From SPARQL Working Group
Revision as of 16:13, 31 January 2012 by Mperry
Please justify/explain all edits to this page, in your "edit summary" text.
15:00:53 <RRSAgent> RRSAgent has joined #sparql 15:00:53 <RRSAgent> logging to http://www.w3.org/2012/01/31-sparql-irc 15:00:54 <swh> Zakim, this will be SPARQL 15:00:55 <trackbot> RRSAgent, make logs world 15:00:55 <Zakim> Zakim has joined #sparql 15:00:57 <trackbot> Zakim, this will be 77277 15:00:57 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start now 15:00:58 <trackbot> Meeting: SPARQL Working Group Teleconference 15:00:58 <trackbot> Date: 31 January 2012 15:01:06 <axelpolleres> I have updated the agenda with more links at: http://www.w3.org/2009/sparql/wiki/Agenda-2012-01-31 15:01:07 <AndyS> zakim, who is on the phone? 15:01:07 <Zakim> SW_(SPARQL)10:00AM has not yet started, AndyS 15:01:09 <Zakim> On IRC I see RRSAgent, chimezie, axelpolleres, MattPerry, swh, MacTed, LeeF, ouvasam, SteveH, AndyS, iv_an_ru, ya, NickH, pgearon, kasei, trackbot, ericP, sandro 15:01:19 <AndyS> zakim, this is 77277 15:01:19 <Zakim> ok, AndyS; that matches SW_(SPARQL)10:00AM 15:01:25 <AndyS> zakim, who is on the phone? 15:01:25 <Zakim> On the phone I see kasei, MattPerry, ??P10, +49.897.aaaa, sandro 15:01:27 <bglimm> bglimm has joined #sparql 15:01:32 <AndyS> zakim, ??P10 is me 15:01:32 <Zakim> +AndyS; got it 15:01:36 <Zakim> +??P14 15:01:40 <swh> Zakim, ??P14 is me 15:01:40 <Zakim> +swh; got it 15:02:16 <axelpolleres> Zakim, who is on the phone? 15:02:17 <Zakim> On the phone I see kasei, MattPerry, AndyS, +49.897.aaaa, sandro, swh 15:02:18 <Zakim> +??P19 15:02:30 <bglimm> Zakim, ??P19 is me 15:02:30 <Zakim> +bglimm; got it 15:02:34 <axelpolleres> Zakim, aaaa is me 15:02:35 <Zakim> +axelpolleres; got it 15:02:38 <bglimm> Zakim, mute me 15:02:38 <Zakim> bglimm should now be muted 15:02:51 <axelpolleres> Zakim, who is on the phone? 15:02:51 <Zakim> On the phone I see kasei, MattPerry, AndyS, axelpolleres, sandro, swh, bglimm (muted) 15:02:54 <cbuilara> cbuilara has joined #sparql 15:03:06 <chimezie> chimezie has joined #sparql 15:03:13 <Zakim> +??P21 15:03:20 <cbuilara> zakim, ??P21 is me 15:03:20 <Zakim> +cbuilara; got it 15:03:22 <axelpolleres> Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2012-01-31 15:03:35 <axelpolleres> regrets: Nicholas, Alex 15:03:44 <axelpolleres> chair: Axel Polleres 15:04:01 <chimezie> chimezie has joined #sparql 15:04:10 <axelpolleres> scribe: MattPerry 15:04:32 <axelpolleres> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2012-01-24 15:04:34 <MattPerry> topic: admin 15:04:39 <Zakim> +chimezie 15:04:40 <ouvasam> ouvasam has left #sparql 15:05:00 <axelpolleres> RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2012-01-24 15:05:36 <axelpolleres> Next regular meeting: 2012-02-07 @ 15:00 UK / 10:00 EST (scribe: Alex or Nicholas) 15:05:52 <MattPerry> axel: anything from the RDF working group 15:06:02 <MattPerry> AndyS: nothing to report 15:06:15 <axelpolleres> topic: Graph Store Protocol 15:06:38 <axelpolleres> http://www.w3.org/2009/sparql/wiki/Agenda-2012-01-31 15:06:49 <Zakim> +pgearon 15:07:00 <axelpolleres> Resolutions 2 and 3 from: http://www.w3.org/2009/sparql/meeting/2012-01-10#Graph_Store_Protocol 15:07:06 <axelpolleres> Consensi from last week: http://www.w3.org/2009/sparql/meeting/2012-01-17#line0051 15:08:07 <MattPerry> chimezie: resolutions apply only to SPARQL 1.1. graph stores, what about things that are not graph stores? 15:08:46 <MattPerry> ... several things were still being discussed in email 15:09:22 <axelpolleres> Zaki, who is on the phone? 15:09:24 <axelpolleres> q? 15:09:35 <MattPerry> ... BASE URI resolution and mt-uri 15:10:12 <axelpolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2012Jan/0008.html 15:10:23 <axelpolleres> (comment PF-1 ) 15:10:45 <axelpolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Nov/0048.html 15:10:52 <axelpolleres> (comment JL-2) 15:11:13 <axelpolleres> response before from sandro... will use similar wording. 15:12:05 <AndyS> http://www.w3.org/2009/sparql/wiki/CommentResponse:AL-1 15:12:05 <axelpolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Jun/0000.html 15:12:09 <axelpolleres> (AL-1) 15:12:33 <MattPerry> axel: do you need approval for that response 15:12:47 <AndyS> DB-3 is editorial - response can be sent. 15:13:29 <MattPerry> chimezie: response mentions linked data protocol question about how to incorporate this into graph store protocol 15:13:38 <AndyS> SJ-1? The response looks ready to go (and is OK). 15:13:53 <MattPerry> axel: we need to fix that before publishing the document 15:14:06 <axelpolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011May/0033.html 15:15:30 <AndyS> q+ to talk (at some point) about base, append. Not comment related. 15:16:17 <MattPerry> axe: let's get DB3 and SJ1 out of the way 15:16:26 <MattPerry> s/axe/axel 15:16:35 <MattPerry> axel: go ahead with DB3 15:16:39 <MattPerry> ... anyone looked at SJ1 15:17:29 <MattPerry> ... from myside SJ1 looks ok as well 15:17:34 <chimezie> http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011May/0033.html 15:17:37 <chimezie> SJ-1 15:17:44 <AndyS> http://www.w3.org/2009/sparql/wiki/CommentResponse:SJ-1 15:17:57 <axelpolleres> sandro: case spelling? 15:18:11 <Zakim> +LeeF 15:18:20 <MattPerry> sandro: looks ok to me 15:18:39 <MattPerry> axel: anything we can do about AL1? 15:18:51 <MattPerry> s/AL1/AL-1 15:19:38 <MattPerry> axel: this is not included in the draft, right? 15:20:14 <MattPerry> chimezie: heard a few different suggestions, don't have a good idea what the consensus of the WG is 15:20:22 <LeeF> I think the WG stands that if the doc is scoped to graph store, then we're OK 15:20:27 <LeeF> right, what Sandro says 15:20:32 <MattPerry> sandro: the scope of the document was the way to handle this 15:20:48 <MattPerry> ... POST append is find if this only applies to SPARQL 1.1 graph stores 15:20:55 <MattPerry> s/find/fine 15:21:17 <AndyS> q? 15:21:24 <MattPerry> chimezie: I can add some text to say what the scope is 15:22:12 <MattPerry> AndyS: I think what we need is a restriction to a certain CASE (i.e. posting to graph rather than to a container) 15:22:21 <AndyS> ackme 15:22:24 <AndyS> ack me 15:22:24 <Zakim> AndyS, you wanted to talk (at some point) about base, append. Not comment related. 15:23:19 <MattPerry> axel: anything to discuss related to summarizing consensus from last telecon? 15:23:42 <axelpolleres> http://www.w3.org/2009/sparql/meeting/2012-01-17#line0051 15:24:59 <MattPerry> chimezie: I understand wording changes for PATCH 15:26:06 <LeeF> _not necessarily_ 15:26:09 <LeeF> right 15:26:20 <MattPerry> axel: graph store protocol is not the same as the SPARQL endpoint URI 15:26:48 <MattPerry> sandro: service IRI is used as the indirect identifier 15:27:07 <axelpolleres> s/is not/does not need to be/ 15:27:34 <LeeF> 2 weeks ago :) 15:27:37 <MattPerry> ... confusion on this issue is a problem 15:28:09 <MattPerry> axel: this needs to more explicit in the document 15:29:04 <MattPerry> chimezie: one question was what is the relationship between the service for the graph store and the service for the SPARQL protocol 15:29:30 <sandro> q? 15:29:31 <MattPerry> ... the other question is how do we describe/discover the graph store service 15:30:12 <MattPerry> sandro: for question 1, the SPARQL endpiont URI is not necessarily related to the graph store service URI 15:30:30 <MattPerry> ... a lot of people may think they are the same 15:31:28 <MattPerry> sandro: does this get rid of the notion of a servcie URI? 15:31:35 <MattPerry> chimezie: yes 15:31:56 <MattPerry> sandro: if you do a get on the service, you get a service description 15:32:09 <MattPerry> chimezie: all service description references will be removed 15:32:17 <sandro> (in the old text, to go away) 15:32:25 <MattPerry> ... there will be a few lines about how to address the graph store 15:32:56 <MattPerry> axel: when do you think these changes will be implemented? 15:32:59 <sandro> chime: Planning to remove the notion of Service IRI, and just use the GraphStore IRI. 15:33:02 <MattPerry> ... first we need another review 15:33:05 <sandro> sandro: sounds good 15:33:28 <MattPerry> chimezie: probably Thursday 15:33:38 <axelpolleres> ACTION: chime to send a mail to the group when final wording on GSP is in place 15:33:39 <trackbot> Created ACTION-582 - Send a mail to the group when final wording on GSP is in place [on Chimezie Ogbuji - due 2012-02-07]. 15:34:08 <axelpolleres> ACTION: sandro to review GSP upon completion of ACTION-582 15:34:08 <trackbot> Created ACTION-583 - Review GSP upon completion of ACTION-582 [on Sandro Hawke - due 2012-02-07]. 15:34:19 <MattPerry> axel: second reviewer? 15:34:46 <axelpolleres> Zakim, pick a victim 15:34:46 <Zakim> Not knowing who is chairing or who scribed recently, I propose bglimm (muted) 15:34:54 <bglimm> Zakim, unmute me 15:34:54 <Zakim> bglimm should no longer be muted 15:35:08 <MattPerry> bglimm: not sure I'm really qualified to review it 15:35:19 <MattPerry> axel: I think it will be ok 15:35:29 <MattPerry> bglimm: I can do a quick review 15:35:39 <bglimm> Zakim, mute me 15:35:39 <Zakim> bglimm should now be muted 15:35:51 <axelpolleres> ACTION: birte to review GSP upon completion of ACTION-582 15:35:51 <trackbot> Created ACTION-584 - Review GSP upon completion of ACTION-582 [on Birte Glimm - due 2012-02-07]. 15:36:27 <MattPerry> topic: Test Cases 15:36:33 <LeeF> Right, the ones that are passed by ARQ and RDF::Query, in particular 15:36:56 <MattPerry> axel: looking to approve those that are passed by 2 implementations 15:37:00 <axelpolleres> http://www.w3.org/2009/sparql/implementations/ 15:37:03 <kasei> not represented on that page, the federation tests all pass on both ARQ and RDF::Query. 15:37:14 <MattPerry> axel: any updates to that list? 15:37:44 <MattPerry> kasei: test status are current, but implementation status has changed a bit 15:37:57 <LeeF> :-) 15:38:11 <MattPerry> AndyS: couple of small details that were changed in the test 15:38:20 <MattPerry> .... just fixing mistakes 15:38:35 <MattPerry> kasei: those changes look fine to me 15:38:46 <MattPerry> ... would be good if someone else looks at them 15:38:58 <MattPerry> LeeF: they look fine to me, but I have not looked in detail 15:39:27 <kasei> if we have time, I'd like to discuss the COPY/MOVE tests that I emailed the list about. 15:39:40 <axelpolleres> PROPOSED: approve all non-approved test cases at http://www.w3.org/2009/sparql/implementations/#sparql11-federated-query 15:39:58 <LeeF> Seconded. 15:39:58 <kasei> +1 15:40:03 <AndyS> +1 15:40:03 <axelpolleres> +1 15:40:06 <cbuilara> +1 15:40:18 <axelpolleres> RESOLVED: approve all non-approved test cases at http://www.w3.org/2009/sparql/implementations/#sparql11-federated-query 15:40:45 <AndyS> (cbuilara? - were the changes OK with you?) 15:40:56 <MattPerry> axel: also have 100% coverage on the query language tests 15:41:04 <LeeF> big kudos to Andy and Greg as shining examples for the rest of the SPARQL implementation world! 15:41:18 <axelpolleres> PROPOSED: approve all non-approved test cases at http://www.w3.org/2009/sparql/implementations/#sparql11-query 15:41:22 <LeeF> seconded 15:41:33 <AndyS> +1 15:41:38 <MattPerry> +1 15:41:44 <kasei> +1 15:42:05 <axelpolleres> RESOLVED: approve all non-approved test cases at http://www.w3.org/2009/sparql/implementations/#sparql11-query 15:42:22 <kasei> q+ 15:42:30 <MattPerry> axel: Update Tests: still don't have 100% coverage 15:42:43 <LeeF> ack kasei 15:43:10 <MattPerry> kasei: found some significant problems with the update shortcuts 15:43:45 <MattPerry> ... the text says COPY/MOVE equivelent to e.g. DROP + MOVE, but is this really the case? 15:43:52 <LeeF> I've reviewed Greg's notes and agree that the definitions are broken as-is. 15:44:11 <MattPerry> s/DROP + MOVE/DROP + ADD 15:44:15 <AndyS> Suggestion - let's not approve the update tests today, discuss on email, and move on to the easy ones we can pass (results formats). Need Paul as update editor. 15:44:35 <AndyS> The tests are excellent in finding this. 15:45:00 <MattPerry> kasei: copy graph to itself intuitively should be a no-op but strictly following the spec gets rid of the graph 15:45:00 <axelpolleres> COPY/MOVE definitions need to be fixed in terms of calling them shortcuts 15:45:20 <pgearon> I agree with kasei on the problems with the language inconsistency 15:46:17 <MattPerry> pgearon: The language inconsistency should be fixed 15:46:53 <chimezie> Zakim, mute me 15:46:53 <Zakim> chimezie should now be muted 15:47:04 <MattPerry> pgearon: I'd be inclined to use the equivalent operations and say if it's the same graph it becomes a no-op 15:47:30 <MattPerry> kasei: still have the issue of 1 operation vs a set of operations 15:47:38 <AndyS> q+ 15:47:58 <MattPerry> ... are the shortcuts atomic? 15:48:19 <MattPerry> LeeF: greg, would you prefer atomic or not? 15:48:26 <axelpolleres> ACTION: Paul to suggest some fixing of the wording regarding COPY/MOVE equivalent operations (reagrding atomicity and remarking that some are no-op) and about failure on empty graphs 15:48:26 <trackbot> Created ACTION-585 - Suggest some fixing of the wording regarding COPY/MOVE equivalent operations (reagrding atomicity and remarking that some are no-op) and about failure on empty graphs [on Paul Gearon - due 2012-02-07]. 15:48:59 <MattPerry> kasei: I'd be happy to not require the shortcuts but don't think this is an option 15:49:49 <MattPerry> AndyS: the word atomic is used in 2 different places 15:50:04 <MattPerry> .. for update requests and in the formal definitions 15:50:17 <MattPerry> ... I don't think it needs to be in the formal section 15:50:37 <axelpolleres> "By 'atomic operation' we mean that the operation performs the described transformation of the Graph Store either completely or leaves the Graph Store unchanged, i.e., the result is either GS' or GS (in case of error)." 15:51:15 <MattPerry> ... If we remove atomic from the formal section, this may fix the problem 15:51:50 <LeeF> kasei, if the word "atomic" were removed from the formal definition of an operation, would you be OK with that? 15:52:15 <kasei> LeeF: the word? or that whole sentence? 15:52:41 <MattPerry> AndyS: if there is no reference to SILENT in the formal section, then error handling is outside of the formal definition 15:52:49 <kasei> s/LeeF:/LeeF,/ 15:53:00 <LeeF> the word, I guess? 15:53:07 <LeeF> But I guess the whole sentence 15:53:08 <LeeF> now that I read it 15:53:09 <LeeF> :) 15:53:43 <MattPerry> LeeF: removing "atomic" removes a requirement 15:53:47 <kasei> q+ 15:53:54 <LeeF> ack AndyS 15:54:07 <MattPerry> AndyS: you could say these operations describe the before and after state of the graph store 15:54:11 <pgearon> +q 15:54:54 <MattPerry> axel: I think you're right that we can drop it 15:55:06 <axelpolleres> q? 15:55:16 <AndyS> ack kasei 15:55:23 <MattPerry> kasei: I'm worried that we are going to require code changes 15:55:34 <LeeF> It's a bug fix 15:55:36 <MattPerry> ... for people that used the equivalent expanded form 15:56:14 <axelpolleres> Agreement to remive "atomic" from http://www.w3.org/TR/sparql11-update/#def_updateoperation? 15:57:19 <MattPerry> s/remive/remove 15:57:53 <LeeF> ack pgearon 15:58:13 <MattPerry> pgearon: I'm concerned about a non-atomic implementation during concurrent queries 15:58:21 <MattPerry> ... is this ok? 15:59:07 <MattPerry> AndyS: is it the overall request that should be atomic? 15:59:26 <MattPerry> pgearon: what about while the update is processing 15:59:52 <AndyS> ping - formats tests? I can stay for +10 mins. 16:00:29 <axelpolleres> ACTION-585, Paul to remove "atomic" from http://www.w3.org/TR/sparql11-update/#def_updateoperation 16:00:29 <trackbot> ACTION-585 Suggest some fixing of the wording regarding COPY/MOVE equivalent operations (reagrding atomicity and remarking that some are no-op) and about failure on empty graphs notes added 16:01:28 <LeeF> want to discuss prop path comments next time - saw some IRC discussion - would be good to have it on mailing list 16:01:31 <Zakim> -LeeF 16:01:48 <Zakim> -chimezie 16:01:49 <axelpolleres> adjourned 16:01:52 <Zakim> -MattPerry 16:01:53 <bglimm> bze 16:01:54 <Zakim> -sandro 16:01:56 <bglimm> bye 16:01:59 <Zakim> -swh 16:02:04 <Zakim> -AndyS 16:02:06 <Zakim> -bglimm 16:02:08 <Zakim> -cbuilara 16:02:12 <axelpolleres> rrsagent, makre records public 16:02:12 <RRSAgent> I'm logging. I don't understand 'makre records public', axelpolleres. Try /msg RRSAgent help 16:02:22 <axelpolleres> rrsagent, make records public 16:02:33 <Zakim> -axelpolleres 16:02:35 <Zakim> -kasei 16:03:37 <Zakim> -pgearon 16:03:38 <Zakim> SW_(SPARQL)10:00AM has ended 16:03:44 <Zakim> Attendees were kasei, MattPerry, +49.897.aaaa, sandro, AndyS, swh, bglimm, axelpolleres, cbuilara, chimezie, pgearon, LeeF # SPECIAL MARKER FOR CHATSYNC. DO NOT EDIT THIS LINE OR BELOW. SRCLINESUSED=00000253