13:48:45 RRSAgent has joined #sparql 13:48:45 logging to http://www.w3.org/2010/05/04-sparql-irc 13:48:47 RRSAgent, make logs world 13:48:47 Zakim has joined #sparql 13:48:49 Zakim, this will be 77277 13:48:49 ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 12 minutes 13:48:50 zakim, this will be SPARQL 13:48:50 Meeting: SPARQL Working Group Teleconference 13:48:50 Date: 04 May 2010 13:48:51 ok, LeeF; I see SW_(SPARQL)10:00AM scheduled to start in 12 minutes 13:48:52 Chair: LeeF 13:48:57 Scribenick: ivan 13:49:09 Regrets: AxelPolleres, SteveH, dcharboneau2 13:49:30 Agenda: http://www.w3.org/2009/sparql/wiki/Agenda-2010-05-04 13:57:21 Prateek has joined #sparql 13:57:47 MattPerry has joined #sparql 13:58:05 SW_(SPARQL)10:00AM has now started 13:58:12 +pgearon 13:58:56 +kasei 13:59:06 chimezie has joined #sparql 13:59:20 +Lee_Feigenbaum 13:59:49 +MattPerry 14:00:00 zakim, dial ivan-voip 14:00:02 ok, ivan; the call is being made 14:00:08 +Ivan 14:00:14 +??P14 14:00:20 zakim, ??P14 is me 14:00:20 +AndyS; got it 14:02:05 zakim, who's on the phone? 14:02:05 On the phone I see pgearon, kasei, Lee_Feigenbaum, MattPerry, Ivan, AndyS 14:03:19 scribenick:ivan 14:03:28 +??P27 14:03:34 minutes from last week: http://www.w3.org/2009/sparql/meeting/2010-04-27 14:03:59 zakim, ??P27 is Orri 14:03:59 +Orri; got it 14:04:19 OlivierCorby has joined #sparql 14:05:28 +OlivierCorby 14:06:15 Regrets: Axel, Steve, David, Souri 14:06:20 Regrets: Axel, Steve, David, Sandro 14:06:27 Regrets: Axel, Steve, David, Souri, Sandro 14:06:48 PROPOSED: Approve minutes from http://www.w3.org/2009/sparql/meeting/2010-04-27 14:06:51 LeeF: anybody concerns with minutes of last week? 14:06:55 ... 1 14:06:57 .... 2 14:07:00 .... 3 14:07:04 .... 4 14:07:21 ... discussions about variables, queries, next round of documents... 14:07:31 RESOLVED: Approve minutes from http://www.w3.org/2009/sparql/meeting/2010-04-27 14:07:41 Next meeting: 2010-05-11 @ 15:00 UK / 10:00 EST (scribe: Sandro) 14:07:45 +chimezie 14:07:56 Zakim, mute me 14:07:56 chimezie should now be muted 14:08:24 Lee: other admin issues 14:08:32 ... couple of comments, they are taken care of 14:08:35 ... same for actions 14:08:43 ... any questions on those? 14:08:51 (no reply from anybody...) 14:09:13 Other wgs: RDB2RDF WG finalizing UCR document 14:09:24 eGov: anybody has questions? 14:09:26 q+ 14:09:30 ack ivan 14:10:34 topic: update on sparql update (sic!) 14:10:53 Lee: seeking immediate feedback 14:11:01 ... or get people to look online 14:11:16 s/online/at offline/ 14:12:30 topic: Recent additions/updates to SPARQL Update 14:12:55 LeeF: issue is the update model 14:13:01 ... what the underlying formalism is 14:14:07 61 14:14:22 i think 61# 14:14:49 60# i think 14:15:00 1 and 0, on and off if I remember right. 14:15:47 ... are those the significant issues to the spec? 14:16:30 update model is at: http://www.w3.org/2009/sparql/docs/update-1.1/Overview.xml#t32 14:16:37 pgearon: the other part was the conversation we had on how we should datasets in insert and delete statements 14:16:53 ... the keywords were not really appropriate for this 14:16:59 ... we wanted to avoid ambiguity 14:17:07 (keywords FROM and FROM NAMED are confusing with DELETE) 14:17:10 ... at the same time nobody wanted to use new keywords 14:17:36 ... USING and USING NAMES came up as a semi-consensus suggestion, I put it in the document 14:18:07 LeeF: and the update model, how final do you think it is, do you expect to go deeper? 14:18:29 pgearon: if we go further we might end up enforcing things on implementations and turn them invalid 14:18:44 ... i do not see that coming any time soon, but it is a danger 14:18:55 ... maybe a little more than what we have now is fine, but not much 14:19:12 ... so long as it is internally consistent the level of details we have is on the right track 14:19:23 ... i mostly copied the points that were on the wiki 14:19:39 If the impls have different outcomes on the same spec, surely the spec is not accurate enough. 14:19:42 (discussions whether that page was done by lee, he does not believe that...) 14:20:11 LeeF: to be clear on what has been put in 14:20:28 -Orri 14:20:29 ... using and using names are synonymous with from and from names, but with different keywords? 14:20:32 pgearon: yes 14:20:55 """ If a USING clause appears, then this will override any effect that WITH may have on the WHERE clause, and only the WHERE clause.""" 14:21:59 LeeF: i guess of this agenda topic was to point out this issue 14:22:02 ... and changes 14:22:13 ... we can close the corresponding issue 14:22:24 ... we may have people review the document to see if this is the final design 14:22:27 trackbot, close ACTION-206 14:22:27 ACTION-206 Coordinate with Paul, Steve, and Andy to form a coherent proposal re: datasets, FROM, FROM NAMED, WITH, default-graph-uri, and named-graph-uri closed 14:22:34 Why doesn't USING apply to DELETE template WHERE pattern and similarly INSERT on their own? 14:22:43 LeeF: anybody any questions? 14:22:57 ... otherwise people should take a look at them and comment 14:23:37 pgearon: in answer to Andy, it should apply there as well, oversight 14:23:38 pgearon: because I was not thorough enough:-( and I will have to put in the right words... 14:24:32 AndyS: i was more worried when you were talk about the update model to avoid implementations to change... (scribe lost) 14:25:00 pgearon: I am not really sure on how much more details are required 14:25:16 ... eg, in general, do you have any formalism on how update work 14:25:29 AndyS: no, because I will have to change my implementation on what the spec says:-) 14:25:50 pgearon: what we have now are more general guidelines rather than being formal 14:26:11 LeeF: my gut feeling is that we will need something more detailed, but I am not sure how much 14:26:37 ... when we begin to collect test cases we will know more by following the spec in more details 14:26:43 ... we will know then if we have a good enough model 14:27:02 ... we will have to rely on the query algebra, but we will not know until we are at the test cases 14:27:28 pgearon: i have a limited experience for this type of formalism, I would need help and I am not the person stipulating what is required 14:27:52 LeeF: we will have to figure what we need and who is the right person to help with it 14:28:11 LeeF: is there anything else of things that are missing 14:28:30 pgearon: return cases for operations, most of them are success or failure, but, eg, you can have a partial load 14:28:34 LeeF: when? 14:28:42 pgearon: should be done by tomorrow night 14:28:50 LeeF: did we pick reviewers last week? 14:28:56 don't think so 14:28:58 (stunned silence) 14:29:09 (or everybody is ducking) 14:29:28 LeeF: we will need at least one reviewer for all documents 14:29:39 ... this is to catch issues 14:29:50 ... and they are in a reasonable stage for publication 14:30:00 LeeF: anybody volunteers to review sparql update 14:30:12 LeeF: I will volounteer 14:30:17 ACTION: LeeF to review SPARQL Update once pgearon gives the go ahead 14:30:17 Created ACTION-229 - Review SPARQL Update once pgearon gives the go ahead [on Lee Feigenbaum - due 2010-05-11]. 14:30:48 LeeF: anybody else reviewing update? 14:31:36 AndyS: for the grammar for publication; there is one grammar that is shared for update and query 14:31:43 http://www.w3.org/2009/sparql/docs/sparql-grammar-11.html 14:31:48 LeeF: have we not published the combined grammar yet? 14:31:55 AndyS: the update stuff is not done yet 14:32:20 LeeF: do we need to wrap it into a document structure with a proper URI in TR space? 14:32:34 ... what is the intention for publishing it 14:32:50 AndyS: the easy way is to put into, say, the query document and make it clear that it is also for update 14:33:00 ... pgearon, would that work for you? 14:33:19 pgearon: yes 14:33:40 LeeF: are there outstanding question on the list to add this to the document? 14:33:52 AndyS: there are couple issues open yet 14:33:59 LeeF: in the update syntax thread 14:34:16 ... let us resolve the syntax issues next week 14:34:29 topic: federated queries 14:34:40 zakim, dial ericP-mobile 14:34:40 ok, LeeF; the call is being made 14:34:41 +EricP 14:35:33 zakim, drop ericp 14:35:33 EricP is being disconnected 14:35:35 -EricP 14:35:58 http://www.w3.org/2009/sparql/docs/fed/service 14:36:23 http://www.w3.org/2009/sparql/docs/fed/service#bindings 14:36:28 LeeF: the issues was the binding stuff 14:36:39 -OlivierCorby 14:36:59 ... short summary is that binding specifies a result set that should be used to constraint the answers to a federated query piece 14:37:24 ... if you have two service nodes, it binds the ones coming from that clause and send that to another call 14:37:34 ... that is the general idea behind bindings 14:37:41 +OlivierCorby 14:37:58 ... andy, most of the bindings from arq, does it include bindings 14:38:08 AndyS: no, that comes from eric's system 14:38:16 http://www.w3.org/2009/sparql/docs/fed/service#update 14:38:38 q+ 14:38:47 LeeF: the other thing is section 3.1 where you can do the same thing for an insert/delete statement as a way to have a compact syntax for ground triples 14:39:05 ack AndyS 14:39:05 q- 14:39:06 ... this was an area where we may not have concensus 14:39:11 q+ 14:39:28 ack ivan 14:39:31 ivan: why is that different than having select in select? 14:39:58 i think BINDINGS is important, but hesitant about its use with UPDATE 14:40:13 LeeF: I am not sure I understand 14:40:21 ivan: then I may not understand 14:40:42 I think you could do this with subselects, but it would be nasty: { SELECT 1 AS ?id {} } UNION { SELECT 2 AS ?id {} } 14:40:56 LeeF: it specifies a result set to join with a query 14:41:05 I'm hestitate about update - seems to mean update needs changing as well and overlaps with INSERT DATA etc. 14:41:20 LeeF: there is an hesitation about using this with update 14:41:42 +q 14:41:46 ack pgearon 14:41:47 ... AndyS also points out that interferes with insert/delete 14:42:00 pgearon: i would mention a use case for binding 14:42:15 q+ to ask about BINDING position (after update discussion) 14:42:18 also, BINDINGS with UPDATE probably doesn't belong in the fed document (SERVICE isn't defined specifically for UPDATE, right?) 14:42:51 ... if you have a query for federation, you may execute part of the query on one host and use the result of that you can go to another place with some of the variables already bound, and that can dramatically reduce time 14:43:12 ... so having binding has a real use case where can have dramatic effects 14:43:21 ... in terms of insert/delete I do not see that too much 14:43:34 kasei - good point - sees to be about the input to the endpoint, not the federating query engine. 14:43:41 in fact, unless BINDINGS is defined as normative in the query doc, it puts a burden on FED extensions to send off speculative queries first but require support for falling back to a non-BINDINGS-enhanced query. 14:43:43 ... we can emulate bindings with subselects and unions and it is messy 14:44:13 ... this particular syntax is useful, but only for select 14:44:29 ack AndyS 14:44:29 AndyS, you wanted to ask about BINDING position (after update discussion) 14:44:53 AndyS: minor point, i have binding in the grammar but after all clauses 14:45:20 ... we had a discussion on the mailing list but it is not yet in the document 14:45:36 ... if everybody on the call has a consensus, then I can catch eric on that 14:45:48 ... i will ask him to just make it so 14:46:16 q+ to ask about moving BINDINGS to the query document 14:46:30 ack kasei 14:46:30 kasei, you wanted to ask about moving BINDINGS to the query document 14:46:57 kasei: i believe that the federated is not required for implementations 14:47:12 ... but the service keyword is provided locally with the endpoint you are talking to 14:47:23 ... the use of binding should be on the other endpoint 14:47:31 ... it should be more on the query side of things 14:48:48 can we say that service descriptions are normative? 14:49:09 LeeF: what if the bindings belong to the query document 14:49:15 kasei: unless we require BINDINGS support for all query implementations, federated systems will likely have to send off speculative queries but may have to drop back to queries not using BINDINGS. 14:49:44 ... are we ok requiring binding for all conformant sparql 1.1 implementations 14:49:45 kasei: either way, this is probably something that should go in the service description. 14:50:17 Zakim, mute me 14:50:17 kasei should now be muted 14:50:32 LeeF: what do other people think? 14:50:42 (stunned silence, they do not think anything) 14:51:30 LeeF: I am inclined to... what am I inclined to do? I am inclined to peek eric's brain... 14:51:36 ... hesitantly think it should also be required. 14:51:52 s/peek/peek into/ 14:52:17 topic: ISSUE-54 14:52:33 mailing list thread: http://lists.w3.org/Archives/Public/public-rdf-dawg/2010JanMar/0487.html 14:52:35 LeeF: putting property functions into service descriptions 14:52:48 Zakim, please dial ericP-mobile 14:52:48 ok, ericP; the call is being made 14:52:50 ... this includes a proposed text 14:52:50 +EricP 14:53:25 -EricP 14:53:38 ... AndyS proposed text how to include that into the description 14:53:49 ... steve and I indicated that we were happy with the proposed text 14:53:49 for some reason I thought we had already resolved this issue... 14:54:01 (already in the document) 14:54:07 ... I would propose to accept this and add it to the document 14:54:25 ... but it seems that greg is faster than anybody else (or travels in time) and has already done it 14:54:33 http://www.w3.org/2009/sparql/docs/service-description-1.1/xmlspec.xml#id0x2c8fdce0 14:54:49 """ 14:54:51 3.4.10 sd:propertyFeature 14:54:51 Releates an instance of sd:Service to a resource representing an implemented feature to the SPARQL Query or Update language that is accessed by using the named property. 14:54:51 sd:propertyFeature is an rdfs:subPropertyOf sd:feature. The rdfs:domain of sd:propertyFeature is sd:Service. 14:54:51 sd:propertyFeature 14:54:52 """ 14:55:15 Zakim, mute me 14:55:15 kasei was already muted, kasei 14:55:28 LeeF: I propose to close this issue 14:55:57 PROPOSED: Close ISSUE-54 based on the current text for sd:propertyFeature in http://www.w3.org/2009/sparql/docs/service-description-1.1/xmlspec.xml#id0x2c8fdce0, noting consensus in mailing list thread http://lists.w3.org/Archives/Public/public-rdf-dawg/2010JanMar/0487.html 14:56:10 +1 14:56:19 +1 14:56:21 RESOLVED: Close ISSUE-54 based on the current text for sd:propertyFeature in http://www.w3.org/2009/sparql/docs/service-description-1.1/xmlspec.xml#id0x2c8fdce0, noting consensus in mailing list thread http://lists.w3.org/Archives/Public/public-rdf-dawg/2010JanMar/0487.html 14:56:31 trackbot, close ISSUE-54 14:56:31 ISSUE-54 Do we need (descriptions of) property functions in SD? Is this in scope for us? closed 14:56:38 ISSUE-54: RESOLVED: Close ISSUE-54 based on the current text for sd:propertyFeature in http://www.w3.org/2009/sparql/docs/service-description-1.1/xmlspec.xml#id0x2c8fdce0, noting consensus in mailing list thread http://lists.w3.org/Archives/Public/public-rdf-dawg/2010JanMar/0487.html 14:56:38 ISSUE-54 Do we need (descriptions of) property functions in SD? Is this in scope for us? notes added 14:56:59 topic: check in our document for readiness 14:57:21 LeeF: we ask our editors to get the documents into reviewable stage and move to publication asap 14:57:55 .... andy: summary on query? 14:58:37 AndyS: i added some more stuff on negation, minus and not exist, and added examples for minus, steve added stuff on aggregates (some formatting problems) 14:59:35 LeeF: it sounds that we can have somebody to review now 14:59:46 i can review it 14:59:50 AndyS: yes, it is messy, but it is only a working draft, so that should fine 15:00:03 ACTION: MattPerry to review SPARQL Query document for issues, concerns, praise, and publication-readiness 15:00:03 Sorry, couldn't find user - MattPerry 15:00:08 ACTION: Matt to review SPARQL Query document for issues, concerns, praise, and publication-readiness 15:00:08 Sorry, couldn't find user - Matt 15:00:13 ACTION: Matthew to review SPARQL Query document for issues, concerns, praise, and publication-readiness 15:00:15 Created ACTION-230 - Review SPARQL Query document for issues, concerns, praise, and publication-readiness [on Matthew Perry - due 2010-05-11]. 15:00:31 LeeF: property path document? 15:01:06 AndyS: I have not done any work since last publication 15:01:20 ... duplicates and negated property paths are to be added 15:01:32 LeeF: we are not required to publish every document for a heartbeat, so that is fine 15:01:34 Zakim, unmute me 15:01:34 kasei should no longer be muted 15:01:40 LeeF: Greg, descriptions 15:02:02 kasei: I wanted to have a link to the serialization format, ivan just did that, will do it 15:02:22 ... the naming of named graphs in datasets is still missing 15:02:38 ... I would like to hear sandro's response on andy's examples, but they look good 15:02:45 ... the document is in a good shape 15:03:03 ... we can sort out the naming issue for the next version, it is to go out today 15:03:08 LeeF: any reviewer? 15:03:27 *crickets* 15:03:32 (deep, deep silence) 15:03:40 Zakim, mute me 15:03:49 LeeF: I will try to find somebody on the mailing list 15:03:51 kasei should now be muted 15:03:51 ACTION: LeeF to find 1 or more reviewers for the SD document 15:03:51 Created ACTION-231 - Find 1 or more reviewers for the SD document [on Lee Feigenbaum - due 2010-05-11]. 15:04:41 I'd be interested in reviewing fed and property paths whenever they're ready for that. 15:05:15 LeeF: I have to put up a poll for the negation issue, will do soon 15:05:19 adjurned 15:05:24 -chimezie 15:05:43 -MattPerry 15:05:44 zakim, drop me 15:05:44 Ivan is being disconnected 15:05:44 bye 15:05:44 -Lee_Feigenbaum 15:05:46 -Ivan 15:05:48 -pgearon 15:05:50 -kasei 15:05:57 -OlivierCorby 15:06:03 -AndyS 15:06:06 SW_(SPARQL)10:00AM has ended 15:06:07 Attendees were pgearon, kasei, Lee_Feigenbaum, MattPerry, Ivan, AndyS, Orri, OlivierCorby, chimezie, EricP 15:32:32 bglimm has joined #sparql 15:33:06 ups, I missed the call right? 15:33:18 I am in Canada, so got time zone confused... 15:34:02 yup. over for 30 minutes now :\ 15:34:11 :-( 17:17:43 Zakim has left #sparql 18:04:37 bglimm has joined #sparql 19:59:56 bglimm has joined #sparql