13:56:59 RRSAgent has joined #sparql 13:56:59 logging to http://www.w3.org/2012/10/02-sparql-irc 13:57:01 RRSAgent, make logs world 13:57:01 Zakim has joined #sparql 13:57:03 Zakim, this will be 77277 13:57:03 ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 3 minutes 13:57:04 Meeting: SPARQL Working Group Teleconference 13:57:04 Date: 02 October 2012 13:57:12 zakim, this will be SPARQL 13:57:12 ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 3 minutes 13:57:14 Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2012-10-02 13:57:16 Chair: LeeF 13:57:24 zakim, code? 13:57:24 the conference code is 77277 (tel:+1.617.761.6200 sip:zakim@voip.w3.org), LeeF 13:57:56 SW_(SPARQL)10:00AM has now started 13:58:03 + +1.310.729.aaaa 13:58:08 +EricP 13:58:11 Zakim, aaaa is me 13:58:11 +kasei; got it 13:58:23 + +1.617.553.aabb 13:58:26 zakim, aabb is me 13:58:26 +LeeF; got it 13:59:10 MattPerry has joined #sparql 13:59:30 + +33.4.92.96.aacc 14:00:01 zakim, aacc is me 14:00:01 +Olivier; got it 14:00:04 + +1.603.897.aadd 14:00:13 +[IPcaller] 14:00:19 +??P7 14:00:19 zakim, aadd is me 14:00:20 +MattPerry; got it 14:00:25 zakim, IPcaller is me 14:00:25 +AndyS; got it 14:00:30 AxelPolleres has joined #sparql 14:00:33 Zakim, ??P7 is me 14:00:33 +SteveH; got it 14:00:35 +Sandro 14:00:58 +??P13 14:01:00 zakim, ??P13 is me 14:01:00 +cbuilara; got it 14:01:24 + +49.897.aaee 14:01:40 Zakim, aaee is probably me 14:01:40 +AxelPolleres?; got it 14:01:51 +pgearon 14:02:12 chimezie has joined #sparql 14:02:18 Zakim, who is on the phone? 14:02:18 On the phone I see kasei, EricP, LeeF, Olivier, MattPerry, AndyS, SteveH, Sandro, cbuilara, AxelPolleres?, pgearon 14:02:22 bglimm has joined #sparql 14:02:54 +??P19 14:03:06 zakim, who's on the phone? 14:03:06 On the phone I see kasei, EricP, LeeF, Olivier, MattPerry, AndyS, SteveH, Sandro, cbuilara, AxelPolleres?, pgearon, ??P19 14:03:07 Zakim, ??P19 is me 14:03:08 +bglimm; got it 14:03:25 Zakim, it;s really me! 14:03:25 I don't understand 'it;s really me!', AxelPolleres 14:03:27 +Chimezie 14:03:34 Zakim, who is talking? 14:03:45 kasei, listening for 10 seconds I could not identify any sounds 14:03:46 Zakim, mute me 14:03:46 bglimm should now be muted 14:04:13 Zakim, unmute me 14:04:13 bglimm should no longer be muted 14:04:17 greg - where in the test suite are these? 14:04:35 scribe: sandro 14:04:35 scirbenick: sandro 14:04:39 topic: admin 14:04:40 Zakim, mute me 14:04:40 Chimezie should now be muted 14:04:47 ACTION: axel to clean up the scribelist 14:04:47 Created ACTION-689 - Clean up the scribelist [on Axel Polleres - due 2012-10-09]. 14:04:47 PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2012-09-25 14:05:22 +1 14:05:28 RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2012-09-25 14:05:41 AndyS, grepping I see agg tests (avg-02, err-02, min-0[12]), coalesce01, plus-1 and plus-2 14:06:12 lee: no open comments for query, right? 14:06:23 andy: comments page is up to date, and shows nothing open. 14:06:56 lee: tests? 14:06:59 http://www.w3.org/2009/sparql/implementations/ 14:07:12 ericP, I hear plh in the background 14:07:54 lee: Based on the feedback, do any of the query tests need to change? 14:08:15 eric: It seems like the goal is to have the generated values be in the canonical form? 14:08:22 http://www.w3.org/TR/xmlschema11-2/#decimal-lexical-representation 14:08:37 eric: So, looking at this, I see the text..... 14:08:37 q+ 14:08:42 In all cases, leading and trailing zeroes are prohibited subject to the following: there must be at least one digit to the right and to the left of the decimal point which may be a zero. 14:08:57 eric: This text says to me there should be a trailing zero. 14:09:43 AndyS:in the algorithms section, it says if an integer, then use ... map, without decimal point 14:10:21 sandro: Eric, is this about the specs or the test suite? 14:10:24 NB "xmlschema11-2/#decimal-lexical-representation" is different to the email ref of EricP 14:11:02 eric: If the spec allows any valid lex rep, then, no, this isnt about the spec. 14:11:49 I think we define result set equivalence by RDF simple entailment...http://www.w3.org/2009/sparql/docs/tests/README.html#queryevaltests 14:12:08 sandro: I'm proposing this as a post-PR test suite cleanup thing. 14:12:08 ... not more, not less. 14:12:28 eric: I don't see how people can pass the tests like this. 14:12:44 lee: We have impls that *do* pass, which is evidence to the contrary. 14:12:55 eric: What result set do we get? 14:13:29 q+ 14:13:39 ack kasei 14:13:39 eric: Were the tests passed by human intervention? 14:14:11 kasei: Everyone already has that burden, because we're not providing a test harness. 14:14:15 ArthurK has joined #sparql 14:14:48 kasei: We can't assume things here -- we don't control how import is done, or when/how/if canonicallization is done. 14:14:49 q+ 14:15:02 sandro: Is that written down -- that the tests are numeric equivalence. 14:15:06 q- 14:15:11 q+ 14:15:12 lee: Let's take an action to document and move on. 14:15:29 http://www.w3.org/2009/sparql/docs/tests/README.html#howto would be a good place to document this. 14:15:33 + +1.512.651.aaff 14:15:34 eric: I'd like to resolve whether a zero is needed, but we can do that later. 14:15:51 ack AxelPolleres 14:16:05 Zakim, +1.512.651.aaff is me 14:16:05 +ArthurK; got it 14:16:24 ack bglimm 14:16:42 ACTION: Axel and Lee to document expectation of comparing numeric results in test results in http://www.w3.org/2009/sparql/docs/tests/README.html#howto 14:16:42 Created ACTION-690 - And Lee to document expectation of comparing numeric results in test results in http://www.w3.org/2009/sparql/docs/tests/README.html#howto [on Axel Polleres - due 2012-10-09]. 14:17:20 lee: Eric, any other problems you found? 14:17:41 eric: one of the other tests, mentioned in my email, gives a result that's very hard to explain. 14:17:52 eric: graph constraint on outside of subselect. 14:18:43 eric: graph constraint on outside of graph constraint is done via a stack of graph constraints. but for subquery (and maybe service?) is you preserve the graph constraint. 14:18:56 sq03.rq: graph ?g { {select ?x where {?x ?p ?g}} } 14:20:24 lee: I propose this is a corner case, unlikely to come up in practice. you wont have to teach this. I propose to change nothing. 14:20:49 eric: i just wanted to see if this resonates with anyone 14:20:56 lee: you want to remove the test? 14:21:19 eric: we could say the graph constraint, { ... { ... we could have the same behavior. 14:21:29 lee: we're not open to a substantive change at this point. 14:22:25 sandro: this is out of order, too late eric 14:22:33 eric: no, this is implementation feedback. 14:22:44 eric: (argument from email) 14:23:14 lee: personally, I'd object to that change, as unintuitive to me. a subquery inside a graph constraint, I can't see why I'd want anything else. 14:23:22 lee: anyone support eric's proposal 14:23:27 14:23:40 PROPOSED: Publish http://www.w3.org/2009/sparql/docs/query1.1/rq25.xml as PR 14:23:49 +1 14:23:50 I agree that this is a corner case, were there might be different opinions, but the test case clarifies what behaviour the group expects. 14:23:53 +1 14:23:53 +1 14:23:55 +1 14:24:00 +1 14:24:01 +1 ASF 14:24:04 +1 14:24:07 +1 14:24:08 +1 14:24:14 +1 14:24:16 +1 14:24:22 RESOLVED: Publish http://www.w3.org/2009/sparql/docs/query1.1/rq25.xml as PR 14:24:33 +1 14:24:40 topic: Update 14:24:56 RV-10 14:26:40 axel: somewhat arguable results. I proposed test cases yesterday, there were some opinions. 14:26:59 lee: steve and olivier think... andy thinks...... 14:27:18 lee: Did the commenter (Rob) express a preference here? 14:27:20 steve and olivier think it should be different bnodes, and thinks the same 14:27:28 *andy 14:27:30 INSERT DATA { ... _:b1 ... } ; INSERT DATA { ... _:b1 ... } 14:27:33 One bnode or two? 14:27:42 lee: steve and olivier think that multiple inserts using the same blank node label should result in different bnodes. andy thinks it should be the same 14:28:04 Third option - ban it. 14:28:30 sandro: think i agree with Andy, based on discussion in RDF group and resolution that in its multigraph syntax, this would be the same blank node 14:28:52 RV would expect something even "more different" ... "distinct fresh blank nodes will be generated for each usage in each block." 14:28:57 I'm not sure there's a parllell with TriG, that more like within one ; 14:29:03 lee: That RDF-WG argument has some weight for me, yes. 14:29:09 FYI: http://www.w3.org/2011/rdf-wg/track/issues/21 14:29:30 AxelPolleres: Rob (commenter) expects different ones, even in nested blocks. 14:29:40 I agree with Steve. TriG is more like GRAPH :g1 {} GRAPH ?g2 {}, not multiple operations in the same request. 14:29:40 AxelPolleres: Which is not what we want, I think. 14:29:52 SteveH: That's not how I recall it. link to message? 14:30:01 bglimm: Is it because of the semicolon? 14:30:47 lee: I've always thought of blank node labels as document scope. 14:30:56 +1 to strawpoll 14:31:46 q+ to answer on sandro 14:31:51 ack AxelPolleres 14:31:51 AxelPolleres, you wanted to answer on sandro 14:32:11 sando gave the argument that convinced the RDF-WG 14:32:21 sandro: I think the RDF-WG was swayed by the notion of serializing a dataset with shared blank nodes between graphs. 14:32:46 straw poll: do the occurrences of _:b1 in "INSERT DATA { ... _:b1 ... } ; INSERT DATA { ... _:b1 ... }" refer to ONE blank node or TWO blank nodes? 14:32:47 AxelPolleres: This is different -- you can do that, with one request with multiply GRAPH clauses. 14:32:51 sandro: ah, okay. 14:33:05 two 14:33:08 two 14:33:10 one blank - or ban it. 14:33:13 one 14:33:14 one 14:33:15 one 14:33:16 two (slight preference) 14:33:18 one 14:33:19 don't care 14:33:19 2 14:33:38 one or ban it 14:33:44 ban it would be ok 14:33:56 lee: even split. when we had a similar issue on spanning BGPs with entailment semantics, the end result was to ban it. 14:34:08 if you ask me as an editor, I'd switch to "one" (because then I don't have to edit anything 14:34:11 … ot make it implementation defined *ducks* 14:34:12 ;-) 14:34:33 PROPOSED: It's an error to use the same blank node label in two different parts. We'd not test this, just assert it. 14:34:52 s/parts/operations in the same request/ 14:35:20 PROPOSED: It's an error to use the same blank node label in two different operations in the same request. We'd not test this, just assert it. 14:35:28 we're omly sort-of testing for it - the test is a bit strange 14:35:30 AndyS: We're already testing for it. 14:35:43 AndyS: This also re-opens the query document. 14:35:49 I don't think we have an *approved test case* for that, do we? 14:36:14 AndyS: Hmmmm, what was Rob running against.... 14:37:14 RV-10 (http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2012Sep/0022.html) doesn't mention a test case, and the exmple he mentions isn't conflicting with the ban across requests, is it? 14:37:29 lee: let's do that proposal, then unapprove test 53. 14:37:44 we could make it a negative syntax test if we wanted to 14:37:54 lee: the publish updates and query as PR, contingient on this modification. 14:38:07 Zakim, mute me 14:38:07 bglimm should now be muted 14:38:15 lee: not time for a new test. 14:38:52 PROPOSED: It's an error to use the same blank node label in two different operations in the same request. Remove test 53. At some point in the future, add negative syntax test. 14:38:55 Test was added 2012/05/26 14:39:08 +1 14:39:10 unapprove .... go for PR ... add and re-approve negative syntax test in the future. 14:39:28 +1 14:39:30 +1 14:39:32 +1 14:39:32 +1 14:39:33 abstain 14:39:41 abstain 14:39:42 abstain 14:39:42 +1 (also changing unapproved new basic-update tests to negative syntax tests) 14:40:30 +1 14:40:45 +1 14:41:02 RESOLVED: It's an error to use the same blank node label in two different operations in the same request. Remove test 53. At some point in the future, add negative syntax test. 14:41:39 test 53 == syntax-update-1/syntax-update-53.ru 14:41:56 PROPOSED: Publish http://www.w3.org/2009/sparql/docs/query1.1/rq25.xml as PR contingent on the text changes to note that it's an error to use the same blank node label in two different operations in the same request 14:42:14 +1 W3C 14:42:18 +1 14:42:20 +1 ASF 14:42:21 +1 (UUlm) 14:42:21 +1 14:42:21 +1 14:42:23 +1 14:42:23 +1 14:42:23 +1 Experian 14:42:24 how do we check/review the change? 14:42:30 -EricP 14:42:44 +1 (plus action to approval by review) 14:42:57 +1 14:43:39 RESOLVED: Publish http://www.w3.org/2009/sparql/docs/query1.1/rq25.xml as PR contingent on the text changes to note that it's an error to use the same blank node label in two different operations in the same request 14:44:19 sandro has changed the topic to: SPARQL WG : http://www.w3.org/2009/sparql/wiki/Agenda-2012-10-02 14:44:20 ACTION AndyS make change to rq25 to reflect banning use of the same bNode label across operations. Notify WG when done. 14:44:20 Created ACTION-691 - Make change to rq25 to reflect banning use of the same bNode label across operations. Notify WG when done. [on Andy Seaborne - due 2012-10-09]. 14:44:20 ACTION: Axel to approve the change re: banning multiple same blank node label in update requests 14:44:20 Created ACTION-692 - Approve the change re: banning multiple same blank node label in update requests [on Axel Polleres - due 2012-10-09]. 14:45:45 ACTION: Axel to update wording in update document to be consistent wiht change re: blank node labels in query doc 14:45:45 Created ACTION-693 - Update wording in update document to be consistent wiht change re: blank node labels in query doc [on Axel Polleres - due 2012-10-09]. 14:45:48 q+ 14:46:11 AxelPolleres: We need an okay from Rob. 14:46:28 AndyS: What part of Query changes? 14:46:48 LeeF: I think it's just editorial, removing an editorial bit about update scope. 14:46:50 ack me 14:47:42 PROPOSED: Publish http://www.w3.org/2009/sparql/docs/update-1.1/Overview.xml as PR contingent on acknowledgement of response to RV-10 and contingent on ACTION-693 completion 14:47:59 q+ to maybe approve RV-10 just now? http://www.w3.org/2009/sparql/wiki/CommentResponse:RV-10 14:48:21 PROPOSED: Publish http://www.w3.org/2009/sparql/docs/update-1.1/Overview.xml as PR contingent on acknowledgement of response to RV-10 and contingent on ACTION-692 and ACTION-693 completion 14:48:22 AndyS: Continent on 692 as well? 14:50:14 kasei: +1 14:50:19 +1 (siemens) 14:50:22 +1 ASF 14:50:24 +1 14:50:26 +1 (UUlm) 14:50:26 +1 14:50:27 +1 14:50:28 +1 14:50:29 +1 Oracle 14:50:35 +1 14:50:36 +1 14:50:39 RESOLVED: Publish http://www.w3.org/2009/sparql/docs/update-1.1/Overview.xml as PR contingent on acknowledgement of response to RV-10 and contingent on ACTION-692 and ACTION-693 completion 14:52:49 -Chimezie 14:52:52 lee: GSP? later, Chime isn't here. 14:53:01 lee: FedQ ? 14:53:17 +Chimezie 14:53:29 cbuilara: One comment, just a suggestion. 14:53:34 http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2012Jul/0013.html 14:54:27 RV-10 draft response updated at http://www.w3.org/2009/sparql/wiki/CommentResponse:RV-10 (sending contingent to completion of ACTIONS on edits in query and update doc) 14:55:39 lee: Since we need another week anyway, might as well go around with Peter again. Carlos can you tell Peter that given where we are in the WG, we'd prefer to handle this in a future WG? 14:55:46 cbuilara: yes, I'll do that today. 14:56:08 PROPOSED: Publish http://www.w3.org/2009/sparql/docs/json-results/json-results.xml as PR 14:56:12 +1 14:56:35 +1 (siemens) 14:56:36 lee: 3 100% test results 14:56:36 +1 (UUlm) 14:56:40 +1 14:56:41 +1 14:56:45 +1 (Oracle) 14:56:46 +1 Experian 14:56:47 +1 14:56:53 RESOLVED: Publish http://www.w3.org/2009/sparql/docs/json-results/json-results.xml as PR 14:56:59 +1 ASF 14:57:09 +1 Algebraix 14:57:27 PROPOSED: Publish http://www.w3.org/2009/sparql/docs/csv-tsv-results/results-csv-tsv.html as PR 14:57:36 lee: csv/tst -- also three at 100% 14:57:39 +1 14:57:41 +1 (siemens) 14:57:43 +1 14:57:43 +1 14:57:44 +1 (UUlm) 14:57:44 +1 14:57:49 +1 ASF 14:57:50 +1 Algebraix 14:57:52 +1 (Oracle) 14:57:56 +1 Experian 14:58:02 +1 14:58:04 RESOLVED: Publish http://www.w3.org/2009/sparql/docs/csv-tsv-results/results-csv-tsv.html as PR 14:58:32 lee: Anything pending with Overview? 14:58:37 -Chimezie 14:58:43 PROPOSED: Publish http://www.w3.org/2009/sparql/docs/sparql11-overview/Overview.xml as PR 14:58:47 axel: Nope. 14:58:48 +1 14:58:51 +1 14:58:51 +1 (siemens) 14:58:54 +1 Algebraix 14:58:55 +1 (UUlm) 14:58:55 +1 14:58:57 +1 (Oracle) 14:58:59 +1 ASF 14:59:08 RESOLVED: Publish http://www.w3.org/2009/sparql/docs/sparql11-overview/Overview.xml as PR 14:59:08 +1 15:00:26 -LeeF 15:00:29 -cbuilara 15:00:30 sandro: I'll send an email on a little thing with XML Results format 15:00:31 -SteveH 15:00:32 ADJOURN 15:00:32 -kasei 15:00:33 -bglimm 15:00:33 -Olivier 15:00:35 -MattPerry 15:00:41 -pgearon 15:02:20 -ArthurK 15:05:14 http://www.w3.org/TR/rdf-sparql-XMLres/ 15:05:24 http://www.w3.org/2001/sw/DataAccess/XMLres-errata 15:06:19 -AxelPolleres? 15:06:20 -AndyS 15:06:30 -Sandro 15:06:31 SW_(SPARQL)10:00AM has ended 15:06:31 Attendees were +1.310.729.aaaa, EricP, kasei, +1.617.553.aabb, LeeF, +33.4.92.96.aacc, Olivier, +1.603.897.aadd, MattPerry, AndyS, SteveH, Sandro, cbuilara, +49.897.aaee, 15:06:31 ... AxelPolleres?, pgearon, bglimm, Chimezie, ArthurK 16:04:38 Enacting bNode label in update operations "change"/"clarification" : looks like one test failure - not syntax-update-53.ru which is actually OK -- one of Axel's new tests which are not approved. insert-data-same-bnode2.ru 16:56:19 Zakim has left #sparql 17:26:44 MacTed has joined #sparql 18:59:45 AxelPolleres has joined #sparql 19:01:46 AxelPolleres has left #sparql 19:30:23 what's the status of the same-bnode tests. are they all expected to fail now?