so 14:48:51 kendall, Thanks, I owe him a phone call to discuss the UDDI use case, will follow up this week. Thanks. 14:48:53 kendall: no use case clearly motivates disjunction right now 14:49:12 please also take a look at my comments on wording of disjunction requirement 14:49:12 Dan: no strong preference on publication date 14:49:31 Yoshio: can you paste a pointer to those comments here? 14:49:42 I sent a mail today 14:49:43 Yoshio, is that a request to add that to the agenda for today? 14:49:56 no 14:49:56 yoshio: cool, haven't seen it yet. 14:49:56 Real-world use case for 3.10(a) http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2004Jul/0000.html 14:50:31 it is http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0129.html 14:50:53 who is that? 14:50:59 Tom to take action from Simon 14:51:05 Tom is taking Simon's action; which action was that? 14:51:25 rob: to reply to Chris Wilper's comments on the public-rdf-dawg-comments list 14:51:29 iirc 14:52:27 hmm, Yoshio's comments re: disjunction are worth putting on our agenda here, if possible 14:52:28 Tom: I like the provenance objective 14:52:55 As implementors, I think Tucana should make comments on this. 14:53:11 More specifically, Andy's re-wording 14:53:18 zakim, mute me 14:53:18 Kendall_Clark should now be muted 14:53:23 zakim, unmute me 14:53:23 Kendall_Clark should no longer be muted 14:53:28 Zakim, unmute me 14:53:28 RobS should no longer be muted 14:53:49 Zakim, mute me 14:53:49 RobS should now be muted 14:54:15 rob: I don't like the provenance objective 14:54:41 q+ to point out the named graphs "movement" 14:55:01 As an objective, I do like provenence 14:55:04 (scribe isn't following the Andy-Dan discussion) 14:55:17 q? 14:55:42 Named Graphs, Provenance and Trust 14:55:44 http://www.hpl.hp.com/techreports/2004/HPL-2004-57.html 14:55:49 AndyS has rewritten provenance requirement; we'll put it in the doc and see what people think. 14:55:51 q- 14:55:54 disjunction? 14:56:05 Zakim, unmute me 14:56:06 RobS should no longer be muted 14:56:44 kendall likes Yoshio's points 14:56:45 It is a followup for what I tried to say in the f2f 14:57:07 tried, but failed (^_^;) 14:57:10 Yoshio's comments on disjunction requirement 14:57:11 http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0129.html 14:58:12 Well I'd like everybody to take closer look taking time 14:58:21 Not sure about rewording of requirement; we'll have a bit more email discussion. 14:58:23 Zakim, mute me 14:58:23 RobS should now be muted 14:58:24 ACTION DanC: follow up to http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0129.html 14:59:06 Zakim, unmute me 14:59:06 RobS should no longer be muted 14:59:14 what grows is the results. 15:00:38 Yes 15:00:42 Zakim, mute me 15:00:42 RobS should now be muted 15:00:51 Zakim, next agendum 15:00:51 agendum 5. "Test case maintenance" taken up [from DanC] 15:02:01 Dan: Jos should relay to SteveH feedback on maintenance of tests 15:02:07 (can't hear...) 15:03:24 Steve, Eric, and Dan to coordinate on getting SteveH CVS write access 15:03:32 Zakim, next agendum 15:03:32 agendum 6. "objective 4.2 provenance/data management" taken up [from DanC] 15:03:47 andy's action done; discussion continues 15:03:51 Zakim, next agendum 15:03:51 agendum 6 was just opened, DanC 15:03:56 Zakim, close agendum 6 15:03:56 agendum 6 closed 15:03:57 I see 1 item remaining on the agenda: 15:03:58 7. XQuery requirement/objective [from DanC] 15:03:59 Zakim, next agendum 15:03:59 agendum 6 was just opened, DanC 15:04:42 Zakim, take up agendum 5 15:04:42 agendum 5. "Test case maintenance" taken up [from DanC] 15:04:50 eric: tests have been kept in n3ql doc; it would be good to integrate with Steve's tests 15:05:28 BRQL doc is in DAWG space; it will be edited there 15:05:40 eric: i'm willing to help edit it, if needed 15:06:16 Erk, my calling card is about to drop me out of the teleconference. 15:06:25 -RobS 15:06:28 http://www.w3.org/2001/sw/DataAccess/issues 15:06:31 eek! 15:06:35 -SimonR 15:06:46 shall I adjourn the meeting and we just chat? 15:06:56 +RobS 15:07:02 Wow, that was good timing. :) 15:07:34 ADJOURN. 15:07:34 meeting adjourned. Yippee. 15:08:31 I'm willing to make that a real offer now, FWIW. 15:08:43 -Kevin 15:08:44 KevinW has left #DAWG 15:09:10 http://www.w3.org/2001/sw/DataAccess/BRQL/ 15:09:40 rq23 15:09:49 ql432987 15:11:10 What about just: http://www.w3.org/2001/sw/DataAccess/QL/ ?? 15:11:20 or just ../design 15:11:34 we might have different design docs for protocol 15:11:48 good pt 15:11:57 -RobS 15:12:10 bloody mobile phones... 15:12:44 never mind. How/when does the RSSAgent post a log? 15:13:10 s/RSS/RRS/ 15:13:56 RRSAgent, pointer? 15:13:56 See http://www.w3.org/2004/07/20-dawg-irc#T15-13-56 15:14:27 Does anyone else feel that 4.2 (Provenance -> Support for RDF Aggregation graphs) is beginning to blur into 4.5 (Aggregate query -> union query)? 15:15:18 I might object to both, but if 4.5 can be pure protocol I might find a way to abstain. 15:15:26 there's a connection there; but i don't see any blurring. 15:16:04 re http://www.w3.org/2004/07/08-BRQL/ put "obsolete" with a pointer to the new one under "version"... 15:16:14 danc: re current BRQL spec: no redirect or tombstone 15:16:18 ... and under "status" just say: this was discussed at ... Is this going to be a canonical URI for the QL? 15:17:40 No, presumably not. 15:17:46 ? 15:17:47 canonical in what sense? 15:19:45 Tom: I doubt there can be a canonical URI before we finally name the QL. :) 15:20:14 Yeah, that's why I was suggesting a generic "QL" for query language... 15:21:14 that presumes the "ql" doc will never have any protocol stuff in it. 15:21:23 http://www.w3.org/2001/sw/DataAccess/rq23/ 15:21:25 not really 15:21:35 I think it'd make sense to call it BRQL for now. 15:21:38 Ah, I see. OK. Understood. 15:21:40 it just presumes that the protocol stuff doesn't get 'named' in the URI :> 15:22:06 i like /working-draft-design personally, but i guess that's too guessable or something 15:22:54 On the topic of protocol, I completely flaked out at the F2F and forgot that Kowari has a SOAP interface. That really should've been mentioned during the tail-end protocol discussion. 15:23:07 heh. 15:23:11 -DanC 15:23:15 SimonR: I was gonna raise that but was deferring to you :> 15:23:28 bye bye 15:23:33 -Kendall_Clark 15:23:35 -Yoshio 15:23:56 I didn't write it myself, so naturally it doesn't exist. ;) 15:23:59 Bye! 15:24:14 -Tom_Adams 15:24:16 -EricP 15:24:18 TomAdams has left #dawg 15:24:27 bye 15:24:27 -HP 15:24:31 -SteveH 15:24:32 SW_DAWG()10:30AM has ended 15:24:33 Attendees were EricP, Yoshio, SteveH, Kendall_Clark, Kevin, RobS, +1.714.539.aaaa, DanC, AndyS, DaveB, SimonR, Tom_Adams 15:25:18 SteveH has left #dawg 15:25:53 wow... "No match for disjunction" -- http://foldoc.doc.ic.ac.uk/foldoc/foldoc.cgi?query=disjunction&action=Search 15:32:24 while conjunction has a match ??? 15:34:03 and http://en.wikipedia.org/wiki/Horn_clause 15:55:11 ?! do you happen to think I don't understand the general meaning of "disjunction", Dan? 15:56:21 what I don't think is clear is the notion of "disjunction of graph patterns" 15:56:45 Is it a graph pattern? I don't think so. 15:56:56 Then what is it? 15:57:44 It's a collection of graph patterns, with the definition clarified in the next phrase: at least one of the patterns must match. 15:58:14 If the notion of "matching a graph pattern" makes any sense, then I'd say that matching a disjunction of graph patterns makes sense. 15:59:02 (but it should be noted that I think the whole "graph pattern" terminology is quite convoluted) 15:59:44 I'm trying to figure out what it is you don't understand, yoshio. The phrase "disjunction of graph patterns" is quite a conventional use of the word "disjunction". 16:00:58 I'd like to move away from "graph pattern" but I think it's not all that wacky; it's kinda like "filename patterns" foo.* and such 16:01:41 We more or less agreed that in this case it means "conjunction of triples", with some notion that variables fit in somehow... 16:02:10 yes 16:02:26 It's really the "variable" thing that makes it confusing, but it's not explicit whether variables must be bound, free, or something else. 16:03:35 (The bound-versus-free issue is a concern to me, and even bigger is the problem that our "optional triples" are an entirely different thing.) 16:04:38 well, if the "disjunction of graph patterns" is a collection of graph patterns, then it is not obvious what "restrict matches based on a COLLECTION of graph patterns" means (or it is at least not defined) 16:04:59 yes, "restrict matches" is problematic. 16:05:12 That's why it's clarified as part of the requirement: at least one must match. 16:05:50 in that saying what the "one" is meant for? 16:06:07 a match? 16:06:09 my next meeting's starting; got to go... 16:06:26 ok thanks, talk in email 16:12:15 what's the problem with matching a collection of patterns, yoshio? 16:13:05 results(disjunction(P1, P2, ...), G) = union(results(P1, G), results(P2, G), ...) 16:13:46 I want to say, "That is what to be defined!" 16:14:20 I mean, it is what we want to define EXPLICITLY 16:14:55 but not what is said explicitly in the original wording 16:15:12 yes, it is what is said in that wording, if you take the conventional meaning of the words and phrases. 16:16:03 change "restrict" to "compute", perhaps. 16:16:51 well are you using ",at least one of which must be satisified by the query" part? 16:17:06 s/by the query/by the queried graph/ as we discussed 16:18:03 oh, I didn't put it into my record... I should have done so 16:22:50 then the antecedent of "which" is "the graph patterns", right? 16:26:31 yes 16:28:39 OK, it solves my third question. Now returning to the point of explicity, 16:29:44 what is currently said is "to restrict matches BASED on a disjunction of graph patterns" 16:30:22 I don't think wha we want to define is cleary defined with just saying "BASED on" 16:30:32 right; we're not writing a specification 16:30:34 yet 16:30:45 yes? 16:30:59 ?? 16:31:15 we don't need to completely specify how disjunction works in the requirements document. 16:31:44 eventually, we'll get to that part of the design, and we'll either claim to meet the requirement or change it. 16:34:05 Hmm, the notion made things look new to me (I'm not sure what I want to say...but) 16:35:13 Ok, I understand we don't have to specify HOW (the procedure) it works, but I think we have to specify WHAT is needed 16:37:07 yes, and we have: WHAT is: disjunction. 16:37:24 it might have been easier to just use the one word. 16:37:58 yosi has joined #dawg 16:38:13 yosi has left #dawg 16:38:16 Hmm, but what we have to do is to define what we mean by "disjunction" here? 16:38:51 s/here/here, isn't it?/ 16:38:54 what we mean by disjuction isn't different from the normal dictionary definition, so I don't see why we need to say much. 16:41:57 Hmm, so we didn't share the goal, I wanted to make it clear to the reader what we mean by "disjunction", but you don't (you think leaving it to the conventional reading, or common sence, is Ok) 16:43:04 there's a time and place for being 100% crystal clear what we mean by disjunction, but I don't think it's in the requirements document. The requirements document is more for saying _why_ disjuction is a requirement. 16:43:41 You're right! 16:47:15 OK, I'm convinced. So our next step should be to ensure _why_ each requirement is needed is well said in our document. 16:47:28 yes 16:49:23 I'll leave my personal interest in defining words to other documents. 16:49:49 it's great to have your careful review. 16:50:52 Thank YOU for helping me (re-)notice the aim of the UC&R 16:50:52 RRSAgent, stop