Warning:
This wiki has been archived and is now read-only.
Chatlog 2011-12-13
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.
14:59:49 <RRSAgent> RRSAgent has joined #sparql 14:59:49 <RRSAgent> logging to http://www.w3.org/2011/12/13-sparql-irc 14:59:54 <LeeF> RRSAgent, make logs world 14:59:57 <trackbot> RRSAgent, make logs world 14:59:58 <LeeF> zakim, this is SPARQL 14:59:59 <swh> swh has joined #sparql 15:00:01 <trackbot> Zakim, this will be 77277 15:00:01 <Zakim> ok, LeeF; that matches SW_(SPARQL)10:00AM 15:00:03 <trackbot> Meeting: SPARQL Working Group Teleconference 15:00:04 <LeeF> zakim, this is SPARQL 15:00:04 <Zakim> ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start now 15:00:05 <trackbot> Date: 13 December 2011 15:00:07 <Zakim> ok, LeeF; that matches SW_(SPARQL)10:00AM 15:00:09 <kasei> Zakim, who is talking? 15:00:16 <cbuilara> zakim, who is on the phone 15:00:16 <Zakim> I don't understand 'who is on the phone', cbuilara 15:00:19 <AndyS> zakim, who is on the phone 15:00:19 <Zakim> I don't understand 'who is on the phone', AndyS 15:00:21 <trackbot> Sorry, LeeF, I don't understand 'trackbot, get with the program'. Please refer to http://www.w3.org/2005/06/tracker/irc for help 15:00:21 <Zakim> kasei, listening for 10 seconds I heard sound from the following: ??P10 (25%) 15:00:23 <AndyS> zakim, who is on the phone? 15:00:29 <Zakim> On the phone I see +1.310.729.aaaa, [IPcaller], MattPerry, ??P10 15:00:36 <kasei> Zakim, aaaa is me 15:00:37 <AndyS> zakim, IPCaller is me 15:00:45 <cbuilara> zakim, ??P10 is me 15:00:46 <Zakim> +kasei; got it 15:00:47 <Zakim> +AndyS; got it 15:00:51 <AndyS> zakim, go faster 15:00:53 <Zakim> +LeeF 15:00:55 <Zakim> +cbuilara; got it 15:01:01 <Zakim> I don't understand 'go faster', AndyS 15:01:32 <Zakim> +pgearon 15:01:33 <Zakim> +??P14 15:01:37 <Zakim> +??P17 15:01:43 <swh> hm, one of those is me... 15:01:52 <Zakim> +sandro 15:02:10 <swh> Zakim, ??P17 is me 15:02:11 <Zakim> +swh; got it 15:02:31 <LeeF> zakim, who's on the phone? 15:02:31 <Zakim> On the phone I see kasei, AndyS, MattPerry, cbuilara, LeeF, pgearon, ??P14, swh, sandro 15:02:37 <bglimm> bglimm has joined #sparql 15:03:03 <LeeF> zakim, ??P14 is Olivier_ 15:03:03 <Zakim> +Olivier_; got it 15:03:29 <LeeF> Regrets: AxelPolleres, chimezie, bglimm 15:04:38 <LeeF> Chair: LeeF 15:04:43 <LeeF> zakim, pick a scribe, please 15:04:43 <Zakim> Not knowing who is chairing or who scribed recently, I propose Olivier_ 15:04:54 <LeeF> scribenick: Olivier_ 15:05:20 <LeeF> PROPOSED: Accept last week's minutes at http://www.w3.org/2009/sparql/meeting/2011-12-06 15:06:14 <LeeF> RESOLVED: Accept last week's minutes at http://www.w3.org/2009/sparql/meeting/2011-12-06 15:06:29 <LeeF> next week's meeting is 12/20 at 10:00am EST 15:06:46 <swh> I'm at risk for next week 15:06:49 <pgearon> +q 15:06:57 <LeeF> ack ??P17 15:06:59 <LeeF> ack pgearon 15:07:15 <LeeF> pgearon: at risk next week 15:07:52 <LeeF> LeeF: next week we will be voting on most documents publication, let me know if you won't be here 15:08:04 <pgearon> q- 15:09:02 <LeeF> www.w3.org/2009/sparql/wiki/PostLastCall 15:09:08 <LeeF> topic: query 15:10:39 <kasei> q+ 15:10:59 <Olivier_> Some small text changes still needed, minor changes 15:11:27 <bglimm> Sorry for only being on IRC. I just wanted to mention that my part1 has all been addressed and I hope to send part 2 tonight 15:11:48 <AndyS> birte, great, thanks 15:11:50 <LeeF> ack kasei 15:11:50 <bglimm> Talking about the Query review in case that wasn't clear 15:11:53 <swh> q+ 15:12:10 <Olivier_> Reread aggregate section needed ? 15:12:43 <bglimm> I read the aggregate section again and I am mostly happy (some minor comments) 15:12:48 <LeeF> ack swh 15:13:20 <Olivier_> Concerns exist about aggregate section 15:15:25 <Olivier_> Include a note about it 15:15:53 <AndyS> LeeF: Will buy everyone a beer if there is a serious bug in aggregates 15:16:06 <Olivier_> We may publish it as LC next week 15:16:19 <sandro> sandro: It might be good to include a note like, "This algebra has not yet been thoroughly reviewed. If you think you found a mistake, please send in a comment." 15:16:21 <LeeF> LeeF: birte to get in her review tonight, andy and steve to address points, group to decide on query LC2 next week 15:16:43 <LeeF> topic: Entailment 15:17:17 <bglimm> Markus review is addressed and I am awaiting Axel's 15:17:32 <LeeF> bglimm, Axel's review was sent 2 days ago 15:17:40 <LeeF> have you had a chance to look at it yet? 15:17:46 <bglimm> Ups, then I overlooked it. Will check again 15:17:54 <bglimm> No, I didn't see it 15:18:00 <LeeF> Here's the link: http://lists.w3.org/Archives/Public/public-rdf-dawg/2011OctDec/0338.html 15:18:18 <LeeF> If you could take a quick look and let us know if it looks manageable to resolve these comments this week, that would be great 15:18:24 <Zakim> -AndyS 15:18:30 <bglimm> I only saw a question about synchronising the wording between Query and Ent. Reg. regarding the conditions (E-consitency etc) 15:18:49 <LeeF> topic: Update 15:18:51 <Zakim> +??P1 15:18:57 <AndyS> zakim, ??P1 is me 15:18:57 <Zakim> +AndyS; got it 15:19:46 <Zakim> -MattPerry 15:20:01 <LeeF> pgearon: working through the 2nd half of Matt's review 15:20:04 <Olivier_> Some minor changes still needed 15:20:06 <LeeF> LeeF: please check if there is anything non-editorial 15:20:07 <bglimm> Seems all managable 15:20:09 <Zakim> +MattPerry 15:20:28 <bglimm> I'll have Ent. Reg. ready next week then 15:20:34 <LeeF> MattPerry: mainly editorial - one thing was a clarification of words of SHOULD/MUST 15:20:48 <pgearon> just finished looking, and it's all fine 15:21:14 <LeeF> LeeF: pgearon to finish matt's review shortly and matt to review the changes to see if OK to publish 15:21:34 <LeeF> topic: service description 15:22:08 <LeeF> LeeF: chimezie sent his review last week, what's outstanding? 15:22:15 <LeeF> kasei: remaining issue is wording on property feature 15:23:12 <Olivier_> Some open issues 15:23:20 <Olivier_> not substantive 15:25:28 <LeeF> LeeF: for property features, suggest changing to "Relates an instance of sd:Service to a resource representing an implemented feature that extends the SPARQL Query or Update language and that is accessed by using the named property." 15:27:05 <Olivier_> It may be published next week 15:27:33 <Olivier_> topic: graph protocol 15:27:33 <LeeF> topic: graph store protocol 15:27:41 <LeeF> LeeF: sandro, do you have anything to add beyond what's in email? 15:27:50 <LeeF> sandro: not much; probably OK with most things but concerned with POST being append 15:28:19 <LeeF> sandro: the question here is what should happen if you POST to a graph resource 15:28:28 <LeeF> ... in general in REST, that's up in the air - you send a message to the thing and what happens depends on the thing 15:28:47 <LeeF> ... if the thing is a piece of a SPARQL database, saying that a reasonable thing to do is append to the graph seems ok 15:29:06 <LeeF> ... at linked enterprise workshop, people are using RDF and REST with other applications / data models 15:29:13 <LeeF> ... there are a lot of other things that you might want to do 15:29:27 <LeeF> ... so i'd like to keep POST open - give the ability to POST a message to something asking it to do something 15:29:32 <kasei> q+ 15:29:33 <LeeF> ... can always append with PATCH 15:29:41 <LeeF> ... suggest we back off on this specification 15:29:51 <LeeF> ... recognize problems with PATCH - it's new, not implemented by IE9 (from JavaScript) 15:29:57 <LeeF> ack kasei 15:30:30 <LeeF> kasei: worried about making this change because we explicitly went through this and made PATCH informative because we were worried about the newness/widespread support with PATCH, and thought we had consensus on POST as append 15:30:53 <LeeF> sandro: i supporeted that, but 2 new bits of information 15:31:04 <LeeF> ... 1) people using this with things other than RDF store 15:31:09 <LeeF> ... 2) people at workshop happy with using PATCH 15:33:13 <LeeF> LeeF: steve, do you implement POST to a graph? 15:33:16 <LeeF> swh: yes, it's append for us 15:33:32 <LeeF> LeeF: sandro, the IBM implementation handles POST to a collection, not a graph, right? 15:33:50 <LeeF> sandro: yes, i'm thinking about the implications of that general idea extended to POSTing to graph 15:33:58 <LeeF> sandro: steve, would you have problems changing to using PATCH for append? 15:34:00 <kasei> PATCH seems pretty underspecified in the document right now. I assume this would require a big change to the document from its current state...? 15:34:14 <LeeF> swh: not enthusiastic about it given lack of support in HTTP libraries and need to change existing code 15:35:50 <LeeF> AndyS: I believe that PATCH is the open-ended one, lets you send any sort of change to a resource, not just append 15:37:31 <LeeF> sandro: will think about this & see if there is a more compelling argument by next week 15:38:06 <LeeF> sandro: ...a collection is also a graph 15:38:20 <LeeF> sandro: ...so POST to that particular graph means something different 15:38:28 <LeeF> AndyS: Where is that specified? 15:46:11 <sandro> sandro: I think this protocol does or should apply to RDF+REST everywhere -- not just being SPARQL. And collections should be a kind of graph. 15:47:18 <sandro> sandro: so POST to collecition (a kind of graph) is different from POST to other graphs? 15:49:53 <sandro> "Protocol enhancements for update. The group will also define protocol to update RDF graphs using ReSTful methods. " 15:50:21 <sandro> from http://www.w3.org/2011/05/sparql-charter 15:50:32 <AndyS> AndyS: I see no evidence presented that a collection IS a graph. It may have a partial representation as a graph (c.f. RDFa) 15:50:36 <LeeF> LeeF: Suggestion is to ask Sandro to produce a proposed change and decide on it next week; also happy to facilitate an out-of-band discussion between now and then for interested parties 15:50:40 <sandro> sandro: If that's what we're doing here, I think we need to keep in mind the bigger picture, beyond just SPARQL. 15:51:19 <sandro> LeeF: Anyone on the call interested in joining a call on this? 15:51:34 <sandro> Lee: Sandro, Andy, Lee, Steve... anyone else? 15:52:27 <sandro> kasei: I'm with Andy in being kind of horrified that the SD is returned on a GET from the GraphStore 15:53:09 <sandro> kasei: I don't know how we got to there 15:53:23 <AndyS> There is an example of GET on the graph store URI in the POST section. 15:54:56 <sandro> kasei: The text seems to say you should return the SD *even if* the GraphStore doesn't implement SPARQL. 15:55:54 <LeeF> LeeF: need to discuss 5.8 of GSP -- Lee, Andy, kasei are surprised and worried about what's there 15:57:11 <kasei> the graph store protocol document also seems to suggest that SD be returned from an OPTIONS requests, which also might deserve its own discussion. 15:57:16 <LeeF> ACTION: Lee to try to organize a conversation on POSTing to a graph this week 15:57:17 <trackbot> Created ACTION-571 - Try to organize a conversation on POSTing to a graph this week [on Lee Feigenbaum - due 2011-12-20]. 15:57:59 <LeeF> topic: Protocol 15:58:06 <LeeF> LeeF: remaining issues are editorial, specifically cleaning up references 15:59:32 <Zakim> -LeeF 15:59:33 <Zakim> -sandro 15:59:34 <Zakim> -swh 15:59:35 <Zakim> -MattPerry 15:59:38 <Zakim> -pgearon 15:59:39 <Zakim> -AndyS 15:59:40 <Zakim> -kasei 15:59:41 <Zakim> -Olivier_ 15:59:42 <LeeF> LeeF: decide on most documents next week - please let me know if you won't be here 15:59:42 <Zakim> -cbuilara 15:59:43 <LeeF> Adjourned. 15:59:44 <Zakim> SW_(SPARQL)10:00AM has ended 15:59:46 <Zakim> Attendees were +1.310.729.aaaa, MattPerry, kasei, AndyS, LeeF, cbuilara, pgearon, sandro, swh, Olivier_ 16:01:03 <AndyS> AndyS has left #sparql 16:21:29 <AndyS> LeeF - transport endpoint -- you mean http://services.data.gov.uk/transport/sparql ? 16:23:56 <LeeF> AndyS, maybe, checking :-) 16:24:07 <LeeF> no, but let me try that one 16:24:36 <AndyS> IIRC There are 2+ hosted copies. Which one did you use? 16:24:46 <LeeF> AndyS, i meant choosing "transportation" from the drop-down at http://data.gov.uk/sparql 16:25:30 <LeeF> the one you gave me gives a similarly empty response to my apparently broken / bit rotten query :-) 16:25:40 <LeeF> basically, i have a query in my tutorial that used to work, and now doesn't return any results 16:25:49 <LeeF> and was trying to short-circuit the process of debugging it by asking those in the know :) 16:26:27 <AndyS> They are the same actually. 16:26:53 <LeeF> ah ok 16:27:09 <LeeF> this query used to work but now returns no results: 16:27:10 <LeeF> PREFIX skos: <http://www.w3.org/2004/02/skos/core#> 16:27:10 <LeeF> PREFIX roads: <http://transport.data.gov.uk/0/ontology/roads#> 16:27:10 <LeeF> SELECT ?cat_name (COUNT(DISTINCT ?thing) AS ?roads) 16:27:10 <LeeF> WHERE { 16:27:10 <LeeF> ?thing a roads:Road ; roads:category ?cat . 16:27:11 <LeeF> ?cat skos:prefLabel ?cat_name 16:27:14 <LeeF> } 16:27:15 <LeeF> GROUP BY ?cat_name 16:27:18 <LeeF> and not sure what in particular changed... 16:30:17 <AndyS> I see no roads:Road 16:30:56 <LeeF> right... so question is whether there is something that replaced it, or moved, or what 16:31:09 <LeeF> i tried to do a general query for the rdf:types in the dataset, but it timed out for me 16:32:05 <AndyS> The endpoint is run by Talis - you can ask on IRC freenode #talis (or maybe #kasabi) The time out is the platofmr's 30s time out but try LIMIT / OFFSET for slicing which might help. 16:35:18 <AndyS> Their should be a dump somewhere close. I can't the ontology either to check URIs. Bad LD. 16:35:33 <LeeF> thanks very much, Andy! 16:35:46 <AndyS> SELECT * { ?s <http://jena.hpl.hp.com/ARQ/property:version> ?y } works :-) 16:36:31 <AndyS> Some SPARQL 1.1 features are turned off IIRC (to stop accidental DOS-isms). 16:36:53 <AndyS> Did the query once work? 16:41:09 <AndyS> Ah - see it did. 16:41:29 <AndyS> I'm not in the know but I know who to ask ... 16:43:11 <AndyS> I do know what code it being run .. but the data? A bit above my level of the stack. 16:43:43 <LeeF> :-) 16:43:53 <LeeF> would you be able to point me to who to ask about the data? 16:47:13 <AndyS> beobal (Sam Tunnicliffe - head of platform team at Talis) - we're on freenode #jena talking about it 16:51:20 <LeeF> LeeF has joined #sparql 16:51:34 <LeeF> having trouble connecting to freenode for some reason 16:52:09 <AndyS> That's freenode. It's free and 17:21:38 <bglimm> bglimm has joined #sparql 17:32:18 <Zakim> Zakim has left #sparql 17:56:05 <bglimm> SteveH, are you still on IRC? 17:56:36 <SteveH> hi bglimm, yeah 17:56:47 <bglimm> What did Olivier mean with 17:56:47 <bglimm> Olivier_: Concerns exist about aggregate section 17:56:47 <bglimm> [15:15] Olivier_: Include a note about it 17:57:18 <bglimm> I more or less finished reading the algebra and don't see major problems 17:57:25 <bglimm> am I overlooking something? 17:57:50 <bglimm> Is that refering to the agg_i issue that also Andy pointed out in his email? 17:58:24 <bglimm> E.g., what happens if a user uses ?agg1 in the query? 18:13:35 <SteveH> bglimm, yes, just that conversation between you and Andy 18:13:42 <bglimm> ok 18:13:44 <SteveH> bglimm, nothing new 18:14:15 <bglimm> I was worried that there is something new since the scribe wasn't to informative 18:14:24 <bglimm> s/to/too/ 18:16:41 <SteveH> yeah, sure 19:53:18 <AndyS> AndyS has joined #sparql 20:36:48 <AndyS> LeeF - did you get to the bottom of the cause of the lack of results of your query? 20:37:42 <LeeF> *just* this sec found time to start looking at it again 20:37:49 <LeeF> basically just looks like totally different data then it used to be 20:55:33 <AndyS> A small difference esp. if it wasn't reloaded. I did see they are running a customized ARQ 20:56:12 <AndyS> ... which looks mildly recent. But the mod (which is now in the codebase) is only for "order by" according to the logs. 22:48:48 <LeeF> LeeF has joined #sparql # SPECIAL MARKER FOR CHATSYNC. DO NOT EDIT THIS LINE OR BELOW. SRCLINESUSED=00000233