14:56:34 RRSAgent has joined #sparql 14:56:34 logging to http://www.w3.org/2010/11/16-sparql-irc 14:56:39 Zakim has joined #sparql 14:56:50 trackbot, start meeting 14:56:52 RRSAgent, make logs world 14:56:54 Zakim, this will be 77277 14:56:54 ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 4 minutes 14:56:55 Meeting: SPARQL Working Group Teleconference 14:56:55 Date: 16 November 2010 14:56:59 i think steveh was strongly opposed to it 14:57:10 as was I 14:57:37 yes, and matt as well. 14:58:06 Bad zakim day. 14:58:10 nickH, are you ok to scribe today? 14:58:21 zakim, who is on the phone? 14:58:21 SW_(SPARQL)10:00AM has not yet started, AndyS 14:58:22 On IRC I see RRSAgent, NickH, cbuilara, karl, OlivierCorby, AndyS, AxelPolleres, SteveH, pgearon, ivan, AlexPassant, kasei, iv_an_ru, trackbot, ericP, sandro 14:58:28 zakim, start SPARQL 14:58:28 I don't understand 'start SPARQL', AndyS 14:58:29 AlexPassant: sure 14:58:49 wish tab completion of nicknames was working though 14:58:50 NickH, I am axel ;-) 14:59:05 argh! exactly 14:59:32 zakim, this is SPARQL 14:59:32 ok, AndyS; that matches SW_(SPARQL)10:00AM 14:59:41 +[IPcaller] 14:59:56 zakim, IPcaller is me 14:59:56 +AndyS; got it 15:00:02 Zakim, who is on the phone? 15:00:02 On the phone I see ??P0, ??P2, +1.310.729.aaaa, AxelPolleres, AndyS 15:00:11 Zakim, aaaa is me 15:00:11 +kasei; got it 15:00:11 zakim, dial ivan-voip 15:00:12 ok, ivan; the call is being made 15:00:12 +Ivan 15:00:13 I just called 15:00:15 zakim, mute AndyS 15:00:15 AndyS should now be muted 15:00:21 + +33.4.92.38.aabb 15:00:28 MattPerry has joined #sparql 15:00:29 zakim, unmute AndyS 15:00:29 AndyS should no longer be muted 15:00:54 -AndyS 15:01:07 -??P2 15:01:07 carlos, you were the IPCaller ;-) 15:01:11 + +1.603.897.aacc 15:01:22 zakim, aacc is me 15:01:26 +MattPerry; got it 15:01:31 +??P7 15:01:34 zakim, ??P0 is me 15:01:39 +NickH; got it 15:01:41 zakim, ??P7 is me 15:01:50 +cbuilara; got it 15:01:53 +??P13 15:01:56 zakim, ??P13 is me 15:01:57 +AndyS; got it 15:02:21 41# on your telephone keypad is "raise hand" for those who need to identify themselves 15:02:30 SteveH_ has joined #sparql 15:02:44 40# on your telephone keypad is "hand down" 15:02:55 Zakim, who is on the phone? 15:02:55 On the phone I see NickH, kasei, AxelPolleres, Ivan, +33.4.92.38.aabb, MattPerry, cbuilara, AndyS 15:03:06 Zakim, aabb is me 15:03:06 +OlivierCorby; got it 15:03:29 +pgearon 15:03:35 agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2010-11-16 15:03:41 chair: Axel Polleres 15:03:47 scribe: NickH 15:03:54 regrets: souri, Lee 15:04:23 + +44.208.439.aadd 15:04:30 Zakim, who is on the phone? 15:04:32 regrets+ sandro 15:04:34 Zakim, aadd is me 15:04:37 On the phone I see NickH, kasei, AxelPolleres, Ivan, OlivierCorby, MattPerry, cbuilara, AndyS, pgearon, +44.208.439.aadd 15:04:42 +SteveH; got it 15:05:10 -NickH 15:05:10 PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2010-11-09 15:05:33 I can try 15:05:40 scribe: cbuilara 15:05:48 scribenick: cbuilara 15:05:49 RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2010-11-09 15:06:06 topic: admin 15:06:25 topic: ISSUE-16 15:06:25 Topic: issue 15, aggregates 15:06:31 ISSUE-16 15:06:36 +??P27 15:06:43 zakim, ??P27 is me 15:06:43 +NickH; got it 15:06:56 PROPOSED: Close ISSUE-16 by noting that aggregates define their semantics over mixed data types which can include errors, errors that propagate to a SELECT list result in an unbound variable, errors that propagate to GROUP BY clause result in an "unbound" key. 15:06:56 sorry, phone troubles :( 15:07:21 Lee's proposal: close issue 16 aggregates and errors, 15:07:39 I think we should close 15:07:46 Axel: any opinions? 15:08:13 Axel: Paul wrote lots about context 15:08:30 Alex: is that covered by this proposal 15:08:37 +1 15:08:41 Alex: any support for this proposal? 15:09:05 s/Alex/Axel/ 15:09:16 AndyS: are you planning in putting in expanitory text? 15:09:38 +1 15:09:43 +1 15:09:45 Any other support? 15:09:46 +1 15:09:46 +1 15:09:48 +1 15:09:49 +1 15:10:00 RESOLVED: Close ISSUE-16 by noting that aggregates define their semantics over mixed data types which can include errors, errors that propagate to a SELECT list result in an unbound variable, errors that propagate to GROUP BY clause result in an "unbound" key. 15:10:00 zakim, who is on the phone? 15:10:00 On the phone I see kasei, AxelPolleres, Ivan, OlivierCorby, MattPerry, cbuilara, AndyS, pgearon, SteveH, NickH 15:10:10 Axel: resolve to close 15:10:12 close ISSUE-16 15:10:12 ISSUE-16 Dealing with aggregates over mixed datatypes closed 15:10:32 topic: Potential shortcuts in update (ISSUE-59) 15:10:52 http://www.w3.org/2009/sparql/docs/update-1.1/shortcuts.xml 15:11:09 Axel: Alex proposed three shortcuts 15:11:20 Axel: Copy, Move and Add 15:11:33 Axel: Copy would allow you to copy one graph into another graph 15:11:49 Axel: Move would do the same but remove the original graph 15:12:16 PROPOSED: Close ISSUE-59 by including the shortcut operations from http://www.w3.org/2009/sparql/docs/update-1.1/shortcuts.xml within SPARQL 1.1 Update 15:12:17 Axel: the proposal is the following 15:12:28 Who has user experience? 15:12:31 can we do a strawpoll on this? I'm worried that we were too few last week to be representative of the group. 15:12:31 q? 15:12:34 me 15:12:42 AndyS, can you say a little bit about it? 15:12:49 sure ... 15:12:50 3/1/2 was the vote from last time 15:13:04 q? 15:13:14 Axel: in the interests of going ahead, we should reach concensus or not do it 15:13:44 AndyS: there is the redundency factor 15:14:12 andy has implemented them, finds it handy 15:14:18 AndyS: secondly there is the easy of optimisation 15:14:34 AndyS: much easier to spot the optimiations with the shortcuts 15:14:54 SteveH: would be happier if more people had implemented them 15:15:18 SteveH: very similar but not exactly the same and feels very early to be doing this 15:15:43 Axel: we have a number of opinions and one concern 15:15:44 "probably harmless" 15:16:04 PROPOSED: Close ISSUE-59 by including the shortcut operations from http://www.w3.org/2009/sparql/docs/update-1.1/shortcuts.xml within SPARQL 1.1 Update 15:16:18 abstain 15:16:27 -1 15:16:41 Axel: lets put in the proposal and would like a concensus 15:17:07 q+ 15:17:08 Axel: kasie: can you tell us your objection 15:17:28 kasie: not sure if we have the experience to know if these are the shortcuts that users wants 15:17:46 ivan: we are adding and adding and to the terms in SPARQL 15:17:46 s/kasie/kasei 15:18:00 ivan large number of things to consume and understand 15:18:04 abstain 15:18:06 -0.5 15:18:06 +1 15:18:09 +1 15:18:10 q+ 15:18:11 +1 (on behalf of DERI) 15:18:11 0 15:18:14 q+ 15:18:15 abstain 15:18:21 ack ivan 15:18:24 abstain 15:18:50 (somebody is breathing very loudly into the phone) 15:19:05 3/1.5/4 15:19:38 Ivan: I won't lie down on the road to oppose this 15:20:06 We usually write yes/0/no i.e. 3/4/1.5 15:20:17 SteveH: we can't resolve with 4 abstainsions 15:20:29 Axel: how can we go forward? 15:21:10 Ivan: is does not reflct consensus, therefore it is rejected 15:21:10 I'd be very happy to postopone it 15:21:22 ...make it the next groups problem 15:21:56 +1 to postpone 15:21:57 Axel: who would be happy to postopone issue 59? 15:21:58 +1 to postpone 15:22:01 not really 15:22:05 +1 to postpone 15:22:07 strawpoll: who wants to postpone ISSUE-59? 15:22:10 q+ 15:22:17 +1 15:22:18 -1 (on behalf of DERI) 15:22:23 q? 15:22:50 ack cbuilara 15:23:08 sorry, I can't hear very well 15:23:40 Axel: if it is not in the spec, then it doesn't really exist 15:23:57 Axel: should we put it in the spec or not? 15:24:10 cbuilara: I would leave it out 15:24:25 ack SteveH 15:24:27 ack me 15:24:36 cbuilara: would postpone it and keep it as an unofficial extention 15:24:59 SteveH: is it a strawpoll or an official vote? 15:25:34 ack ivan 15:25:35 SteveH: not obvious what type of vote the original poll was 15:25:48 ivan: what does postponing mean? 15:25:55 q+ 15:26:27 ivan: does it mean that we need more implementation experience? or that the complexity of the language is too much? 15:26:30 ivan: two concerns, one about implementation experience, one about complicating the language 15:27:12 AndyS: my interpretation is that it is put on a list of things to look at later 15:27:16 andy: postponed was meant as "put it on the list of Postponed ISSUES" 15:27:45 ivan: i agree with postponement 15:28:07 summary: majority seems to be for postponing ISSUE-59 15:28:29 Axel: we are likely not to such ISSUE-59 untl the end 15:28:36 q? 15:28:38 I do not support postponing. 15:29:14 AndyS: I think it is an importand issue and it isn't something that should be postponed 15:29:26 AndyS: take it to the community and see what they think 15:30:08 Axel: that alternative would be to put it into the draft get feedback and then take it out again later 15:30:20 +1 to avoid idioms 15:30:57 Andy: how about putting it in as "At risk" feature into LC 15:31:00 Axel: I don't see any problems with adding it to the next draft if we don't have agreement - it can be 'At risk" 15:31:05 Ivan: even possible in CR. 15:31:22 Axel: that would be a way that we could go forward 15:31:26 q+ 15:31:29 ack AndyS 15:31:29 q- 15:32:06 Ivan: that is a way forward, I think that in the document we should make it clear the reason that it is At Risk and that we would like feedback 15:33:23 PROPOSED: add update shortcuts in LC marked explicitly "AT RISK" and asking for feedback, explicitly about potentially complicating the language, and implementation experience 15:33:27 -1, I think it's a bad idea 15:33:40 this is a strawpoll 15:33:44 1 15:33:54 0 15:34:03 0 15:34:07 AndyS: can you say why this is a bad idea? 15:34:48 SteveH: this is quite a big chunk of text to add to the document. But is mainly a 'faster horses' problem. If you ask people if they want faster horses, they wil say they want it 15:35:23 LeeF has joined #sparql 15:35:24 q+ to speak with chairhat off for Alex 15:35:26 SteveH: the 'faster horses' propblem is a quote from Henry Ford about why User Focus groups don't work 15:35:46 +1 15:35:49 SteveH: is a an example of when you will get bad results from asking users what they want 15:36:00 q? 15:36:05 ack iv_an_ru 15:36:05 SteveH: it will add 4 or 5 extra pages of specification 15:36:09 ack ivan 15:36:09 ack ivan 15:36:14 maybe 3 or 4 15:36:49 It's 1.5 pages currently for me in FF 15:36:50 Axel: this is nor really makeing the language more complex but making it easier for users 15:37:03 Axel: the spec is already there 15:37:05 q+ 15:37:14 ack AxelPolleres 15:37:14 AxelPolleres, you wanted to speak with chairhat off for Alex 15:38:31 greg: we have already described that there are two possible ways of storing, either as graphstores or quadstores. I don't think this is idiomatic in any way. 15:39:18 PROPOSED: add update shortcuts in LC marked explicitly "AT RISK" and asking for feedback, explicitly about potentially complicating the language, and implementation experience 15:39:28 Axel: now putting this forward for an offical vote 15:39:35 abstain 15:39:37 0 15:39:38 +1 15:39:39 +1 15:39:39 1 15:39:40 abstain 15:39:42 abstain 15:39:42 +1 (for DERI) 15:39:43 +1 15:39:45 0 15:39:54 what does 0 mean? 15:40:09 you can either abstain, object, or be silent 15:40:13 abstain 15:40:34 4/5/0 15:40:58 Axel: there are still a lot of abstaintions, but no objections 15:41:10 SteveH: still can't move forward from this 15:41:33 AndyS: you can move forward from this, abstain could mean "don't care" 15:41:58 Ivan: my feeling right now is to move on and postpone the issue 15:42:24 Axel: I need to talk to Alex and see what he says 15:42:37 Axel: and see if he objects to this 15:42:54 summary: majority seems to be for postponing ISSUE-59 15:42:57 Axel: what is your opinion AndyS? 15:43:29 AndyS: I see no evidence of that conclusion 15:44:23 Suggest to bring these two questions forward again when Alex is back: (a) put in as AT RISK (b) POSTPONE ISSUE-59 15:44:39 Axel: moving on to next topic 15:44:45 topic: Grouping by expressions 15:45:04 PROPOSED: The GROUP BY clause can include expressions with or without AS aliases. These aliases can be re-used in the HAVING clause and query result clause. 15:46:02 Axel: last time there was some discussion if the BIND in the where clause might be redundant 15:46:14 PROPOSED: The GROUP BY clause can include expressions with or without AS aliases. These aliases can be re-used in the HAVING clause and query result clause (i.e. SELECT or CONSTRUCT clause). projections created in the SELECT clause are not accessible for GROUP BY expressions. 15:46:43 +1 15:46:54 Axel: projections after groupings 15:46:55 q? 15:46:59 is the idea that { ... BIND( expr AS ?x ) } GROUP BY ?x exatly= { ... } GROUP BY (expr AS ?x) ... 15:47:02 ack kasei 15:47:34 q+ 15:47:49 except that GROUP BY (expr) wouldn't cause an extra variable binding 15:48:45 sorry, having trouble hearing/understanding this bit 15:49:57 SteveH: there could be a variety of different things happening after the SELECT 15:50:27 AndyS: if you do SELECT COUNT(*) then there is an implicit grouping 15:50:37 PROPOSED: The GROUP BY clause can include expressions with or without AS aliases. These aliases can be re-used in the HAVING clause and query result clause (i.e. SELECT or CONSTRUCT clause). projections created in the SELECT clause are not accessible for GROUP BY expressions. 15:50:49 -NickH 15:50:59 stupid phone 15:51:05 sure 15:51:42 AxelPolleres: I do not see any objection, vote proposal 15:51:43 +??P27 15:51:47 +1 15:51:53 +1 15:51:54 +1 15:51:54 +1 15:51:55 +1 15:51:56 0 15:51:58 0 15:52:03 +1 15:52:07 +1 15:52:18 8 +1, 2 abstentions 15:52:22 Zakim, ??P27 is me 15:52:22 +NickH; got it 15:52:43 sorry, 7 +1 15:52:44 7/2/0 15:53:05 Axel: I guess we can accept it 15:53:08 8/2/0, I didn't say anything 15:53:10 I didn't abstain 15:53:10 forever hold my piece 15:53:25 Axel: anyone want to explain their abstention? 15:53:29 RESOLVED: The GROUP BY clause can include expressions with or without AS aliases. These aliases can be re-used in the HAVING clause and query result clause (i.e. SELECT or CONSTRUCT clause). projections created in the SELECT clause are not accessible for GROUP BY expressions. 15:54:01 Axel: a few minutes left... 15:54:15 topic: function library 15:54:25 Axel: see if we can agree anything on the function library 15:54:27 http://lists.w3.org/Archives/Public/public-rdf-dawg/2010OctDec/0188.html 15:54:55 no objections against including these 15:55:28 17.4.1 bound .... 17.4.24 NOT EXISTS and EXIST 15:55:34 Axel: we have a set of functions already in the current draft that are non-controversial 15:55:48 plus RAND() RAND (seed) 15:56:24 Axel: would there be any objection to the functions in Lee's email? 15:56:40 Axel: silence is taken as agreement 15:56:54 fn:concat 15:57:20 http://www.w3.org/TR/xpath-functions/#func-concat 15:57:20 Axel: there was some discussion about concat, and if it needs to be redefined for SPARQL 15:57:27 we have variable arity already 15:57:49 AndyS: the problem with concat is that it takes an existing type and casts it to string 15:58:07 s/existing/xsd atomicType/ 15:58:19 I think we should use fn:concat as is 15:58:28 Axel: it would mean rather than casting to string, it would use our special string function 15:58:30 but not for + 15:58:35 AXEL: CONCAT() would be either implicitly use STR() or only acccept strings 15:59:43 AndyS: it can either only accept a string or it can blow up 16:00:14 Axel: there are two options to add our own function 16:00:23 variable arity is fine 16:00:35 q+ 16:00:39 q- 16:00:52 Axel: we are running out of time, going to finish the queue 16:00:56 ack AndyS 16:01:23 http://www.w3.org/TR/2010/REC-rif-dtb-20100622/ 16:01:28 q+ 16:01:37 ack ivan 16:01:41 Ivan: the RIF people have already been through this, can't we just use their definition 16:02:24 Axel: there are no errors in RIF, but SPARQL does have errors, so we can return an error 16:02:50 Ivan: there should be compatibility between RIF and SPARQL 16:03:36 +1 to using a similar set of functions as RIF 16:03:41 Ivan: I understand that there are semantic different, but if I was a dumb user, I would want them to behave in the same way 16:03:53 or at least a similar set of names 16:04:30 Axel: the things that are mentioned in Lee's email do not contrdict that 16:05:14 zakim, drop me 16:05:14 Ivan is being disconnected 16:05:16 -Ivan 16:05:19 -MattPerry 16:05:20 -SteveH 16:05:22 -kasei 16:05:22 ACTION: Axel to take functions discussion on email. 16:05:22 Created ACTION-333 - Take functions discussion on email. [on Axel Polleres - due 2010-11-23]. 16:05:23 -OlivierCorby 16:05:25 -AxelPolleres 16:05:26 -pgearon 16:05:29 -cbuilara 16:05:35 AndyS, why are you in favour of type strictness for CONCAT()? 16:05:42 maybe I'm missing something 16:05:44 rrsagent, make records public 16:05:54 cbuilara1 has joined #sparql 16:06:10 cbuilara1 has left #sparql 16:06:15 I'm not - was expressing alternatives as I recall discussion 16:06:18 Axel: sorry about calling you Alex! Im not the best at typing fast! 16:06:24 AndyS, ah, ok 16:06:24 -AndyS 16:06:30 cbuilara1 has joined #sparql 16:06:58 -NickH 16:06:59 SW_(SPARQL)10:00AM has ended 16:07:01 We are strict for REGEX, say. 16:07:01 Attendees were +1.310.729.aaaa, AxelPolleres, AndyS, kasei, Ivan, +33.4.92.38.aabb, +1.603.897.aacc, MattPerry, NickH, cbuilara, OlivierCorby, pgearon, +44.208.439.aadd, SteveH 16:07:32 I do think strict for + makes sense as a conservative spec choice. 16:07:41 Axel: I think it was the UK phone number gateway that kept disconnecting me - it is still far from ideal 16:16:21 OlivierCorby has left #sparql 16:20:55 SteveH has joined #sparql 16:21:20 AxelPolleres has left #sparql 16:27:27 cbuilara1 has left #sparql 18:10:18 Zakim has left #sparql 19:02:03 LeeF has joined #sparql 20:24:41 LeeF (or anyone), do you know how to invoke the an xmlspec transformation to give a WD instead of an editor's draft? 20:59:03 ericP: I'm not sure if it's multiple things, but I think this covers the process: http://www.w3.org/2009/sparql/wiki/Pub-Process 20:59:24 either ENTITY maturity.level, or spec w3c-doctype="wd", or some combination... 21:09:14 kasei, rock. tx! 21:24:51 AndyS has joined #sparql 21:27:29 wsirc_735156 has joined #sparql 21:57:06 LeeF has joined #sparql