From SPARQL Working Group
Revision as of 15:04, 8 June 2010 by Apollere2
Please justify/explain all edits to this page, in your "edit summary" text.
13:58:45 <RRSAgent> RRSAgent has joined #sparql 13:58:45 <RRSAgent> logging to http://www.w3.org/2010/06/08-sparql-irc 13:58:52 <AxelPolleres> trackbot, start meeting 13:58:55 <trackbot> RRSAgent, make logs world 13:58:57 <trackbot> Zakim, this will be 77277 13:58:57 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 2 minutes 13:58:58 <trackbot> Meeting: SPARQL Working Group Teleconference 13:58:58 <trackbot> Date: 08 June 2010 14:00:03 <Zakim> SW_(SPARQL)10:00AM has now started 14:00:10 <Zakim> +kasei 14:00:42 <MattPerry> MattPerry has joined #sparql 14:00:49 <Zakim> +pgearon 14:00:56 <Zakim> +AxelPolleres 14:01:19 <AxelPolleres> Zakim, who is on the phone? 14:01:19 <Zakim> On the phone I see kasei, pgearon, AxelPolleres 14:01:20 <Zakim> +[IPcaller] 14:01:26 <AndyS> zakim, IPCaller is me 14:01:26 <Zakim> +AndyS; got it 14:01:35 <Zakim> + +1.603.897.aaaa 14:01:47 <MattPerry> zakim, aaaa is me 14:01:47 <Zakim> +MattPerry; got it 14:01:48 <ivan> zakim, dial ivan-voip 14:01:48 <Zakim> ok, ivan; the call is being made 14:01:49 <Zakim> +Ivan 14:01:54 <AxelPolleres> Zakim, who is on the phone? 14:01:54 <Zakim> On the phone I see kasei, pgearon, AxelPolleres, AndyS, MattPerry, Ivan 14:01:56 <SteveH__> SteveH__ has joined #sparql 14:02:10 <Zakim> + +1.603.897.aabb 14:02:12 <Souri> Souri has joined #sparql 14:02:14 <Zakim> +[Garlik] 14:02:17 <AxelPolleres> agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2010-06-08 14:02:29 <SteveH> Zakim, [Garlik] is temporaily me 14:02:29 <Zakim> I don't understand '[Garlik] is temporaily me', SteveH 14:02:39 <SteveH> Zakim, [Garlik] is temporarily me 14:02:39 <Zakim> +SteveH; got it 14:02:39 <AxelPolleres> scribe: SteveH 14:02:48 <AxelPolleres> chair: AxelPolleres 14:02:52 <Zakim> +Sandro 14:03:03 <Zakim> +??P33 14:03:08 <AxelPolleres> Zakim, who is on the phone? 14:03:08 <Zakim> On the phone I see kasei, pgearon, AxelPolleres, AndyS, MattPerry, Ivan, +1.603.897.aabb, SteveH, Sandro, ??P33 14:03:22 <sandro> sandro has changed the topic to: http://www.w3.org/2009/sparql/wiki/Agenda-2010-06-08 14:03:30 <Souri> zakim, aabb is me 14:03:30 <Zakim> +Souri; got it 14:03:30 <NicholasHumfrey> Zakim, ??P33 is me 14:03:31 <Zakim> +NicholasHumfrey; got it 14:03:33 <SteveH> there's too much noise on the line 14:03:41 <SteveH> Zakim, who's speaking? 14:03:44 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/Agenda-2010-06-08 14:03:51 <Zakim> SteveH, listening for 10 seconds I heard sound from the following: AxelPolleres (8%), Sandro (38%), Ivan (20%) 14:04:05 <sandro> better? 14:04:11 <SteveH> sandro, yes, thanks 14:04:12 <NicholasHumfrey> yes 14:04:23 <AxelPolleres> PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2010-06-01 14:04:40 <SteveH> +1 14:04:43 <NicholasHumfrey> +1 14:04:48 <AxelPolleres> RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2010-06-01 14:05:14 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/Comments 14:05:19 <SteveH> AxelPolleres: all comments are assigned, look at page please 14:05:28 <SteveH> ... some old ones are still open 14:05:37 <Zakim> +[IPcaller] 14:05:43 <LeeF> zakim, IPcaller is me 14:05:43 <Zakim> +LeeF; got it 14:06:19 <SteveH> ... IRI creating comment, I will reply 14:06:33 <SteveH> ... doesn't look too bad, but please update 14:06:58 <AxelPolleres> topic: liaisons 14:07:29 <SteveH> Souri: rdf2rdb has progress, almost ready to publish 14:07:36 <AxelPolleres> q? 14:07:47 <AxelPolleres> ack sandro 14:07:52 <LeeF> (I don't think the RDB group needs a UC&R review from the SPARQL WG) 14:08:13 <SteveH> sandro: rif, today is the last day of the review period 14:08:28 <SteveH> ... egov, working on charter for linked data for govt. working group 14:08:45 <SteveH> ... lots of enthusiasm, put together some education and outreach, plus more serious stuff 14:08:52 <SteveH> ... might produce recommendations 14:09:16 <OlivierCorby> Not able to join the telcon, the phone is out of order here ... 14:09:29 <AxelPolleres> topic: http rdf update TC 14:09:33 <AxelPolleres> http://www.w3.org/2009/sparql/meeting/2010-06-07 14:09:38 <SteveH> AxelPolleres: update minutes 14:10:15 <SteveH> ... no resolutions, but base URI agreement on how it's addressed, proposed text from chime 14:10:31 <SteveH> ... will add example for when there no BASE in the payload 14:10:46 <SteveH> ... resolved terminology change re. networked rdf knowledge 14:11:04 <Souri> UC&R FPWD for RDB2RDF: http://www.w3.org/2001/sw/rdb2rdf/use-cases/ 14:11:09 <AxelPolleres> http://www.w3.org/2009/sparql/track/issues/49 14:11:16 <SteveH> ... discussion about graph URIs er. ISSE-49 14:11:32 <SteveH> ... not sure what to do about URIs which [sth] when you reference them 14:11:52 <AxelPolleres> http://www.w3.org/2001/tag/issues#httpRange-14 14:12:30 <SteveH> ... want to point out that we dont want to discourage, we expect that some might allow HTTP and other protocol 14:12:30 <bglimm> bglimm has joined #sparql 14:12:56 <SteveH> ... want to say that it might lead to ambiguity, give general warning, point to TAG issue 14:13:03 <SteveH> ... how/can we do that? 14:13:20 <SteveH> ... can we reference a TAG issue in a Rec track document? 14:13:26 <Zakim> + +0186528aacc 14:13:33 <bglimm> Zakim, mute me 14:13:33 <Zakim> sorry, bglimm, I do not know which phone connection belongs to you 14:13:38 <AndyS> q+ 14:13:42 <SteveH> ivan: I think the answer is yes, but not sure 14:13:46 <bglimm> Zakim, +0186528aacc is me 14:13:46 <Zakim> +bglimm; got it 14:13:52 <bglimm> Zakim, mute me 14:13:52 <Zakim> bglimm should now be muted 14:14:06 <SteveH> AndyS: the TAG findings aren't recs, what standing do they have 14:14:26 <SteveH> ivan: "TAG findings" - they have no other official standing 14:14:28 <AxelPolleres> Can we give it as an informal reference? 14:14:57 <SteveH> sandro: we can't normatively reference them, but can informatively 14:15:25 <SteveH> ... can't say "to find out how to do this look [at TAG finding]" cos it's not a Rec 14:16:03 <Souri> :-) 14:16:16 <sandro> Basically it's like with Turtle and CURIEs ---- we wanted the concept, but there was no REC, so we had to write it ourselves. 14:16:16 <SteveH> AxelPolleres: either quote text, or reference it informatively 14:16:34 <sandro> Yes, we do. 14:16:36 <AxelPolleres> normative and informative references. 14:17:05 <AxelPolleres> best solution is probably a combination. 14:17:07 <SteveH> AndyS: we take textual forms from Turtle, and include them in the document 14:17:22 <SteveH> AxelPolleres: explain possible ambiguity, and informally reference 14:17:37 <bglimm> But the Wuery spec says: This document uses the http://www.w3.org/TeamSubmission/turtle/ [TURTLE] data format to show each triple explicitly. 14:17:50 <bglimm> s/Wuery/Query/ 14:18:04 <AxelPolleres> ACTION: axel to convey summary to chime: explain possible ambiguity, and informally reference TAG issue. 14:18:04 <trackbot> Created ACTION-255 - Convey summary to chime: explain possible ambiguity, and informally reference TAG issue. [on Axel Polleres - due 2010-06-15]. 14:18:29 <kasei> Zakim, mute me 14:18:29 <Zakim> kasei should now be muted 14:18:37 <SteveH> AxelPolleres: there's a red box in the document, it's enough to drop it 14:18:38 <Souri> s/chime/Chime/ 14:18:49 <SteveH> ... discussion about other terms, and distinctions 14:19:01 <AxelPolleres> unclear "network-manipulable RDF dataset" and distinction from "graph store" and "endpoint" needs clarification 14:19:18 <AndyS> bglimm, that's for the data in examples; not the SPARQL syntax (although that isn't as formal as I'd like) 14:19:21 <SteveH> AxelPolleres: overall doc on good track 14:19:39 <SteveH> Topic: find possible slot for testcases call 14:19:44 <bglimm> Ah, I thought it is for the whole doc including where parts of queries 14:19:49 <AndyS> bglimm, c.f. "4.2 Syntax for Triple Patterns" 14:20:19 <AxelPolleres> http://www.doodle.com/ctb2a6t3bxup8t4d 14:20:26 <bglimm> AndyS, I see what you mean 14:20:41 <SteveH> AxelPolleres: there are several actions on collecting testcases 14:20:51 <AxelPolleres> http://www.w3.org/2009/sparql/track/actions/191 14:21:11 <SteveH> ... would be helpful to have emails pointing to testcases 14:21:27 <SteveH> ... don't need to be formal, just want to track them 14:21:40 <SteveH> ... people are responsible for collecting them 14:22:11 <SteveH> ... david, Olivier, AxelPolleres have outstanding actions 14:22:26 <SteveH> ... make sure that you fill out the poll 14:23:04 <SteveH> Topic: Function Library 14:23:06 <AxelPolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0309.html 14:23:18 <SteveH> AxelPolleres: AndyS has added several functions 14:23:44 <SteveH> ... main question is what's open 14:23:53 <SteveH> ... put some more links about this 14:24:11 <SteveH> ... one question is what about xquery and xpath F&O, we have initial list proposed by AndyS 14:24:11 <AxelPolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg/2009JulSep/0244.html 14:24:25 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/Feature:FunctionLibrary#Starting_Points 14:24:29 <SteveH> ... incorporated in wiki doc ^ 14:24:38 <SteveH> ... and we have one more doc on the wiki 14:24:41 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/Design:FunctionLibrary 14:24:51 <SteveH> ... also half baked 14:25:15 <SteveH> ... do we reuse URIs in F&O? 14:25:58 <SteveH> ... AFAIR, we have had some discussion about namespaces 14:26:16 <SteveH> ... keywords v's namespaced URIs 14:26:30 <SteveH> AndyS: I don't remember that dicsussion 14:26:39 <Zakim> -Sandro 14:26:41 <SteveH> ... there are some functions that dont exsists elsewhere 14:26:55 <SteveH> ... have to assign URIs 14:27:14 <Zakim> +Sandro 14:27:31 <SteveH> AxelPolleres: do we have a namespace already for the sparql F&O 14:27:44 <SteveH> ivan: I did set one up. 14:28:04 <SteveH> AxelPolleres: could look at initial draft 14:28:10 <ivan> http://www.w3.org/ns/sparql.html 14:28:36 <AxelPolleres> that should be fine... 14:28:53 <AxelPolleres> http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#func-coalesce 14:29:10 <SteveH> ivan: it's obvious what you have to do to add more items 14:29:26 <SteveH> AxelPolleres: that would affect all the operators that are new in SPARQL 14:29:34 <SteveH> ivan: yes 14:29:49 <SteveH> AxelPolleres: things in §15.4 need URIs 14:30:10 <SteveH> ivan: I will update doc if you send me changes 14:30:29 <AndyS> For Ivan: http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#SparqlOps 14:30:35 <kasei> ivan, might be nice to assert rdf:type sd:Function for the functions in /ns/sparql 14:30:43 <AxelPolleres> 15.4.14 COALESCE 14:31:42 <AxelPolleres> COALESCE, IF, IN, NOT IN, IRI, URI, BNODE, STRDT, STRLANG 14:31:44 <ivan> q+ 14:31:57 <AxelPolleres> ack AndyS 14:31:58 <AndyS> ack me 14:32:24 <SteveH> ivan: better decided around a table with beer 14:32:31 <SteveH> ... new ones are all caps, old ones are not 14:32:41 <ivan> if, in 14:32:45 <ivan> and not IF and IN 14:32:47 <AndyS> SPARQL is case-insenitive keywords (except for "a") 14:32:48 <SteveH> AxelPolleres: it would make sense to keep with what we had 14:32:52 <kasei> they're all case insensitive in the language, though. 14:33:18 <SteveH> AxelPolleres: it's difficult, for the short names we have camel case 14:33:20 <SteveH> q+ 14:33:26 <ivan> ack ivn 14:33:29 <ivan> ack ivan 14:33:40 <kasei> yes. was just pointing it out because we don't have URIs for the old ones yet 14:34:20 <SteveH> ack me 14:34:23 <AxelPolleres> camelCase seems to be the URI candidates, capital used in the definition 14:34:35 <AndyS> Yes - the grammar uses UC for clarity. Examples use same case as section name (usually) 14:34:35 <SteveH> capital used in grammar/examples 14:34:43 <Zakim> -NicholasHumfrey 14:34:49 <SteveH> AndyS, oh, sorry, thought it was UC in examples 14:34:56 <AndyS> "if" looks silly 14:35:26 <AxelPolleres> notIn 14:35:36 <SteveH> +1 to ivan 14:35:49 <SteveH> AxelPolleres: just make it uniform 14:36:01 <AxelPolleres> ACTION: ivan to buy beers in SF 14:36:01 <trackbot> Sorry, amibiguous username (more than one match) - ivan 14:36:01 <trackbot> Try using a different identifier, such as family name or username (eg. ivan, imikhail) 14:36:52 <SteveH> AxelPolleres: take this to mail 14:37:07 <SteveH> AndyS: are we putting the F&O set in this doc, or another, or somewhere else 14:37:26 <SteveH> AxelPolleres: extension to core functions which we endorse should be in Q document 14:37:36 <SteveH> AndyS: who's going to write it 14:37:46 <Souri> 15.4.21 example needs a closing paranthesis: STRDT("123", xsd:integer 14:37:49 <SteveH> ... it will take longer if I do it 14:38:01 <SteveH> AxelPolleres: what new text do we need? 14:38:12 <SteveH> ... do we need to add examples for all these functions 14:38:14 <AndyS> souri - already fixed (in my copy) 14:38:43 <SteveH> AxelPolleres: what's missing is the formulation, and adding examples [?] 14:38:44 <AxelPolleres> http://www.w3.org/2009/sparql/docs/query-1.1/rq25.xml#func-coalesce 14:40:08 <AndyS> http://www.w3.org/2009/sparql/wiki/Design:FunctionLibrary 14:40:21 <SteveH> AxelPolleres: not sure about issues around sequences 14:40:34 <SteveH> AndyS: functions in [uri] are directly relevant to SPARQL, no seuences 14:40:44 <SteveH> ... arguments all values that we grok 14:41:18 <SteveH> ... unclear what the proposed way forwards is about this list, how do we deal with rel. to F&O 14:41:24 <AxelPolleres> 3 options 14:41:25 <SteveH> ... see 3 proposals 14:41:39 <SteveH> ... 1) start from SQL 14:41:45 <LeeF> I think the 2nd of those proposals is the best 14:41:46 <SteveH> ... 2) import a speicific set 14:41:56 <SteveH> ... 3) point to F&O equiv 14:42:32 <SteveH> ... in one there are two URIs 14:43:06 <SteveH> AxelPolleres: LeeF thinks we should import a specific set 14:43:06 <LeeF> Is anyone wanting to advocate taking from SQL instead of XPath/XQuery ? 14:43:24 <SteveH> ... if we import we can change them slightly if needed 14:43:40 <sandro> (I'm wondering about taking from RIF, which took option 3.) 14:43:41 <SteveH> LeeF: not sure how much support there is to import from SQL 14:44:05 <SteveH> AxelPolleres: We have COALESCE and IF from SQL 14:44:12 <LeeF> (i'd be fine with that approach as well, sandro. I don't know that much about the RIF function basis.) 14:44:14 <SteveH> ... strawpoll on 2) vs 3) 14:44:16 <AndyS> Interesting -sandro - why did it do that? 14:44:28 <sandro> q+ 14:44:33 <sandro> zakim, unmute me 14:44:33 <Zakim> Sandro should no longer be muted 14:44:44 <SteveH> sandro: story from RIF, having same debate 14:45:01 <SteveH> ... use URIs or make own, decided to make own URIs 14:45:16 <SteveH> ... needed to change semantics in a few places, watned to add a few things, users would like in same namesoace 14:45:24 <SteveH> ... operators don't have URIs anyway 14:45:28 <SteveH> q+ 14:45:34 <AndyS> q+ 14:45:39 <SteveH> ... should sparql and rif use same URIs 14:45:46 <SteveH> q- 14:46:04 <SteveH> ... sparql and rif could coalesce on common uris, and they're both RDFy 14:46:13 <SteveH> AndyS: where did you have to change the smeantics 14:46:22 <AxelPolleres> q+ to speak chair-hat off 14:46:30 <SteveH> sandro: hoping AxelPolleres would remember, something about sequences 14:46:48 <SteveH> ... sequences have unusual semantics 14:46:51 <AxelPolleres> erorrs, sequences, played a role 14:46:55 <sandro> zakim, mute me 14:46:55 <Zakim> Sandro should now be muted 14:46:58 <AxelPolleres> ack me 14:46:58 <Zakim> AxelPolleres, you wanted to speak chair-hat off 14:46:59 <SteveH> ... outside of that I don't remember the differences 14:47:06 <AndyS> q- 14:47:12 <SteveH> AxelPolleres: we had problems with errors too, no errors in RIF 14:47:30 <SteveH> ... what would we do with errors 14:47:36 <AndyS> SPARQL is F&O evaluation 14:47:43 <sandro> (I think the errors bit is not worth doing a different URI over, personally, but I'm not a formalist.) 14:47:55 <AndyS> ... with errors being an extension point 14:48:32 <SteveH> ... if we use own namespace we dont have to think about which ones are keywords, if it has a URI in sparql namespace it can be used in syntax 14:48:40 <AxelPolleres> q? 14:48:47 <sandro> q- 14:49:20 <SteveH> AxelPolleres: poll: keep F&O namespace, or give own [sparql] namespace 14:49:34 <SteveH> 1 - keep namespace 14:49:39 <SteveH> 2 - sparql ns 14:49:45 <AxelPolleres> 2 14:49:46 <pgearon> 2 14:49:48 <AndyS> Option 2 (reuse namespace + alloc a namespace for operators) 14:49:52 <kasei> 2 (mild preference, would like to see rif+sparql convergence) 14:49:56 <MattPerry> 2 14:49:56 <bglimm> 2 14:49:57 <SteveH> 1 14:50:11 <LeeF> 0 14:50:15 <sandro> 0 unsure..... 14:50:30 <AndyS> Option 1 (reuse namespace + alloc a namespace for operators) 14:50:47 <LeeF> Yeah, I agree with that 14:50:50 <LeeF> too many keywords :) 14:52:54 <SteveH> AxelPolleres: op: ones are things that can be moddled with + etc. [?] 14:53:02 <AxelPolleres> op : � + - ! 14:53:19 <AxelPolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg/2009JulSep/0244.html 14:53:22 <SteveH> "op:" has no namespace officially 14:53:37 <SteveH> AxelPolleres: we dont have a URI for them 14:53:44 <AndyS> yes - but what about writing "op:numeric-add(?X, ?Y)" 14:54:15 <SteveH> AxelPolleres: you have to write the short form in xpath etc. 14:54:50 <SteveH> ... datatypes depenends on how it's delegates 14:55:05 <AxelPolleres> http://www.w3.org/2009/sparql/wiki/Design:FunctionLibrary 14:55:23 <SteveH> ... not complete, but we should have op URI + shortcut 14:55:45 <SteveH> ... then shortcuts come from [sth] 14:56:01 <SteveH> ... should take to email, running out of time 14:56:02 <AndyS> My pref: complete Design:FunctionLibrary as a design then take to WG. 14:56:55 <SteveH> AndyS: I don't have any time in next 2 weeks 14:57:01 <AxelPolleres> ACTION: Axel to work out http://www.w3.org/2009/sparql/wiki/Design:FunctionLibrary within 2 weeks 14:57:01 <trackbot> Created ACTION-256 - Work out http://www.w3.org/2009/sparql/wiki/Design:FunctionLibrary within 2 weeks [on Axel Polleres - due 2010-06-15]. 14:57:41 <AxelPolleres> http://lists.w3.org/Archives/Public/public-rdf-dawg/2010AprJun/0311.�html 14:58:03 <SteveH> AndyS: couple of proposals, realtively decided on cardinality, want to work on details 14:58:48 <Zakim> -LeeF 14:59:12 <SteveH> AndyS: don't think there's enough detail to make proposal 14:59:59 <SteveH> ... matter of writing document and coming back with proposal 15:00:14 <SteveH> AxelPolleres: as long as you know where it's going 15:00:36 <SteveH> update next time 15:00:40 <SteveH> + other issues 15:00:51 <AndyS> At risk next time: AndyS 15:01:03 <Zakim> -bglimm 15:01:08 <ivan> zakim, drop me 15:01:08 <Zakim> Ivan is being disconnected 15:01:09 <Zakim> -Ivan 15:01:10 <AxelPolleres> adjourned 15:01:11 <Zakim> -SteveH 15:01:15 <Zakim> -Souri 15:01:15 <MattPerry> quit 15:01:16 <Zakim> -kasei 15:01:18 <Zakim> -AndyS 15:01:19 <Zakim> -MattPerry 15:01:21 <Zakim> -pgearon 15:01:23 <AxelPolleres> thanks steve for scribing... 15:01:23 <Zakim> -Sandro 15:01:26 <SteveH> AxelPolleres, how do I make mins from the IRC? 15:01:37 <AxelPolleres> rrsagent, make records public 15:01:56 <AxelPolleres> steve, will handle that... essentially: http://www.w3.org/2009/CommonScribe/manual.html # SPECIAL MARKER FOR CHATSYNC. DO NOT EDIT THIS LINE OR BELOW. SRCLINESUSED=00000320