Warning:
This wiki has been archived and is now read-only.
Chatlog 2010-06-01
From SPARQL Working Group
See original RRSAgent log and preview nicely formatted version.
Please justify/explain all edits to this page, in your "edit summary" text.
09:57:00 <LeeF> Chair: LeeF 09:57:00 <LeeF> Regrets: Axel, Alex, Souri, SteveH 09:57:00 <sandro> scribe: sandro 09:57:00 <sandro> rrsagent, make log public 09:57:00 <sandro> meeting: SPARQL Working Group 09:57:00 <Zakim> SW_(SPARQL)10:00AM has now started 09:57:00 <Zakim> +??P4 09:58:00 <Zakim> +??P5 09:58:00 <NicholasHumfrey> zakim, ??P4 is me 09:58:00 <Zakim> +NicholasHumfrey; got it 09:58:00 <Zakim> +Lee_Feigenbaum 09:58:00 <AndyS> zakim, ??P5 is me 09:58:00 <Zakim> +AndyS; got it 09:58:00 <Zakim> +OlivierCorby 09:58:00 <Zakim> +Sandro 09:59:00 <LeeF> zakim, who's on the phone? 09:59:00 <Zakim> On the phone I see NicholasHumfrey, AndyS, Lee_Feigenbaum, OlivierCorby, Sandro 09:59:00 <Zakim> +kasei 10:00:00 <ivan> zakim, dial ivan-voip 10:00:00 <Zakim> ok, ivan; the call is being made 10:00:00 <Zakim> +Ivan 10:00:00 <Zakim> +MattPerry 10:00:00 <Zakim> -MattPerry 10:00:00 <chimezie> Zakim, passcode? 10:00:00 <Zakim> the conference code is 77277 (tel:+1.617.761.6200 tel:+33.4.89.06.34.99 tel:+44.117.370.6152), chimezie 10:01:00 <Zakim> +MattPerry 10:01:00 <Zakim> +Chimezie_Ogbuji 10:01:00 <Zakim> +pgearon 10:02:00 <chimezie> Zakim, mute me 10:02:00 <Zakim> Chimezie_Ogbuji should now be muted 10:02:00 <LeeF> zakim, who's on the phone? 10:02:00 <Zakim> On the phone I see NicholasHumfrey, AndyS, Lee_Feigenbaum, OlivierCorby, Sandro, kasei (muted), Ivan, MattPerry, pgearon, Chimezie_Ogbuji (muted) 10:02:00 <sandro> zakim, list attendees 10:02:00 <Zakim> As of this point the attendees have been NicholasHumfrey, Lee_Feigenbaum, AndyS, OlivierCorby, Sandro, kasei, Ivan, MattPerry, Chimezie_Ogbuji, pgearon 10:03:00 <sandro> lee: not a big agenda today, but let's try to make progress on all our issues 10:03:00 <Zakim> +dcharbon2 10:03:00 <LeeF> Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2010-06-01 10:03:00 <sandro> lee: we should have time for extra matters, if they come up. 10:03:00 <LeeF> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2010-05-25 10:04:00 <LeeF> RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2010-05-25 10:04:00 <LeeF> Next meeting: 2010-06-08 @ 15:00 UK / 10:00 EST (scribe: Steve Harris) 10:04:00 <chimezie> i will be out next tuesday 10:05:00 <Zakim> +??P25 10:06:00 <sandro> lee: I'm hoping to advertise this next round a lot better. we're just a round or two from last call. 10:06:00 <AndyS> +1 to more feedback - last call is *last* call :-) 10:07:00 <sandro> sandro: if someone is only going to review it once, then just last call 10:07:00 <Zakim> + +1.540.412.aaaa 10:08:00 <sandro> lee: I don't want a repeat of DAWG's three last calls! 10:08:00 <Zakim> -pgearon 10:08:00 <pgearon> Zakim, aaaa is me 10:08:00 <Zakim> +pgearon; got it 10:08:00 <sandro> sandro: Okay, then in the publicity now, highlight whatever is likely to be controvercial. 10:08:00 <sandro> lee: yeah, I was going to contact the editors to talk about that. 10:09:00 <LeeF> sandro: AC reps encouraged to give feedback on RIF PR if interested 10:11:00 <AndyS> q+ 10:12:00 <LeeF> ack AndyS 10:13:00 <sandro> sandro: I think linked data has turned the corner and is rolling downhill in the govt space now 10:13:00 <NicholasHumfrey> I will pass things on to other SPARQL users in the BBC 10:13:00 <sandro> lee: Any particular groups we should contact there? 10:13:00 <kasei> I can send the draft pointers to the data.gov group for comments. 10:13:00 <sandro> sandro: I can't think of any right now. I suggest a blog post and careful use of twitter to get the word out, 10:14:00 <sandro> (rdb2rdf stuff missed) 10:15:00 <AndyS> AndyS: maybe data.gov.uk (Government Linked Data Kernel Project) 10:15:00 <sandro> Lee: Dedicate call on something planned, did scheduling poll. 10:15:00 <LeeF> Monday, June 7th 10:15:00 <LeeF> at 10am ET / 3pm UK time 10:15:00 <sandro> topic: # HTTP RDF Update dedicate TC scheduling 10:15:00 <sandro> lee: all interested parties be there. 10:16:00 <AndyS> Is this HTTP update or all update? 10:16:00 <sandro> lee: This is specifically HTTP update, not "all update". 10:17:00 <sandro> topic: Property paths dedicated TC 10:17:00 <sandro> lee: very successful. thanks Andy for the suggestion of these dedicated TCs. 10:17:00 <sandro> lee: we got consensus around all the issues 10:17:00 <LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0285.html 10:18:00 <sandro> lee: I'd like to run through these proposals and get WG approval now. 10:18:00 <Zakim> +bglimm 10:18:00 <LeeF> PROPOSED: The ^ inverse path operator is strictly a unary operator. 10:19:00 <LeeF> . /^ 10:20:00 <AndyS> q+ 10:20:00 <LeeF> ack AndyS 10:20:00 <sandro> sandro: Could this change be adopted back into n3 ? 10:20:00 <LeeF> AndyS: N3 doesn't use forward slash (/) anyway for paths, it uses ! 10:20:00 <LeeF> ... so the differences in syntax are already there 10:21:00 <LeeF> ... N3 uses ^ both as unary reverse operator and as a combining operator 10:21:00 <sandro> andy: n3 uses ! instead of / so there are differences there anyway. n3 uses ^ in a different way, both in paths and in an binary sense, where we're not using it. 10:21:00 <LeeF> ... SPARQL WG prefers single combining syntax (/) and operator (^) for reversing 10:22:00 <kasei> +1 10:22:00 <bglimm> +1 10:22:00 <sandro> +0 10:22:00 <pgearon> abstain 10:22:00 <AndyS> +0 10:22:00 <ivan> 0 10:22:00 <OlivierCorby> +1 10:22:00 <dcharbon2> 0 10:22:00 <MattPerry> +1 10:23:00 <sandro> lee: do we need more +1s than +0s? 10:23:00 <LeeF> RESOLVED: The ^ inverse path operator is strictly a unary operator, with sandro, AndyS, ivan, dcharbon2, pgearon abstaining 10:23:00 <sandro> sandro: no, it's okay 10:23:00 <LeeF> PROPOSED: Property paths do not preserve the order of underlying graph structures (no change to spec). 10:24:00 <sandro> lee: To returns list items in order using property paths would require deep changes to the algebra, etc. 10:25:00 <sandro> ivan: I am disappointed that this is a problem that SPARQL 1.1 cannot solve. 10:25:00 <sandro> lee: We all wish we had a solution. 10:25:00 <AndyS> Needs more than SPARQL to solve! 10:26:00 <MattPerry> +1 10:26:00 <AndyS> +1 10:26:00 <pgearon> +1 10:26:00 <bglimm> +1 10:26:00 <ivan> 0 10:26:00 <ivan> 0 :-( 10:26:00 <sandro> +0 sounds okay, but I don't really understand the issues 10:26:00 <OlivierCorby> +1 10:26:00 <LeeF> RESOLVED: Property paths do not preserve the order of underlying graph structures (no change to spec), ivan, sandro abstaining 10:27:00 <sandro> thanks, pgearon, but the real problem is that I don't have time to think about it. 10:27:00 <LeeF> PROPOSED: Postpone (beyond this WG) any work on returning the length of a matched property path. 10:28:00 <sandro> +1 10:28:00 <bglimm> +1 10:28:00 <pgearon> +1 10:28:00 <AndyS> +1 10:28:00 <MattPerry> +1 10:28:00 <LeeF> RESOLVED: Postpone (beyond this WG) any work on returning the length of a matched property path. 10:28:00 <OlivierCorby> +1 10:28:00 <LeeF> PROPOSED: The cardinality of solutions to fixed-length paths is the same as the cardinality of solutions to the path expanded into 10:28:00 <LeeF> triple patterns (with all variables projected); the cardinality of solutions to variable-length paths is the cardinality of solutions 10:28:00 <LeeF> via paths that do not repeat nodes; the cardinality of solutions to paths combining fixed and variable length (elt{n,} ) is a combination 10:28:00 <LeeF> of the fixed definition plus the variable definition for paths longer than the fixed length. 10:30:00 <sandro> PROPOSED: Cardinality of solutions to property paths is as in http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0285 10:30:00 <AndyS> And see also Birte's http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0293 10:31:00 <sandro> lee: Note this does not affect the answers you get, just the cardinality of the answers. 10:31:00 <LeeF> {n,} 10:32:00 <sandro> +1 10:32:00 <bglimm> So what is the cardinality of query with BGP a r+ ?x ? on graph: a r b. b r c . b r d. ? 2? because the path uses b twice, but it is still a different path for a to c and a to d. 10:32:00 <LeeF> PROPOSED: Cardinality of solutions to property paths is as in http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0285 10:32:00 <bglimm> Zakim, unmute me 10:32:00 <Zakim> bglimm should no longer be muted 10:34:00 <LeeF> b r b. 10:34:00 <sandro> lee: ?x bound to c and d, each soln has card 1. 10:34:00 <sandro> (really, lee, where are you goging? :-) 10:34:00 <sandro> chime: resonsible for detecting cycles and excluding paths that use them 10:35:00 <sandro> lee: yes 10:35:00 <sandro> +1 10:36:00 <LeeF> PROPOSED: Cardinality of solutions to property paths is as in http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0285 10:36:00 <sandro> lee: obviously if new information is unearthed in the future, we can revisit these decisions. But let's move forward now unless someone sees an actual problem. 10:36:00 <AndyS> +1 10:36:00 <bglimm> +1 10:36:00 <OlivierCorby> +1 10:36:00 <dcharbon2> +1 10:36:00 <MattPerry> +1 10:36:00 <pgearon> +1 10:36:00 <LeeF> RESOLVED: Cardinality of solutions to property paths is as in http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0285 10:37:00 <sandro> (it seems to me that one needs to keep track of cycles anyway.) 10:38:00 <LeeF> PROPOSED: Property paths include an operator to negate paths consisting of URIs and reverse URIs only. 10:38:00 <sandro> lee: I missed the discussion of this next one, about negating paths of URIs only -- no negation of +/* 10:38:00 <AndyS> Yes - accurate 10:39:00 <LeeF> PROPOSED: Property paths include an operator to negate paths consisting of URIs and reverse URIs only; more complex paths (such as those including * or +) cannot be negated 10:39:00 <AndyS> Douglas Reid / BBN / http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2010Feb/0000.html 10:39:00 <sandro> +1 10:39:00 <AndyS> +1 10:39:00 <MattPerry> +1 10:39:00 <OlivierCorby> +1 10:39:00 <ivan> +1 10:39:00 <bglimm> 0 (I am scared of path negation) 10:39:00 <dcharbon2> +1 10:39:00 <pgearon> +1 10:39:00 <LeeF> RESOLVED: Property paths include an operator to negate paths consisting of URIs and reverse URIs only; more complex paths (such as those incliuding * or +) cannot be negated, bglimm abstaining 10:40:00 <LeeF> ?s :p{0} ?o 10:40:00 <sandro> lee: is this bound to every node in the data set? every node in the universe? nothing? 10:41:00 <sandro> (I think: every node/pair in the dataset.) 10:41:00 <LeeF> zakim, who's on the phone? 10:41:00 <Zakim> On the phone I see NicholasHumfrey, AndyS, Lee_Feigenbaum, OlivierCorby, Sandro, kasei (muted), Ivan, MattPerry, Chimezie_Ogbuji, dcharbon2, ??P25, pgearon, bglimm (muted) 10:41:00 <sandro> lee: anything else on property paths? no.... 10:42:00 <bglimm> I would say that ?s :p{0} ?o is every pair of nodes in the graph 10:42:00 <sandro> lee: no point in talking about Steve's issue on the WHERE part of sparql update, with SERVICE, since Steve isn't here. 10:43:00 <sandro> right, I think so too, bglimm 10:43:00 <LeeF> http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0267.html 10:43:00 <sandro> lee: anyone have any progress on Steve's query issue here? 10:43:00 <sandro> topic: SPARQL Update and SERVICE (move this up) 10:44:00 <sandro> pgearon: I think it would be fine! 10:44:00 <sandro> pgearon: Since they occur one after the other, it would see the deletion. 10:44:00 <sandro> lee: please reply to Steve on the mailing list. 10:44:00 <sandro> topic: Open Issues 10:45:00 <sandro> http://www.w3.org/2009/sparql/track/issues/open 10:45:00 <ivan> q+ 10:45:00 <LeeF> ack ivan 10:46:00 <bglimm> Zakim, unmute me 10:46:00 <Zakim> bglimm should no longer be muted 10:46:00 <sandro> ivan: should we talk about which URIs to use, for the entailment regimes and for import 10:46:00 <bglimm> http://www.w3.org/ns/entailment/RDF is wrong for describing RDF entailment in SDs? 10:47:00 <bglimm> Zakim, unmute me 10:47:00 <Zakim> bglimm was not muted, bglimm 10:47:00 <sandro> sandro: RIF will review the published version and give its feedback in general, and on the URIs in specific. 10:48:00 <bglimm> The namespace and URI used for rif:imports is still under discussion within the group 10:48:00 <bglimm> Should that entailment URI be http://www.w3.org/ns/entailment/RIF instead, since the choice between Core, BLD, or PLD does not mean a difference in entailment semantics but rather a difference in syntactic subsets of RIF? 10:48:00 <sandro> sandro: if the URIs are reasonable straw proposals, then leave them. if we know they're bad, then use example.org. 10:48:00 <bglimm> These are the 2 ed notes we have for RIF 10:49:00 <sandro> chime: we could use entailment/rif and change the note to clarify 10:49:00 <sandro> ivan: but for import we have nothing in mind. 10:49:00 <sandro> ivan: Okay. 10:51:00 <sandro> lee: so are we changing entailment regimes doc today, or not? 10:51:00 <sandro> chime: Yes, change from rdf-core (rif-core?) to rif 10:52:00 <bglimm> http://www.w3.org/ns/entailment/RDF-RIF-Core is the URI to describe the RIF regime in SDs 10:52:00 <bglimm> That I guess should be changed 10:53:00 <bglimm> Zakim, mute me 10:53:00 <Zakim> bglimm should now be muted 10:53:00 <sandro> sandro: for the webmaster, it seems like we should prepare something like: sparql/round_nnn/WD-sparql-update-20100601/(all the files there) 10:53:00 <AndyS> Can I start editing rq25 again now? 10:54:00 <AndyS> There is an HTML copy anyway. 10:54:00 <sandro> lee: Yes, you can start editing rq25 again. # SPECIAL MARKER FOR CHATSYNC. DO NOT EDIT THIS LINE OR BELOW. SRCLINESUSED=00001141