Warning:
This wiki has been archived and is now read-only.
Chatlog 2011-04-26
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.
<kasei> Present: bglimm, kasei, cbuilara, SteveH, AxelPolleres, MattPerry, pgearon, OlivierCorby, AndyS, chimezie, Lee_Feigenbaum 13:55:40 <RRSAgent> RRSAgent has joined #sparql 13:55:40 <RRSAgent> logging to http://www.w3.org/2011/04/26-sparql-irc 13:55:49 <SteveH_> SteveH_ has joined #sparql 13:55:53 <AxelPolleres> Zakim, this will be SPARQL 13:55:53 <Zakim> ok, AxelPolleres; I see SW_(SPARQL)10:00AM scheduled to start in 5 minutes 13:55:59 <bglimm> ok, then I'll dial in... no Easter holidays... 13:56:01 <AxelPolleres> agenda: http://lists.w3.org/Archives/Public/public-rdf-dawg/2011AprJun/0108.html 13:56:08 <AxelPolleres> chair: Axel Polleres 13:57:07 <AxelPolleres> regrets: Alex 13:57:21 <Zakim> SW_(SPARQL)10:00AM has now started 13:57:30 <Zakim> + +44.186.528.aaaa 13:57:31 <cbuilara> cbuilara has joined #sparql 13:57:39 <bglimm> Zakim, +44.186.528.aaaa is me 13:57:39 <Zakim> +bglimm; got it 13:57:40 <Zakim> +??P3 13:57:55 <Zakim> +kasei 13:57:59 <Zakim> +[IPcaller] 13:58:06 <cbuilara> zakim, IPcaller is me 13:58:07 <Zakim> +cbuilara; got it 13:58:24 <Zakim> -??P3 13:58:46 <MattPerry> MattPerry has joined #sparql 13:58:50 <Zakim> +??P3 13:58:54 <SteveH> Zakim, ??P3 is me 13:58:54 <Zakim> +SteveH; got it 13:59:09 <Zakim> +AxelPolleres 13:59:32 <Zakim> +MattPerry 14:01:04 <Zakim> +pgearon 14:01:06 <Zakim> +OlivierCorby 14:01:16 <chimezie> chimezie has joined #sparql 14:01:21 <AxelPolleres> tried to quickly paste things we shall discuss on the wiki: http://www.w3.org/2009/sparql/wiki/Agenda-2011-04-26 14:01:28 <AxelPolleres> agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2011-04-26 14:02:04 <AxelPolleres> let's give it another minute... 14:02:26 <OlivierCorby> pass code is rejected when calling number in France (04.26.46.79.03) 14:03:31 <Zakim> +??P14 14:03:37 <AndyS> zakim, ??P14 is me 14:03:37 <Zakim> +AndyS; got it 14:03:47 <chimezie> chimezie has joined #sparql 14:04:12 <kasei> who is scribing? 14:04:31 <kasei> I can 14:04:36 <kasei> scribenick: kasei 14:04:40 <Zakim> + +1.216.368.aabb 14:04:49 <AxelPolleres> scribe: Greg Williams 14:04:49 <chimezie> Zakim, +1.216.368.aabb is me 14:04:49 <Zakim> +chimezie; got it 14:04:55 <AxelPolleres> topic: admin 14:05:21 <AxelPolleres> PROPOSED: approve minutes form last time http://www.w3.org/2009/sparql/meeting/2011-04-19 14:05:44 <AxelPolleres> RESOLVED: approve minutes form last time http://www.w3.org/2009/sparql/meeting/2011-04-19 14:06:15 <AxelPolleres> topic: document status 14:06:45 <kasei> AxelPolleres: plan is to go to last call for many of the documents next week. 14:06:59 <kasei> ... query, update, service description, graph store protocol, entailment, federated query <kasei> subtopic: query 14:07:38 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/To_Last_Call 14:07:38 <AndyS> http://www.w3.org/2009/sparql/wiki/To_Last_Call#Query 14:07:40 <kasei> SteveH: done all the major things. still some tidying up to be done. 14:07:45 <kasei> ... in reasonable shape. 14:08:03 <kasei> AndyS: I've got comments form Birte to deal with and one comment that needs a reply. 14:08:13 <bglimm> q+ to ask about (NOT) EXISTS FILTER 14:08:15 <kasei> AxelPolleres: is there anything preventing LC next week? is that realistic? 14:08:39 <kasei> AndyS: realistic to make a formal decision. there may be some things outstanding. 14:08:54 <kasei> AxelPolleres: probably need some final approval from reviewers. 14:09:21 <kasei> bglimm: I had a look at the algebra section and still not happy with (NOT) EXISTS 14:09:30 <kasei> ... I can't see how that is supposed to work. 14:09:54 <kasei> AndyS: going to deal with Birte's comment. 14:10:31 <AxelPolleres> Query summary: Birte's comments and one public comment open ? 14:10:38 <Zakim> +Lee_Feigenbaum 14:10:56 <kasei> AxelPolleres: let us know if there are any other issues. otherwise will hopefully move forward next time. 14:10:58 <AxelPolleres> q? 14:11:04 <AxelPolleres> ack birte 14:11:09 <bglimm> Zakim, ack me 14:11:09 <Zakim> unmuting bglimm 14:11:11 <Zakim> bglimm, you wanted to ask about (NOT) EXISTS FILTER 14:11:11 <AxelPolleres> ack bglimm 14:11:12 <Zakim> I see no one on the speaker queue 14:11:20 <AndyS> bglimm, what is "expire"? s//exists/? <kasei> subtopic: update 14:11:42 <pgearon> There seem to be a couple of open questions, particularly around the formal semantics 14:11:47 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/To_Last_Call#Update 14:12:09 <kasei> pgearon: most of the questions are around the formal semantics. 14:12:17 <kasei> ... all outstanding editorial tasks have been done. 14:12:29 <kasei> ... a couple of questions have come up lately that haven't been addressed. 14:12:43 <kasei> AxelPolleres: just sent a mail trying to address greg's comments. 14:12:51 <kasei> ... still a list of open things. 14:13:13 <kasei> ... open issue: semantics of USING concerning blank nodes. 14:13:26 <kasei> ... up to Paul and myself to make a proposal. 14:13:29 <kasei> ... this is the most critical. 14:13:44 <kasei> ... other things are mostly editorial. 14:13:50 <AndyS> And the descriptive text of USING ("identicial" is not acceptable) 14:13:53 <kasei> ... no "glue" between syntax and semantics yet. 14:14:15 <kasei> ... we don't have semantics of multi-operation requests. 14:14:34 <kasei> ... hope to address this by next week, but not realistic to make a decision yet. 14:14:42 <LeeF> q+ to note publication moratorium 14:15:07 <LeeF> We need to have our publication requests ready by May 6, or else we slip until late May - http://lists.w3.org/Archives/Member/chairs/2011JanMar/0001.html 14:15:10 <LeeF> ack me 14:15:10 <Zakim> LeeF, you wanted to note publication moratorium 14:15:36 <kasei> kasei: mostly happy with the response to my comments and how they are being handled. 14:15:38 <LeeF> It's a W3C publication moratorium surrounding the W3C AC Meeting 14:15:44 <AndyS> q+ 14:16:09 <kasei> AxelPolleres: will try to have things ready by Friday, and hope Andy and Greg can have another look. 14:16:24 <LeeF> I think fully ready :-( 14:16:27 <kasei> AndyS: we've got to get through the pubrules. Does this moratorium mean it has to be ready by then? Or just asked for publication by then? 14:16:28 <LeeF> But sandro would know best 14:16:42 <LeeF> I'll see if I cna ask someone else while we continue the call 14:16:57 <kasei> AndyS: might be able to turn this into a virtue. there's still the issue with the RDF WG about changes in turtle. 14:17:06 <kasei> ... if we want to keep turtle and sparql aligned, then more time would be good. 14:17:17 <AndyS> ack me 14:17:20 <kasei> ... have asked that they address the issue, but haven't seen it on an agenda yet. 14:17:22 <LeeF> I will bring that up at CG call ASAP as well (again) 14:17:36 <kasei> AxelPolleres: are there any other issues from the RDF WG that will affect our work? 14:17:49 <LeeF> AndyS, if there's a statement of intent, do we have any other changes we'd need to make? 14:17:51 <kasei> AndyS: worst outcome would be a long discussion for each difference. 14:18:16 <kasei> AxelPolleres: should try to get as ready as we can by next week. 14:18:26 <kasei> ... will clarify with sandro what is needed by the 6th. 14:18:46 <AxelPolleres> ACTION: axel to check with sandro about what's needed for publication by May 6th 14:18:46 <trackbot> Created ACTION-443 - Check with sandro about what's needed for publication by May 6th [on Axel Polleres - due 2011-05-03]. 14:19:04 <kasei> AxelPolleres: will ask editors to check pubrules by next time. 14:19:44 <kasei> ... if you think it's realistic ot be ready by next week, is it possible to have text that can be reviewed by the end of this week? 14:19:50 <LeeF> wedding? 14:19:57 <kasei> AndyS: no. 4-day weekend in UK this week. 14:21:10 <kasei> AxelPolleres: I can propose text for open update issues, but need help for checkign pubrules, etc. 14:21:17 <kasei> pgearon: I can commit to helping with that. 14:21:34 <AxelPolleres> paul and axel to chat later on skype <kasei> subtopic: service description 14:22:11 <bglimm> I wrote an email on that 14:22:55 <AxelPolleres> q? 14:23:41 <kasei> http://lists.w3.org/Archives/Public/public-rdf-dawg/2011AprJun/0102.html 14:23:57 <kasei> bglimm: I suggested one way of handling the entailment issue 14:24:11 <kasei> q+ 14:24:39 <kasei> bglimm: at the moment, you can specify a default regime and a regime per named graph. 14:24:54 <kasei> ... suggested that domain of the entailment regime property should be Graph, not NamedGraph. 14:25:17 <kasei> ... also, right now Profiles are attached to graphs. 14:25:18 <AxelPolleres> a) renaming graph property, b) profiles assigned to service not to graphs 14:27:26 <AxelPolleres> as for b) greg's sees a problem, possibly every profile attached to any regime. 14:27:51 <kasei> AxelPolleres: didn't we already agree that entailment regimes are assigned to graphs? 14:28:18 <kasei> bglimm: yes, but if you support owl direct semantics and rdf-based semantics, saying "OWL RL" isn't enough, because RL can be used in either Direct or RDF-based semantics. 14:28:29 <kasei> ... does RL apply to both semantics if you calim RL is a supported profile? 14:28:40 <kasei> ... at the moment, there's no way of distinguishing. 14:28:57 <kasei> ... the semantics and the profile belong together. 14:29:36 <kasei> ... it's a bit weird to assign the profile to graphs. a graph doesn't "have a profile". 14:30:02 <kasei> AxelPolleres: is there any action here? does anybody else have a strong opinion? 14:30:13 <kasei> bglimm: taking it to email is probably best. 14:30:15 <chimezie> it might help (me) if we had an example, perhaps 14:30:24 <AxelPolleres> Birte and Greg to discuss a path forward 14:30:36 <bglimm> I added an example (of an SD) to my email 14:30:55 <AxelPolleres> ACTION: Birte to report on outcome of entailment and SD discussion on email 14:30:55 <trackbot> Created ACTION-444 - Report on outcome of entailment and SD discussion on email [on Birte Glimm - due 2011-05-03]. 14:30:56 <chimezie> ok, thanks (I missed that) <kasei> subtopic: graph store protocol 14:31:31 <kasei> chimezie: no substantive outstanding issues. 14:31:40 <kasei> ... just sent out a draft response to KjetilK. 14:31:53 <kasei> ... nothing else other than editorial thigns like pubrules checks 14:32:02 <kasei> ... ready for LC by next week. 14:32:23 <AxelPolleres> ACTION: chime to check pubrules on graph protocol doc 14:32:23 <trackbot> Created ACTION-445 - Check pubrules on graph protocol doc [on Chimezie Ogbuji - due 2011-05-03]. 14:32:45 <AxelPolleres> ACTION: greg to check pubrules for SD 14:32:45 <trackbot> Created ACTION-446 - Check pubrules for SD [on Gregory Williams - due 2011-05-03]. 14:33:06 <AxelPolleres> ACTION: paul to do pubrules check for Updates 14:33:06 <trackbot> Created ACTION-447 - Do pubrules check for Updates [on Paul Gearon - due 2011-05-03]. 14:33:37 <AxelPolleres> ACTION: steve to check pubrules on query 14:33:37 <trackbot> Created ACTION-448 - Check pubrules on query [on Steve Harris - due 2011-05-03]. <kasei> subtopic: entailment 14:33:46 <bglimm> wiki page for that: http://www.w3.org/2009/sparql/wiki/Pub-Process 14:34:01 <kasei> bglimm: still waiting for replies. mainly editorial issues. 14:34:24 <kasei> ... could take silence as agreement. 14:34:48 <kasei> AxelPolleres: I suggest you try one more time. Take no response by the end of the week as agreement. 14:35:04 <AxelPolleres> ACTION: Birte to ping jeff one more time to ask for confirmaiton of rewordings, otherwise takes silence as agreement 14:35:05 <trackbot> Created ACTION-449 - Ping jeff one more time to ask for confirmaiton of rewordings, otherwise takes silence as agreement [on Birte Glimm - due 2011-05-03]. 14:35:11 <kasei> AxelPolleres: anything else open? 14:35:15 <kasei> bglimm: no, that's the last thing. 14:35:30 <kasei> AxelPolleres: one small thing from sandro about the RIF in RDF document. 14:35:45 <kasei> ... I assume it can wait until after LC. Just need that published before rec. 14:35:52 <kasei> ... not a roadblock to LC. 14:36:15 <kasei> bglimm: issue on property paths and entailment. 14:36:45 <kasei> AxelPolleres: what's written in the entailment doc is that entailment works at the BGP level (only extension point we have) 14:36:58 <kasei> ... property paths are not affected by entailment regime. 14:37:11 <kasei> ... property paths are first simplified, then matched. 14:37:22 <kasei> ... arbitrary/zero length paths are different. 14:37:42 <kasei> ... some cases where you would expect entailment to work on arbitrary paths. 14:38:11 <kasei> ... worried that we've locked ourselves in on any future entailment work working at this level. 14:38:24 <AndyS> Arbitrary length paths have cardinality=1 makes entailment future easier. (speculation) 14:38:32 <kasei> bglimm: could say that for queries that can't be simplified into BGPs, behaviour is not defined. leaves open for future versions. 14:38:54 <kasei> ... if you use entailment, only some property paths are guaranteed to work. others aren't specified. 14:39:22 <kasei> AxelPolleres: is this section informative? 14:39:31 <kasei> bglimm: yes. 14:39:59 <kasei> AxelPolleres: would feel better if that was made clearer. 14:40:17 <kasei> bglimm: I can do that today. 14:40:46 <kasei> AxelPolleres: would it make sense to add a statement that future work may address this issue? 14:41:03 <kasei> AndyS: how can you make it clearer than "informative" in the title? 14:41:11 <LeeF> <blink> 14:41:20 <kasei> LeeF++ 14:41:22 <LeeF> ...or unicode snowmen 14:41:36 <kasei> AxelPolleres: want to point out the problematic cases. 14:42:05 <Zakim> -bglimm 14:43:05 <AxelPolleres> informative probably enough, but admittedle not entirely happy with it. 14:43:19 <AxelPolleres> take that back to email. 14:43:28 <Zakim> +bglimm 14:43:34 <bglimm> Sorry I dropped out 14:44:07 <kasei> bglimm: will try to make the text clearer. 14:44:28 <AxelPolleres> axel: show what's the limits would be good <kasei> subtopic: federated query 14:44:59 <kasei> cbuilara: waiting for comments. 14:45:12 <LeeF> I'm hoping to get my comments by tonight 14:45:18 <kasei> AxelPolleres: some questions on BINDINGS section? 14:45:45 <kasei> ... not sure if I'll be able to look through it, but will check what cbuilara and LeeF discuss. 14:45:50 <kasei> ... apart form that, ready for LC? 14:46:05 <kasei> cbuilara: yes. I've applied previous comments. 14:46:25 <kasei> ... depends on comments being waited on, but think it can go. 14:46:38 <AxelPolleres> ACTION: carlos to check pubrules 14:46:38 <trackbot> Created ACTION-450 - Check pubrules [on Carlos Buil Aranda - due 2011-05-03]. 14:47:02 <bglimm> Yes 14:47:12 <AxelPolleres> ACTION: birte to check pubrules on entailment 14:47:12 <trackbot> Created ACTION-451 - Check pubrules on entailment [on Birte Glimm - due 2011-05-03]. 14:47:46 <kasei> AxelPolleres: that's it for the documents. 14:48:21 <kasei> ... do we have dataset-merge definition still in query? 14:48:23 <kasei> AndyS: yes. 14:48:42 <kasei> AxelPolleres: where do we go with that? I put in an alternative version in the update document. 14:48:48 <kasei> ... not sure if that addresses Peter's comment. 14:49:08 <kasei> AndyS: I did some editing, new wording. 14:49:28 <kasei> ... nervous about describing a dataset as a bunch of slots. 14:50:16 <kasei> AxelPolleres: is the RDF WG going to take that up? 14:50:40 <kasei> AndyS: don't know. It's there because if there's going to be a distinguished version of dataset-merge, makes sense to have it next to the definition for dataset. 14:51:03 <kasei> AndyS: don't have a picture of what the possible outcomes are. 14:51:46 <kasei> AxelPolleres: should we leave it in optionally? at risk? 14:52:05 <kasei> AndyS: just leave it in. 14:52:27 <kasei> ... does update need a merge? 14:52:45 <kasei> AxelPolleres: not really sure. worried if we describe the LOAD operation in terms of dataset UNION, have to say something extra about blank nodes. 14:52:53 <kasei> ... if I use dataset merge, then I don't. 14:53:08 <kasei> AndyS: LOAD is only on a single graph. so renaming can happen when the graph is read. it's coming from syntax. 14:53:50 <kasei> AxelPolleres: if we don't need union, would also work for me. 14:54:34 <kasei> AndyS: the only problem is if the syntax introduces bnodes with clashing labels. 14:54:42 <kasei> ... somewhere it should say "don't". 14:55:10 <kasei> AxelPolleres: probably don't need merge in update. 14:55:41 <kasei> AxelPolleres: don't know how to keep the blank nodes adopting the definitions from query. 14:55:47 <kasei> ... so came up with the skolemization text. 14:56:32 <kasei> AxelPolleres: we'll get to BINDING tests next week. 14:56:52 <kasei> cbuilara: did tests for BINDINGS only, not yet SERVICE. 14:57:10 <kasei> AxelPolleres: I had an open action for federated tests. If you could suggest something, that would be great. 14:58:05 <AxelPolleres> ACTION: Carlos to think about Federated query testing (essentially helping on ACTION-274) 14:58:05 <trackbot> Created ACTION-452 - Think about Federated query testing (essentially helping on ACTION-274) [on Carlos Buil Aranda - due 2011-05-03]. # SPECIAL MARKER FOR CHATSYNC. DO NOT EDIT THIS LINE OR BELOW. SRCLINESUSED=00000293