13:53:23 RRSAgent has joined #sparql 13:53:23 logging to http://www.w3.org/2009/09/15-sparql-irc 13:53:25 RRSAgent, make logs world 13:53:25 Zakim has joined #sparql 13:53:27 Zakim, this will be 77277 13:53:27 ok, trackbot; I see SW_(SPARQL)10:00AM scheduled to start in 7 minutes 13:53:28 Meeting: SPARQL Working Group Teleconference 13:53:29 Date: 15 September 2009 13:53:30 zakim, this will be SPARQL 13:53:30 LeeF, Team_(sparql-func)13:02Z is already associated with an irc channel; use 'move SPARQL to here' if you mean to reassociate the channel 13:53:34 Chair: LeeF 13:54:40 LukeWM, are you available as backup scribe if necessary? 13:54:48 hmm, the scribe list is getting a little unmaintained 13:55:01 yes, sure LeeF 13:55:27 SimonS - I don't think it has been created. Coudl you create a file in that dir and cvs ci it in (CVS knows nothing about directories :-() 13:55:49 thanks, LukeWM 13:56:03 ok, thanks AndyS 13:56:41 SW_(SPARQL)10:00AM has now started 13:56:47 +??P2 13:57:01 zakim, ??P2 is me 13:57:01 +AndyS; got it 13:57:04 + +1.617.960.aaaa 13:57:23 zakim, aaaa is me 13:57:24 +LeeF; got it 13:57:42 +kasei 13:58:03 AlexPassant has joined #sparql 13:58:19 is today a 90 minute call? 13:58:29 + +049261287aabb 13:58:40 Zakim, aabb is me 13:58:40 +SimonS; got it 13:58:53 +??P22 13:58:56 +??P24 13:58:57 Zakim, ?P22 is me 13:58:57 sorry, AlexPassant, I do not recognize a party named '?P22' 13:59:02 Zakim, ??P22 is me 13:59:02 +AlexPassant; got it 13:59:28 zakim, ??P24 is Garlik 13:59:28 +Garlik; got it 13:59:32 bglimm has joined #SPARQL 13:59:35 yep 13:59:39 Scribe: AlexPassant 13:59:42 Scribenick: AlexPassant 13:59:55 zakim, Garlik has LukeWM,SteveH 14:00:01 zakim, who's on the phone? 14:00:04 +LukeWM, SteveH; got it 14:00:06 On the phone I see AndyS, LeeF, kasei, SimonS, AlexPassant, Garlik 14:00:08 Garlik has LukeWM, SteveH 14:00:14 + +0186528aacc 14:00:25 Zakim, +0186528aacc is bglimm 14:00:30 +bglimm; got it 14:00:31 zakim, dial ivan-voip 14:00:32 ok, ivan; the call is being made 14:00:34 +Ivan 14:00:39 Zakim, mute me 14:00:42 bglimm should now be muted 14:01:12 +AxelPolleres 14:01:24 +EricP 14:01:40 zakim, mute me 14:02:06 Ivan should now be muted 14:02:12 Zakim, mute me 14:02:18 zakim, mute me 14:02:40 kasei should now be muted 14:02:42 + +1.540.412.aadd 14:02:48 AxelPolleres should now be muted 14:03:00 Zakim, aadd is me 14:03:06 muting with Zakim is going to be a hassle if it's always this slow... 14:03:10 Zakim, unmute me 14:03:26 zakim, unmute me 14:03:29 chimezie has joined #sparql 14:03:34 +pgearon; got it 14:03:41 zakim, mute me 14:03:44 AxelPolleres should no longer be muted 14:03:45 LeeF: 60 minutes call (instead of 90 minutes initially planned) 14:03:46 topic: admin 14:03:48 topic: admin stuff 14:03:51 PROPOSED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2009-09-08 14:03:56 Ivan should no longer be muted 14:03:59 q+ to ask if we should schedule 90min calls at 13:30Z 14:04:02 +Chimezie_Ogbuji 14:04:04 Chimezie_Ogbuji should now be muted 14:04:32 AxelPolleres: Regrets to be added to the previous minutes 14:04:42 q- 14:04:42 RESOLVED: Approve minutes at http://www.w3.org/2009/sparql/meeting/2009-09-08, noting Axel's regrets 14:05:06 All: Congrats to Axel :-) 14:05:07 axelsfamily++ 14:05:08 +1 to Axel 14:05:18 congrats 14:05:32 SteveH__ has joined #sparql 14:05:41 http://www.w3.org/2009/sparql/wiki/F2F2 14:05:45 LeeF: Update on the F2F status, please edit the wiki 14:06:10 ... Liaisons ? 14:06:10 q+ 14:06:14 ack ivan 14:06:28 ivan: RIF will go to CR very soon 14:06:39 ... conditional approval need to check with other groups for comments 14:06:43 ... but no comments expected 14:06:55 topic: FPWD 14:06:59 ... happened yesterday 14:07:02 on RIF: Axel still following up, LC comments closed, going to Cand. Rec ... ah, ivan reporting already 14:07:20 http://userpages.uni-koblenz.de/~sschenk/sparql-update-1.0/Overview.xml 14:07:31 SimonS: First draft of the update 14:07:39 ... issues to put it in cvs 14:07:58 q+ to ask about entailment regimes and updates (deletes) 14:08:02 http://www.w3.org/2009/sparql/docs/update-1.0/ 14:08:06 Zakim, unmute me 14:08:07 bglimm should no longer be muted 14:08:21 ... slightly outdated version, current status: started from the member submission + wiki content 14:08:33 ... discussions on the ML that not yet reflected on the doc (besides wiki content) 14:08:36 ... then clean up 14:09:04 LeeF: good idea would be to mention where are the current issues 14:09:20 ... referencing wiki page with open issues 14:09:26 SimonS: Adding an issue section 14:09:31 ... for issues unrelated to the features 14:09:37 ... and mention other issues inline 14:09:39 + +1.937.775.aaee 14:09:44 ummmm... i seem to be listed as an editor on that update-1.0 page(?) 14:10:30 update-1.0 is out of date 14:10:37 the other URL has the most up to date one, kasei 14:10:40 afaict 14:10:44 I wasn't available to help Simon on this iteration. Hoping to pick up the slack with the next one. 14:10:46 ok, just as long as that's been fixed :) 14:11:11 -> http://jena.hpl.hp.com/~afs/sparql-1.1.html is combined query and update. 14:11:51 SimonS: another internal iteration 14:12:02 LeeF: e-mail to be sent later 14:12:09 ... good practice for each document 14:12:42 ... discuss in the group, then publish 14:12:59 ... anybody not involved in X should review X, etc. 14:13:03 Prateek has joined #sparql 14:14:08 LeeF: anything needed from the group to query editors ? 14:14:24 SteveH__: stuff around select syntax 14:14:43 ... can be reolved after FPWD 14:14:47 +1 14:15:19 ack bglimm 14:15:19 bglimm, you wanted to ask about entailment regimes and updates (deletes) 14:15:41 bglimm: Enteilement regime : will it be left open on the update document ? 14:15:50 zakim, please unmute me 14:15:50 Chimezie_Ogbuji should no longer be muted 14:16:02 entailment regimes only appply to 'active' graphs, so I'm not sure if it is impacted 14:17:09 +1 to defer update semantics for higher entailments to entailment regime TF 14:17:15 ISSUE: How can other entailment regimes plug in their semantics to SPARQL/Update? 14:17:15 Created ISSUE-40 - How can other entailment regimes plug in their semantics to SPARQL/Update? ; please complete additional details at http://www.w3.org/2009/sparql/track/issues/40/edit . 14:17:46 Zakim, mute me 14:17:46 bglimm should now be muted 14:17:47 q? 14:18:04 topic: Service Description 14:18:04 LeeF: anything else about the documents themselves ? 14:18:22 LeeF: would like to reach a decision about the discovery mechanism 14:19:03 ... polls - speak if you are concerned by any of the proposals 14:19:08 I'd like to see options 8 and 7' (the RDFa one) be discussed in addition to the ones we've already had strawpolls on. 14:19:33 Zakim, unmute me 14:19:33 kasei should no longer be muted 14:19:48 ... kasei to resume current status 14:20:01 kasei: did my best to summariwe the pros / cons of current proposals 14:20:22 ... optin 1: link header in http and/or html 14:20:28 ... will link to a service description document 14:20:56 ... option 2 : HTTP OPTION 14:20:59 ... * vote against 14:21:03 ... 8 vote against 14:21:10 my objection to OPTION was cache related 14:21:30 ... difficult to query 14:21:41 ... option 7: content negociation on the endpoint URI 14:21:45 ... 4/1/4 14:21:58 ... biggest point against would be the reactions of HTTP purists 14:22:26 ... as not identical resources will be described (e.g. web form) 14:22:30 ... to be discussed: 14:22:45 ... option 7' RDFa version of the Service Description 14:22:56 ... encoded in the HTML page / form 14:23:12 q+ to ask if 7' has the opposite negative of option 7 14:23:19 q+ for RDFa and (non-X)HTML 14:23:25 ... option 8 14:23:31 ... implementation issues mentioned by Steve 14:23:50 LeeF: Straw poll needed 14:24:02 +[IPcaller] 14:24:30 ... option 7' opposite negative of the conneg option 14:24:42 ... return any sort of HTML just to encode the RDFa 14:24:43 and the drafback of the conneg option 14:24:46 ack me 14:24:46 LeeF, you wanted to ask if 7' has the opposite negative of option 7 14:24:46 This was my concern 14:24:56 q? 14:25:00 ack AlexPassant 14:25:00 AlexPassant, you wanted to discuss RDFa and (non-X)HTML 14:25:01 I agree with Lee, 7' is just a more detailed version of 7 14:25:02 q+ on Option 7' variations. 14:25:49 AlexPassant: issue of RDFa in non-X HTML 14:25:57 ... more political than technical decision 14:26:01 ack me 14:26:01 AxelPolleres, you wanted to comment on Option 7' variations. 14:26:08 .. will imply that endpoints / forms should be XHTML 14:26:12 q+ to talk about reverse proxies 14:26:18 ack AxelPolleres 14:26:26 q+ on RDFa and who authors forms 14:26:29 q+ 14:26:37 I suppose GRDDL is also an option. 14:26:45 q- 14:26:55 AxelPolleres: content negociation could return pure RDF or HTML+RDFa 14:27:08 q? 14:27:33 ack AndyS 14:27:33 AndyS, you wanted to comment on RDFa and who authors forms 14:27:53 AndyS: people may not have complete control on the URL of the endpoint 14:28:13 what I meant is, independent of whether content negotiation is supported, we allow both RDFa+HTML or pure RDF/XML as as service descriptions. 14:28:19 q+ to ask AndyS a followup question 14:28:27 ack kasei 14:28:27 kasei, you wanted to ask AndyS a followup question 14:28:44 kasei: will it affect the HTTP headers of the page 14:28:50 AndyS: would have less effet 14:28:57 ... data is more under control of the service provider 14:29:03 AxelPolleres, I don't understand what you mean by independent of content negotiation. How would this look without content negotiation? 14:29:05 +q to ask about RDFa processors in JS 14:29:10 q+ 14:29:18 ... proxy ownership issue 14:29:24 in ruling these out are we considering these mutually exclusive options? 14:29:28 s/.../kasei: 14:29:42 ack pgearon 14:29:42 pgearon, you wanted to ask about RDFa processors in JS 14:30:10 ack ivan 14:30:16 pgearon: any RDFa library in javascript ? 14:30:19 LeeF: yes 14:30:28 ivan: would like to put the RDFa issue aside for a moment 14:31:03 +1 to ivan, allowing RDFa is orthogonal to conneg 14:31:21 ivan: isn't this exactly why encoding in RDFa would be a good thing? just one representation of the resource, but with RDF in it. 14:32:23 LeeF: acknowledge the option of people returning form makes it a little easier 14:32:36 ... purist would argue that there is no for in the conneg 14:32:51 q? 14:33:08 ... separate proposal on the content negociation ? 14:33:18 i think 'self-describing' service resource is more conservative than pure content negotiation 14:33:26 q+ 14:33:30 why cannot we describe the form in the conneg. RDF ? 14:33:40 yes, strawpoll 14:33:51 ack kasei 14:34:04 q+ 14:34:04 kasei: possible proposal, RDFa without conneg 14:34:15 ... complaints of the purists as it's not the same resource 14:35:02 LeeF: do not see RDFa as required 14:35:12 ... strawpoll on connect with RDFa as an option 14:35:14 ack ivan 14:35:47 ivan: issue is not RDF/RDFa but what is the URI of the resource 14:35:56 though i don't think the alternative should require RDFa but be one where we *dont* do con neg but encourage RDFa 14:36:07 zakim, who's on the phone? 14:36:07 On the phone I see AndyS, LeeF, kasei, SimonS, AlexPassant, Garlik, bglimm (muted), Ivan, AxelPolleres, EricP, pgearon, Chimezie_Ogbuji, +1.937.775.aaee, [IPcaller] 14:36:10 Garlik has LukeWM, SteveH 14:36:19 zakim, who's speaking? 14:36:29 LeeF, listening for 10 seconds I heard sound from the following: AndyS (9%), LeeF (14%), Ivan (59%) 14:37:22 ... the resource = the service description 14:37:29 Would that be the Option which is now being discussed? "RDFa+HTML or RDF/XML at the service endpoint URL are both valid ways to serve the service description"? 14:38:03 I think Ivan is saying that unless the RDFa uses the endpoint URI explicitely, the issue of what we are identifying doesn't come up 14:38:16 zakim, aaee is Prateek 14:38:17 +Prateek; got it 14:38:19 zakim, IPCaller is Orri 14:38:19 +Orri; got it 14:38:31 q? 14:38:37 can I talk before the poll 14:38:38 ? 14:38:59 ack SteveH__ 14:38:59 SteveH__, you wanted to talk about reverse proxies 14:39:00 LeeF: strawpoll on option: SD served via conneg w/ pure RDF + RDFa in XHTML 14:39:16 SteveH__: reverse proxies become very common 14:39:22 ... esp w/ webservices, etc. 14:39:39 ... problem is that the SPARQL endpoint cannot get the URL it is called from 14:39:46 zakim, drop me 14:39:46 Ivan is being disconnected 14:39:48 -Ivan 14:39:53 ... unambiguously pointing back to the client 14:39:55 zakim, dial ivan-voip 14:39:55 ok, ivan; the call is being made 14:39:57 +Ivan 14:40:45 q+ to ask steveh whether this isn't an orthogonal issue? remember we discussed that the discription itself can use a bnode for the endpoint. 14:40:54 ericP: any way to talk about the URI even with the proxy 14:41:03 ack me 14:41:03 AxelPolleres, you wanted to ask steveh whether this isn't an orthogonal issue? remember we discussed that the discription itself can use a bnode for the endpoint. 14:41:05 Return a bnode, not (semi :-) 14:41:17 AxelPolleres: orthogonal issue ? 14:41:27 I can have a proxy URI which is where the endpoint is 14:41:35 ... if we get the URI we get the description ? 14:41:46 ... discussions on the ML if the description itself whould be on a blank node 14:42:22 I can see the general confusion with identifying an endpoint behind a rev proxy, but i think empty URI ref handles it (as well a s con neg) but as Axel said , it is orthogonal, because the descriptions don't have to reference the endpoint 14:43:31 SteveH__: nothing suggest you may have a / at the end 14:43:47 Prateek has joined #sparql 14:43:50 chime, the description *can* have a reference to the endpoint, but then rather by an attribute denoting the "official" endpoint URL, I'd say e.g. [a endpoint; endpointURL ] . 14:43:50 how is <> different in this respect from ? 14:43:58 LeeF: SD doesn't need to refer to the service with that URI ? 14:44:07 SteveH__: depends on the exact defintion of the SD 14:44:14 ... able to talk about the endpoint 14:44:28 ... e.g. fetch the service description: needs to know what we're talking about 14:44:56 q? 14:45:11 base + gives 14:45:30 kasei, ^^ 14:45:31 option 7': conneg + option to return RDFa for service descriptin in hTML 14:45:39 +1 14:45:42 +1 14:45:50 0 14:45:51 0 14:45:51 -1 14:45:52 +1 14:45:54 +1 14:45:55 +1 14:45:56 -1 14:45:56 0 14:45:59 +1 14:46:02 +1 14:46:04 0 14:46:31 0 14:47:21 RDFa, doesn't require use of accept headers 14:47:53 q+ 14:47:58 AlexPassant, you need to implement proper conneg to implement the sparql results spec + json anyway, and many people do that 14:48:36 SteveH__: I mean *real* proper conneg (see previous thread on the LOD mailing lists, really a pain to get it clean) 14:49:07 ack kasei 14:49:41 +1 to kasei 14:49:41 q+ 14:49:57 kasei: query form in the HTML is a different issue 14:50:09 ericP: pb is that we can have 2 descriptions of the same resource 14:50:24 kasei: pushing for the HTML form or for the service description ? 14:50:24 ack me 14:50:27 ack LukeWM 14:51:25 LeeF: option 8 - new protocol verb, e.g. endpoing?servicedesc 14:51:26 AlexPassant, I think that kasei had already said it. 14:51:41 s/endpoing/endpoint 14:51:42 LukeWM: can we just mandate that there is no query form. 14:51:53 ... issue: nothing to refer to the URI if using reverse proxy 14:52:29 q+ 14:52:51 ack SteveH__ 14:54:09 . /sparql?serviceDescription 14:54:18 -0.5 14:54:19 +1 14:54:21 +1 14:54:21 0 14:54:23 0 14:54:24 -1 14:54:25 +1 14:54:27 0 14:54:28 '0 14:54:28 0 14:54:29 0 14:54:30 0 14:54:32 0 14:55:33 0 14:55:42 my feeling is that if I want to use the query endpoint to query into the service description it seems to be a bit too convoluted to be worth it... 14:55:54 LeeF: most support for conneg / RDFa: i.e. return SD at the URL of the endpoint 14:56:06 +1 to LeeF's characterization of option 7 14:56:11 +1 14:56:17 we dont need to mention conneg 14:56:33 RDFa is a mention of best practice? Did I hear right? 14:57:03 and/or GRDDL or just serving plain RDF... 14:57:04 ... debate on recipes , which ones do we want to include 14:57:05 ...arguably an HTML form is the way you describe the endpoint access method in HTML 14:57:10 I do think the document should mandate RDFa 14:57:17 ... editorial latitude 14:57:22 +1 to plain RDF as 1st class choice. 14:57:29 +1 for Lees proposal and mention of RDFa in the doc 14:57:56 http://www.w3.org/2001/tag/doc/selfDescribingDocuments.html#RDFSection 14:58:08 topic: AOB 14:58:16 ISSUE-28 14:58:17 SimonS: issue with update + entailment 14:58:20 ... issue 28 14:58:53 ericP, sure 14:59:13 ... e-mnail to be sent in the week 14:59:15 Can you paste the URI? 14:59:19 yep, issues 28 and 40 are the same... 14:59:50 -Chimezie_Ogbuji 14:59:57 LeeF: please update the F2F and SELECT query syntax comments are welcome on the mailing list 14:59:59 better more than none:-) 15:00:01 -Orri 15:00:02 zakim, drop me 15:00:02 Ivan is being disconnected 15:00:04 -Ivan 15:00:04 -LeeF 15:00:05 bye 15:00:10 bye 15:00:13 bye 15:00:14 -Prateek 15:00:17 -bglimm 15:00:19 -AxelPolleres 15:00:20 Paul, could you stay for a second? 15:00:32 AlexPassant, thanks for scribing 15:00:35 ericP, do you mind if I listen in to this discussion? 15:00:40 not at all 15:00:41 -AlexPassant 15:00:45 please 15:00:45 cool 15:04:43 -SimonS 15:10:21 (still listening in) it's not the case that the endpoint necessarily dereferences a FROM 15:11:07 AlexPassant, can you take a shot at the minutes as per http://lists.w3.org/Archives/Public/public-rdf-dawg/2009AprJun/0406.html ? 15:15:47 SimonS has left #sparql 15:17:42 -AndyS 15:18:18 -pgearon 15:20:05 I'm off, bye 15:29:29 I'm off as well. ericP or SteveH__, any chance you could post some thoughts if you resolve any issues? :) 15:29:48 -kasei 15:41:33 SELECT * WHERE { GRAPH ?g { ?a a :Foo } } 15:41:40 SELECT * WHERE { GRAPH ?g { ?g a :Foo } } 15:42:55 SELECT * WHERE { GRAPH ?g { ?g a :SPARQLEndpoint } } 15:45:00 SELECT * WHERE { GRAPH ?g { ?g a :SPARQLEndpoint . ?s ?p ?o } } 15:49:44 FROM SELECT * WHERE { ?x :cpuLoad ?load } 15:51:17