Warning:
This wiki has been archived and is now read-only.
Chatlog 2009-12-15
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.
14:57:00 <LukeWM> Present: Souri, SteveH, AndyS, LukeWM, LeeF, MatthewPerry, kasei, KjetilK, dcharbon2, pgearon, bglimm, AlexPassant, Sandro, Orri, chimezie 14:57:00 <LukeWM> Note - SteveH only present on IRC for first few minutes. 15:07:14 <LeeF> Regrets: SteveH, OlivierCorby, AxelP 14:54:49 <kasei> wasn't entirely sure if the entailment terms had been nailed down in that lengthy email thread. 14:55:21 <MatthewPerry> MatthewPerry has joined #sparql 14:55:36 <LeeF> SteveH, we didn't resolve the scope of alias variables last week because we wanted your input - are you happy with the proposal as put in the agenda? 14:55:39 <AndyS> SteveH, custom aggregate syntax? 14:56:11 <SteveH> LeeF, I have no strong feelings on scope, whatever people think is sensible is fine by me 14:56:19 <LeeF> thanks, SteveH 14:56:25 <SteveH> AndyS, someone mentioned optional AS, apparently, I'm firmly against that 14:57:20 <AndyS> SteveH, reason? Comes up in example applications when not over HTTP. 14:57:27 <LeeF> the F2F discussion on custom syntax for aggregates was "Mild opinion in favor of having no keyword or special syntax for custom aggregate functions (LeeF, Axel, SteveH, kasei). pgearon (and observer dajobe) expressed mild preference for a keyword to introduce custom aggregates. Consensus that this should be advice to the editors. " 14:58:10 <LeeF> Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2009-12-15 14:58:20 <LeeF> LeeF has changed the topic to: http://www.w3.org/2009/sparql/wiki/Agenda-2009-12-15 14:58:44 <kasei> Zakim, mute me 14:58:54 <AndyS> It's down to SteveH in the end but this editor would go for no special syntax unless proven to need it in which case AGG(<uri>, ..args..) 14:59:18 <LeeF> AndyS - I think that's the general feeling of the group then, but we'll talk about it in a bit (hopefully not for too long!) 14:59:59 <SteveH> I prefer no special syntax too 15:06:01 <LukeWM> LeeF: talks to people about whether they can make the third F2F 15:06:54 <LeeF> http://www.w3.org/2009/sparql/wiki/Agenda-2009-12-15 15:07:00 <LeeF> topic: admin 15:07:16 <LeeF> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2009-12-08 15:07:41 <LeeF> RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2009-12-08 15:07:49 <LeeF> Next meeting: 2009-12-22 @ 15:00 UK / 10:00 EST 15:08:07 <LukeWM> LeeF: next meeting is going to be next week. 15:08:18 <LeeF> subtopic: F2F3 15:08:23 <LukeWM> LeeF: get documents in order for publication in January, so it's useful to come if you can. 15:08:42 <LeeF> F2F3 will be March 25 & 26, 2010 15:08:50 <LeeF> with 2 locations - one in Cambridge, MA at MIT 15:09:00 <LukeWM> LeeF: F3F march 25, 26 with video conferencing. 15:09:12 <LeeF> one location at Oxford, UK 15:09:41 <LeeF> ACTION: Sandro to work with Birte to figure out video conference facilities for F2F3 15:09:42 <trackbot> Created ACTION-154 - Work with Birte to figure out video conference facilities for F2F3 [on Sandro Hawke - due 2009-12-22]. 15:10:24 <LukeWM> LeeF: comment handling, some questions on the process. 15:10:38 <LeeF> http://www.w3.org/2009/sparql/wiki/Comments 15:10:58 <LukeWM> LeeF: if you're swapped in on a relevant topic, compose response, put it on a link to the wiki on http://www.w3.org/2009/sparql/wiki/Comments, notify mailing list and wait a few days. 15:11:25 <LukeWM> LeeF: if you have a response saying it's OK, send it out, if saying it's not OK, don't send it out. 15:11:38 <LukeWM> LeeF: if no response, bug the chairs. 15:11:46 <LukeWM> LeeF: no liason business? 15:12:06 <LeeF> topic: scope of alias variables 15:12:57 <LukeWM> LeeF: question on the table is can the variable ?SUM be used in other places in the query. 15:13:38 <LukeWM> LeeF: proposal is that they can be used in HAVING and further to the right in the SELECT clause, but they can't come in the query pattern. 15:13:41 <AndyS> q+ 15:13:50 <LeeF> ack AndyS 15:13:56 <LukeWM> Orri: why not in order by/group by? 15:14:13 <AndyS> will check here 15:14:15 <AndyS> sorry 15:14:27 <AndyS> Should be in ORDER 15:14:53 <AndyS> need example for why it's needed. 15:15:04 <AndyS> (this is GROUP BY in same SLECT block 15:15:05 <kasei> I think we can hear Andy :) 15:15:33 <LukeWM> AndyS: you can't GROUP BY COUNT(*) so the scoping issue is the same 15:15:46 <LukeWM> ... because the AS clause introduces new variables. 15:15:57 <LukeWM> AndyS: you can say ORDER BY COUNT(*) though. 15:15:58 <kasei> but grouping by scalar expressions...? 15:16:47 <LukeWM> LeeF: we have open actions on SteveH to do with GROUP BY. 15:17:10 <LukeWM> LeeF: there are questions on expression equivalence etc. and it's all equivalent. 15:17:25 <LukeWM> LeeF: I'd like to resolve this issue without touching the GROUP BY thing. 15:17:30 <AndyS> http://lists.w3.org/Archives/Public/public-rdf-dawg/2009OctDec/0585.html 15:17:31 <LukeWM> AndyS: fine 15:17:51 <LeeF> PROPOSED: Aliased variables can be used in expressions further to the right in the SELECT clause and in the HAVING clause and in the ORDER BY clause; query strings that have an aliased variable elsewhere (e.g. in the query pattern) are malformed queries. 15:18:11 <AndyS> which says no to GROUP and yes to ORDER at that level of the SELECT nexting 15:19:06 <LeeF> SELECT (?x + ?y AS ?sum) ... GROUP BY ?sum 15:19:28 <LukeWM> kasei: maybe this relates to GROUP BY with scalars? I'm worried by the way it's been phrased. 15:19:32 <AndyS> q+ to say it can be done 15:19:38 <LeeF> ack AndyS 15:19:38 <Zakim> AndyS, you wanted to say it can be done 15:19:52 <LukeWM> LeeF:we can revisit the issues with GROUP BY later. 15:20:32 <kasei> +1 15:20:33 <MatthewPerry> +1 15:20:35 <LukeWM> LeeF: order by now, revisit group bys later. 15:20:41 <LukeWM> LeeF: happy? 15:20:45 <bglimm> +1 15:20:48 <LeeF> RESOLVED: Aliased variables can be used in expressions further to the right in the SELECT clause and in the HAVING clause and in the ORDER BY clause; query strings that have an aliased variable elsewhere (e.g. in the query pattern) are malformed queries. 15:20:48 <AndyS> +1 15:20:51 <dcharbon2> +1 15:20:59 <LeeF> topic: custom syntax for aggregates 15:21:07 <LukeWM> LeeF: ... summary 15:21:23 <LukeWM> LeeF: Aggregate functions are different to scalars in a number of ways. 15:21:39 <LukeWM> LeeF: e.g. having, group by. 15:22:00 <LukeWM> LeeF: sometimes you have to know if something is an aggregate function rather than a scalar. 15:22:40 <LukeWM> LeeF: question is, do we want static syntax errors (grammar doesn't allow error conditions)? 15:22:54 <LukeWM> LeeF: no problem for built in aggregates 15:23:44 <LukeWM> LeeF: but challenging when custom functions are introduced. Arbitrary URIs don't allow differentiation between scalars and aggregates. 15:24:51 <LukeWM> LeeF: Either: 1) a new piece of syntax and static syntax errors, or 2) make error conditions where implementations have to look up URIs and work out whether they're aggregates or not. 15:25:20 <Prateek> Prateek has joined #sparql 15:25:27 <Souri> Souri has joined #sparql 15:25:39 <LukeWM> LeeF: what do you all think? It seems there's a mild leaning towards not having custom syntax. 15:26:07 <LukeWM> LeeF: pgearon seemed pro custom syntax at F2F 15:26:44 <LukeWM> pgearon: happy to go with whatever the group goes with but. 15:26:45 <LukeWM> pgearon: was shooting from the hip earlier, but like to have a way of differentiating between what's built in and what's an extension. 15:27:23 <LukeWM> LeeF: we can still still differentiate between extensions and built in 15:27:43 <LukeWM> pgearon: it could be confusing to people 15:28:04 <LukeWM> LeeF: anyone else want special syntax for custom aggregates? 15:28:52 <LeeF> PROPOSED: Custom aggregate functions are invoked by URIs with no special syntax 15:29:13 <AndyS> seconded 15:29:42 <bglimm> I abstain, I just don't know enough about the consequenes 15:29:49 <bglimm> I go with the group too 15:29:57 <Souri> +0 15:30:01 <LeeF> RESOLVED: Custom aggregate functions are invoked by URIs with no special syntax, no objections or abstentions 15:30:07 <LukeWM> LeeF: are you OK with this pgearon? 15:30:16 <pgearon> yes 15:30:20 <LukeWM> pgearon: Yes, I'll go with the group 15:30:26 <LeeF> topic: publication 15:30:45 <LukeWM> LeeF: had 2 goals. 15:31:10 <LukeWM> LeeF: 1) Resolve issues. 15:31:20 <LukeWM> LeeF: 2) change information in the document. 15:31:39 <LukeWM> LeeF: 3) Make sure reviewers are given a heads up. 15:31:48 <chimezie> change information in the document = change log at bottom or something more? 15:32:19 <LukeWM> LeeF: being aggressive, and would like reviews in first week of January. 15:33:01 <LukeWM> LeeF: changing information in the document means looking at major changes since last time we published. 15:33:39 <LukeWM> LeeF: hopefully publish middle/second half of january. 15:33:39 <LeeF> subtopic: query 15:33:43 <LeeF> http://www.w3.org/2009/sparql/docs/query-1.1/ 15:34:00 <LukeWM> LeeF: where are we on 1), 2), 3) on query? 15:34:28 <LukeWM> AndyS: all but one issue is resolved, the one left is EXISTS or MINUS which won't be resolved with this publication. 15:35:12 <LukeWM> AndyS: Need to put exists material in the next few days. 15:35:12 <LeeF> Query reviewers: Birte, Matt Souri, Axel 15:35:30 <LukeWM> AndyS: not sure what SteveH's status is. 15:35:44 <LukeWM> LeeF: can you give the go ahead to the reviewers in the next week or so? 15:36:01 <bglimm> Is query 1.1 now integrated with query 1.0? 15:36:04 <AndyS> See http://www.w3.org/2009/sparql/wiki/WG-Documents 15:36:17 <LukeWM> AndyS: probably a bit stretched on Steve's end but don't want to put words in his mouth. 15:36:39 <AndyS> http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml 15:36:47 <LeeF> are we using http://www.w3.org/2009/sparql/docs/query-1.1/ or http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml? 15:37:06 <LukeWM> AndyS: haven't got the formatting sorted yet. 15:39:08 <LeeF> Right now, look at http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml for the latest 15:39:27 <LukeWM> Souri: rq25.xml is the latest? 15:39:28 <LeeF> subtopic: Update 15:39:29 <LeeF> http://www.w3.org/2009/sparql/docs/update-1.1/ 15:39:31 <LukeWM> AndyS: yes 15:40:01 <AndyS> ACTION: Andy: Move Overview.html out of the way 15:40:02 <trackbot> Created ACTION-155 - Move Overview.html out of the way [on Andy Seaborne - due 2009-12-22]. 15:40:40 <LukeWM> pgearon: The document links issues, do we keep them? 15:41:05 <LukeWM> LeeF: Could go either way. It would be nice to keep them in the document. 15:41:22 <LukeWM> pgearon: lots of the issues are no longer relevant but not formally resolved. 15:41:32 <LukeWM> LeeF: feel free to remove where appropriate. 15:41:51 <LeeF> Update reviewers: AndyS, Axel 15:42:21 <LukeWM> LeeF: is it going to be ready for reviewers soon? 15:42:27 <LukeWM> pgearon: at least by the weekend? 15:42:48 <LukeWM> LeeF: any issues, just bring them up with the chairs, and we'll give teleconference time to them. 15:42:51 <LeeF> subtopic: protocol 15:42:52 <LeeF> http://www.w3.org/2009/sparql/docs/protocol-1.1/ 15:43:41 <LukeWM> LeeF: david, we should schedule something. 15:43:49 <LukeWM> LeeF: friday? 15:44:02 <LukeWM> dcharbon2: OK 15:44:15 <LukeWM> LeeF & dcharbon2 coordinate a meeting. 15:44:20 <LeeF> Leef & dcharbon2 will meet 11am on Friday to sort through protocol document 15:44:48 <LeeF> subtopic: service description 15:44:48 <LeeF> http://www.w3.org/2009/sparql/docs/service-description-1.1/ 15:44:51 <LukeWM> LeeF: david & I will send out plan for protocol on the mailing list based on Friday's meeting. 15:45:16 <LukeWM> kasei: there are a couple of things to get in this week 15:45:40 <LukeWM> kasei: one of them is allowing the dataset to have a different name for the graph than the graph's IRI. 15:46:19 <LukeWM> kasei, I didn't understand something, should I just pull from emails? 15:46:32 <LukeWM> bglimm: fine by me, not sure about ivanH. 15:46:53 <LeeF> SD reviwers: david, Alex, Axel 15:46:53 <LukeWM> kasei: will have it done in the next couple of days and then people can look at it. 15:47:08 <bglimm> I'll also have a quick look at SDs 15:47:10 <LeeF> subtopic: http rdf update 15:47:10 <LeeF> http://www.w3.org/2009/sparql/docs/http-rdf-update/ 15:47:31 <LeeF> reviewers for rdf-update: andyS, LeeF, Axel 15:47:33 <chimezie> http://www.w3.org/2009/sparql/wiki/CommentResponse:ldodds-http-update#Response 15:48:02 <LeeF> http://www.w3.org/2009/sparql/docs/http-rdf-update/ 15:48:32 <LukeWM> chimezie, added a link to the general idea of a use case for adding a user to a list. 15:49:00 <AndyS> q+ to asks about graph naming 15:49:02 <LukeWM> chimezie: items in red are not issues as far as I'm concerned, are they issues according to the working group? 15:49:17 <LukeWM> chimezie: we should probably assume RDFXML as a payload type. 15:49:52 <LeeF> ack AndyS 15:49:52 <Zakim> AndyS, you wanted to asks about graph naming 15:49:57 <LukeWM> LeeF: in the interests of time, lets leave the red notes in because we might not formally be able to resolve them before publication. 15:50:45 <Souri> q+ 15:50:51 <LukeWM> AndyS: we had a discussion in the working group regarding graph namin. 15:50:52 <LeeF> ack Souri 15:51:16 <LukeWM> AndyS: we should probably talk about it. 15:51:19 <LukeWM> chimezie: OK 15:51:34 <LukeWM> Souri: ntriples will be allowed too, right? 15:51:40 <LukeWM> chimezie: yes 15:51:53 <LukeWM> LeeF: what do you need to do before reviewers can look at the document. 15:52:15 <LukeWM> LeeF: ? 15:52:27 <LukeWM> chimezie: look at the things that have been talked about today. Will take a couple of days. 15:52:31 <LeeF> subtopic: property paths 15:52:32 <LeeF> http://www.w3.org/2009/sparql/docs/property-paths/Overview.xml 15:52:47 <LukeWM> AndyS: I asked reviewers about publishing as is. 15:53:07 <LukeWM> AndyS: as it's the first time, I think getting something out is the most important thing. 15:53:17 <LukeWM> AndyS: would like reviewers to make comments on that position. 15:53:38 <LeeF> subtopic: entailment 15:53:40 <LeeF> http://www.w3.org/2009/sparql/docs/entailment/xmlspec.xml 15:53:46 <LukeWM> LeeF: my opinion is that it's a good idea to let the world know what we're doing on this. 15:54:17 <LukeWM> LeeF: what would you need to do bglimm before it is ready for review 15:54:18 <LeeF> entailment regimes review: ivanh, Axel 15:54:37 <LukeWM> bglimm: should be ready in a day. 15:55:19 <LeeF> close as no longer valid: 15:55:19 <LeeF> http://www.w3.org/2009/sparql/track/actions/19 15:55:19 <LeeF> close as completed: 15:55:19 <LeeF> http://www.w3.org/2009/sparql/track/actions/82 15:55:19 <LeeF> http://www.w3.org/2009/sparql/track/actions/89 15:55:20 <LeeF> http://www.w3.org/2009/sparql/track/actions/117 15:55:22 <LeeF> http://www.w3.org/2009/sparql/track/actions/142 15:56:04 <LukeWM> LeeF: action 19 no longer valid. 15:56:17 <LukeWM> LeeF: action 82 has been dealt with today. 15:56:27 <LukeWM> LeeF: 89, no longer relevant. 15:56:34 <LukeWM> LeeF: 117, greg has done this. 15:56:44 <LeeF> trackbot, close ACTION-19 15:56:44 <trackbot> ACTION-19 Send to the mailing list a few example cases (data, query, results) of SELECT queries in FILTERs closed 15:56:46 <LeeF> trackbot, close ACTION-82 15:56:46 <trackbot> ACTION-82 Start thread on mailing list re: ISSUE-36 closed 15:56:47 <LeeF> trackbot, close ACTION-89 15:56:47 <trackbot> ACTION-89 Go through issues list, figure out status, which need to be discussed in TCs, which have pending actions, which can be resolved closed 15:56:49 <LeeF> trackbot, close ACTION-117 15:56:50 <trackbot> ACTION-117 Incorporate an example for extensibility by the end of the week into SD closed 15:56:51 <LeeF> trackbot, close ACTION-142 15:56:54 <trackbot> ACTION-142 Summarise changes of SD and conclusions from last TC along with open questions. closed 15:56:55 <LukeWM> LeeF: 142 is complete, anyone thing they aren't done? 15:56:56 <bglimm> ACTION-153 is also completed 15:57:05 <bglimm> that's changelog 15:57:14 <LeeF> trackbot, close ACTION-153 15:57:14 <trackbot> ACTION-153 Provide a high-level changelog summarising what has happened since FPWD in entailment closed 15:58:01 <LukeWM> LeeF: happy holiday