14:23:44 RRSAgent has joined #rdf-wg 14:23:44 logging to http://www.w3.org/2011/11/02-rdf-wg-irc 14:23:46 RRSAgent, make logs world 14:23:46 Zakim has joined #rdf-wg 14:23:48 Zakim, this will be 73394 14:23:48 ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 37 minutes 14:23:49 Meeting: RDF Working Group Teleconference 14:23:49 Date: 02 November 2011 14:52:29 SW_RDFWG()11:00AM has now started 14:52:36 +guus 14:53:14 +Scott_Bauer 14:53:54 Guus has joined #rdf-wg 14:54:14 +gavinc 14:54:55 zakim, dial ivan-voip 14:54:55 ok, ivan; the call is being made 14:54:57 +Ivan 14:56:39 Good morning 15:00:03 swh has joined #rdf-wg 15:00:10 AndyS has joined #rdf-wg 15:00:29 mischat_ has joined #rdf-wg 15:00:38 +??P3 15:00:44 Zakim, ??P3 is [Garlik] 15:00:44 +[Garlik]; got it 15:01:10 Zakim, [Garlik] has me, mischat 15:01:10 +swh, mischat; got it 15:01:11 +bhyland 15:01:15 + +1.707.318.aaaa 15:01:20 Zakim, bhyland is me 15:01:20 +davidwood; got it 15:01:41 +??P7 15:01:43 zakim, who is here? 15:01:43 On the phone I see guus, Scott_Bauer, gavinc, Ivan, [Garlik], davidwood, +1.707.318.aaaa, ??P7 15:01:45 [Garlik] has swh, mischat 15:01:45 -??P7 15:01:46 On IRC I see mischat_, AndyS, swh, Guus, Zakim, RRSAgent, Scott_Bauer, gavinc, MacTed, danbri, mischat, mox601, AndyS1, ivan, davidwood, manu1, manu, NickH, ericP, trackbot, sandro 15:01:52 +OpenLink_Software 15:01:52 PatH has joined #rdf-wg 15:01:59 Scribe: Scott_Bauer 15:02:06 Zakim, OpenLink_Software is temporarily me 15:02:06 +MacTed; got it 15:02:08 Zakim, mute me 15:02:08 MacTed should now be muted 15:02:09 +[IPcaller] 15:02:17 zakim, IPCaller is me 15:02:17 +AndyS; got it 15:02:31 zakim, who is here? 15:02:31 On the phone I see guus, Scott_Bauer, gavinc, Ivan, [Garlik], davidwood, +1.707.318.aaaa, MacTed (muted), AndyS 15:02:34 [Garlik] has swh, mischat 15:02:37 On IRC I see PatH, mischat, AndyS, swh, Guus, Zakim, RRSAgent, Scott_Bauer, gavinc, MacTed, danbri, mox601, AndyS1, ivan, davidwood, manu1, manu, NickH, ericP, trackbot, sandro 15:03:02 +PatH 15:03:07 cygri has joined #rdf-wg 15:03:24 AlexHall has joined #rdf-wg 15:04:01 +AlexHall 15:04:18 scribenick: Scott_Bauer 15:04:27 Topic: Admin 15:04:38 zakim, mute me 15:04:38 PatH should now be muted 15:05:18 +mhausenblas 15:05:23 zakim, mhausenblas is temporarily me 15:05:23 +cygri; got it 15:05:25 PROPOSED: accept last weeks minutes 15:05:54 http://www.w3.org/2011/rdf-wg/meeting/2011-10-19 15:06:02 guus: should be 19 rather than 5 15:06:10 … of October 15:06:32 mhausenblas owl:sameAs cygri? 15:06:35 (fixed link to old minutes in agenda) 15:06:40 RESOLVED: minutes are accepted 15:06:43 Souri has joined #rdf-wg 15:06:49 CGI734 has joined #rdf-wg 15:06:57 Topic: Action Items 15:07:02 zwu2 has joined #rdf-wg 15:07:24 +Souri 15:07:31 guus: Action-3 to be closed 15:07:43 +ericP 15:07:56 +zwu2 15:08:09 zakim, unmute me 15:08:09 PatH should no longer be muted 15:08:20 Looks like Sandro and Richard have a lot to do ;) 15:08:22 guus: quite a number of open items 15:09:06 guus: need to discuss the primer with Fabien 15:10:05 guus: will continue all for Richard, Sandro 15:11:17 sandro: issue three was an attempt to close issues at the end of the face to face. 15:11:34 … need to find student to go through old comments 15:11:51 … suggest we reopen issue 3 15:12:13 ISSUE-3? 15:12:13 ISSUE-3 -- Between us, we need to study the feedback we got via http://lists.w3.org/Archives/Public/www-rdf-comments/ on the previous round of specs (and errata) -- open 15:12:13 http://www.w3.org/2011/rdf-wg/track/issues/3 15:13:03 guus: if you could edit issue 3 15:13:18 zakim, mute me 15:13:18 PatH should now be muted 15:13:49 http://www.w3.org/2011/rdf-wg/track/actions/111 updated to be more clear. 15:14:00 Topic: Telecon next week 15:14:01 -ericP 15:14:18 +ericP 15:14:38 guus: time change noted for next week -- back to the same time. 15:14:56 ISSUE-71? 15:14:56 ISSUE-71 -- Reconcile various forms of string literals (time permitting) -- pending review 15:14:56 http://www.w3.org/2011/rdf-wg/track/issues/71 15:15:02 Topic: Issue 71 15:15:31 guus: amendment proposal from Jerermy Carrroll 15:15:51 zakim, unmute me 15:15:51 PatH should no longer be muted 15:16:11 … Richard is the amendment ok for you? 15:16:15 http://lists.w3.org/Archives/Public/public-rdf-wg/2011Oct/0385.html 15:16:53 … Jeremy is an expert in the language labels. 15:17:00 +1 to proposal 15:17:03 +q 15:17:09 cygri: this looks ok, it's just a clarification. 15:17:47 I find the wording a bit strange 15:17:56 +1 to PatH 15:18:00 PatH: I have no objections to the modification but I do to the other. 15:18:22 i guess that's why RIF uses "symbol space" instead of "datatype" 15:19:03 cygri: Two separate concepts that deal with datatype IRIS. Typed datatype iri and a lexicalform 15:19:31 PatH: Data type IRI used as an IRI? 15:19:34 -ericP 15:19:38 gavinc_ has joined #rdf-wg 15:19:41 cygri: it's just an IRI 15:19:45 +ericP 15:19:55 … nothing that requires it to be a datatype. 15:20:07 … not syntactically invalid 15:20:31 … the intention is that this is IRI is exceptional 15:20:40 q+ 15:20:59 … used in a datatype position but just an IRI 15:21:28 PatH: That seems to be wrong to put it in that position 15:21:35 Zakim, mute me 15:21:35 gavinc should now be muted 15:22:07 cygri: I don't think we can really improve on the design 15:22:25 … I don't' think we want to revisit the design 15:22:35 PatH: I agree 15:24:02 cygri: RDF concepts, section 5 on datatypes denoted by one or more uri references. There is nothing that requires them to be connected 15:24:25 Zakim, who is speaking? 15:24:35 swh, listening for 10 seconds I heard sound from the following: AndyS (9%), PatH (70%) 15:24:58 PatH: I find this design unacceptable 15:25:41 PatH: This is called a datatype thats not a datatype. 15:25:47 q+ to ask if we have formally decided on the URI as rdf:langString (helps SPARQL finish) 15:25:53 … let's take it off line 15:26:19 q+ 15:26:34 zakim, mute me 15:26:34 PatH should now be muted 15:26:36 AndyS: Is that going to be the URI? 15:27:00 -ericP 15:27:03 but that iri will not actually occur in the literal... 15:27:04 cygri: There was no discussion of the specific URI 15:27:35 cygri: Should it be called something different? 15:27:45 ack me 15:27:45 AndyS, you wanted to ask if we have formally decided on the URI as rdf:langString (helps SPARQL finish) 15:27:47 +ericP 15:27:48 ack me 15:28:00 no objections to the label.. 15:28:02 guus: I suggest we keep the label. Is that ok Andy? 15:28:19 q- 15:28:34 q- 15:29:29 AlexHall: Datatype IRI that's not and IRI reminds me of how REST defines the simple space. Some of these are also datatypes. 15:29:50 s/REST/RIF/ 15:30:02 Topic: issue 77 15:30:10 s/simple/symbol/ 15:30:16 ISSUE-77? 15:30:16 ISSUE-77 -- Should we mark rdf:Seq as archaic (cf ISSUE-24) -- open 15:30:16 http://www.w3.org/2011/rdf-wg/track/issues/77 15:30:33 guus: RDF seq and lists discussion 15:31:19 I don't think there's much consensus 15:31:56 AndyS: I don't see a consensus forming. 15:32:32 link to wikipage? 15:32:37 http://www.w3.org/2011/rdf-wg/wiki/Ordered 15:32:50 guus: First Proposal just call them sequence and list downplay as containers. Consensus on this? 15:33:14 Andy: I haven't seen push back or support. 15:33:22 people DO call them sequence and list, in fact. 15:33:40 yeah 15:33:41 I think there is some level of consensus on well formedness as well 15:33:43 zakim, unmute me. 15:33:43 PatH should no longer be muted 15:33:50 guus: Semantics proposal is just to remove the semantics. 15:34:32 q+ 15:34:34 well formedness for Seqs was pushed back at : two rdf:_1 to indicate equal rank 15:34:41 PatH: We may not need to remove that. We may just not need so many at once (containers) 15:34:45 -ericP 15:35:26 q- 15:35:31 +ericP 15:35:40 ack AlexHall 15:35:54 ack ivan 15:36:24 Ivan: The current semantics would define an infinite amount of axiomatic triples. 15:37:26 … we get interoperability issues. 15:38:10 … certain conclusions in semantics are used in those properties? 15:38:47 PatH: I agree is would be better if there was a clearer normative statement. 15:38:49 q+ 15:39:30 guus: We should amend the text. 15:39:40 Ivan: We should do this quickly 15:40:07 … sparql does define one approach to this. 15:40:28 q+ 15:40:29 … look at this and see if we can use it to define it. 15:40:40 -ericP 15:40:52 +ericP 15:41:34 -> the SPARQL document to look at http://www.w3.org/TR/sparql11-entailment/ 15:42:28 Action: PatH review the sparql specifications for consistency with rdf semantics for container properties 15:42:28 Created ACTION-114 - review the sparql specifications for consistency with rdf semantics for container properties [on Patrick Hayes - due 2011-11-09]. 15:42:45 +Sandro 15:42:47 zakim, mute me 15:42:47 PatH should now be muted 15:43:23 Ivan: We have one major user that uses the container properties -- Adobe 15:43:35 q+ 15:43:40 ack ivan 15:43:42 … Should we contact them regarding these changes. 15:43:49 Hi Sandro! 15:43:53 +1 to Ivan. We have been presuming that Adobe will care about terminology. 15:44:17 q- 15:44:25 i think there was a consensus not to deprecate or otherwise be rude about containers? 15:44:37 not consensus 15:44:49 OK 15:45:25 Action: sandro will discuss the issue of the container properties and deprecation with Adobe. 15:45:25 Created ACTION-115 - Will discuss the issue of the container properties and deprecation with Adobe. [on Sandro Hawke - due 2011-11-09]. 15:45:49 q+ 15:46:37 it means if they do use it they have to take a shower afterwards. 15:47:04 q? 15:47:11 Topic: Well Formed Lists 15:47:33 q+ 15:47:40 guus: Is there consensus on this 15:47:58 sandro: I think so -- question is what would we do with the definition. 15:48:02 zakim, unmute me 15:48:02 PatH should no longer be muted 15:48:40 SteveH: It's not ok to ask Adobe if its ok to deprectate 15:49:03 q+ 15:49:05 … the containers 15:49:24 q- 15:49:27 +q re wellformed lists. Is the idea to have wellformed triples or a different 'native' structure? 15:49:47 q- 15:49:58 AndyS: Lists cannot be shared. Can't have two triples pointing to the same object. 15:50:26 PatH: can you clarify that? 15:50:48 AndyS: It's linked specifically on the turtle page. 15:50:49 I think PatH it's that IF they are well formed triples they CAN be implemented as a 'native' structure 15:51:17 +ericP.a 15:51:26 -ericP 15:51:33 """ Each list node has exactly one inbound arc; for all but the first node, it's an rdf:rest arc from the previous node. """ 15:51:55 ok, clearly we need to check the definition very carefully. 15:52:19 guus: I don't hear anybody saying it is a problem. 15:52:28 "(Perhaps there is a clearer definition using more math.) " ;) 15:52:45 … leaving this to further list discussion. 15:52:47 gavinc, yes. in fact there are several... 15:53:05 zakim, mute me 15:53:05 PatH should now be muted 15:53:21 sandro: would like to approach Adobe for feedback on the possibility. 15:54:04 Topic: To mark Lists as archaic or not. 15:54:25 sandro: see Steve's summary 15:54:44 guus: If we are stuck we go with the minimal option. 15:54:59 … proposal one. 15:55:09 lists archaic?? 15:55:45 +1 to guus. minimal change. 15:56:18 Sandro you may need to self scribe I'm not hearing you well 15:56:53 sandro: should we get rid of one of them? 15:57:06 … sequences or lists? 15:57:27 not clear to me what the alternative(s?) are. 15:57:40 sandro: no one is happy with containers, no one is happy with collections. should we try to actually fix the problem? is there will in the wg to approach this? 15:57:55 q+ 15:58:03 ack me 15:58:11 ah. not obvious to me that this is a problem to solve here. 15:58:15 q- 15:58:31 +Sandro.a 15:58:53 -Sandro 15:59:07 cygri: The other proposals include a list ontology and a datatype that is a list of UI/s 15:59:24 +1 cygri 15:59:32 +1 to Richard: not really in scope for this WG 15:59:42 +1 15:59:49 … seems like this should be carefully considered and explored in some other venue. 15:59:53 good for an XG (or whatever the current thing is) 15:59:55 +1 to richard 16:00:06 ack me 16:00:33 +☃ to cygri 16:00:33 guus: Proposal one the most likely option. 16:01:18 sandro: We are taking a big step by endorsing turtle and it will be painful to have all these turtle lists out there. 16:01:26 i dont see any simple way to get this tidied up now. whatever we do will be a crock at this stage. 16:02:03 ... ( 1, 2, 3) ?type . 16:02:22 Turtle exists in the wild already - I worry about changing current meaning. 16:02:23 … could perhaps use parenthesis for rdf list 16:02:39 +1 to AndyS 16:02:43 list datatype, maybe. 16:02:47 +1000 to AndyS 16:02:56 +1 andy 16:03:04 ():List ():Seq with () as list MIGHT work. 16:03:34 lists can have sublists? 16:03:50 Topic: Named Graph Issues 16:03:50 PatH - yes 16:04:14 sandro has joined #rdf-wg 16:04:43 PatH, an object of rdf:first can be a list by Sandro's wording and by Turtle syntax. Unusual but legal. 16:04:44 +q what TriG draft currently says 16:04:50 guus: Consensus model for moving forward is to agree on the minimal model. 16:04:52 +q to say what TriG draft currently says 16:05:07 Zakim, unmute me 16:05:07 gavinc should no longer be muted 16:05:25 A graph statement pairs an IRI with a RDF Graph. It is intended that triple statements made about that IRI are being made about the graph. The triple statements that make up the graph are enclosed in {}. 16:05:50 zakim, unmute me 16:05:50 PatH should no longer be muted 16:05:57 +q 16:06:05 I like the approach Guus outlines. Document different practices. 16:06:05 ack gavinc 16:06:05 gavinc, you wanted to say what TriG draft currently says 16:06:09 ack gavinc 16:06:14 ack patH 16:06:26 PatH: The second sentence "made using that IRI" would be better 16:06:40 A graph statement pairs an IRI with a RDF Graph. It is intended that triple statements made using that IRI are being made about the graph. The triple statements that make up the graph are enclosed in {}. 16:06:40 zakim, mute me 16:06:40 PatH should now be muted 16:06:51 sandro: informal that the IRI denotes the graph 16:07:11 gavinc: We match that grammar in a number of places. 16:07:16 can i simplify 16:07:16 SELECT * WHERE { GRAPH { ?s ?o1 } 16:07:16 SERVICE { GRAPH { ?s ?o2 } } } 16:07:16 to 16:07:16 SELECT * WHERE { GRAPH { ?s ?o1 ; ?o2 } } 16:07:19 ? 16:07:39 Yeah, "using that IRI" not "*about* that IRI". 16:08:12 ericP: One persons notion of a graph doesn't have to match another's. Sandro seems to say they probably should. 16:08:36 I'm not even sure that's an ideal world 16:08:37 sandro: That's probably going farther than we can right now. 16:08:44 In a TriG document a graph IRI must not be used to label more then one graph. The IRI of a graph statement may be omitted. In this case the graph is considered the default graph of the RDF Dataset. 16:08:47 -∞ to those being equivalent 16:08:51 zakim, unmute me 16:08:51 PatH should no longer be muted 16:08:57 +q 16:09:04 Not the same - need extra assumptions. 16:09:16 gavinc: TRIG doesn't deal with transactions 16:09:41 can i simplify 16:09:41 SELECT * WHERE { GRAPH { ?s ?o1 } 16:09:41 { GRAPH { ?s ?o2 } } } 16:09:41 to 16:09:41 SELECT * WHERE { GRAPH { ?s ?o1 ; ?o2 } } 16:09:44 ? 16:09:55 IRIs have globsal scope, so they must be the same according to the semantics. 16:10:01 gavinc: It says you can't label them twice. 16:10:02 ack PatH 16:10:13 That is OK. 16:10:14 ericP: I'll retract this line of reasoning. 16:10:14 s/globsal /global / 16:10:33 what if one dataset is a canned state from 1 year ago? 16:10:37 I agree for versions of global that include everyone getting to have their own globe 16:10:40 or just hasn't been updated 16:10:54 yet 16:10:56 PatH: We have an IRI and it denotes a graph. Nothing about only inside a data set. 16:11:03 it's not the deployed RDF world either 16:11:26 it's not representative of how technology actually works 16:11:40 http://dvcs.w3.org/hg/rdf/raw-file/default/trig/index.html#sec-trig-intro current TriG draft 16:11:58 sandro: it would be nice if an IRI refers to the same graph but that's not how sparql works. 16:12:00 +1 to cygri -- not just SPARQL, just easier to write down currently. 16:12:20 … can't just mandate this (Consensus at the face to face) 16:12:21 ... in fact predates SPARQL. 16:13:02 PatH: I thought a resolution was made to have the IRI denote the graph. 16:13:38 History - SPARQL followed/consensus of subset of deployed usage even back then. 16:13:45 gavinc: No one could agree on denotes 16:13:50 +1 to AndyS 16:14:32 sandro: You can't stop the IRI's from denoting something in RDF, but does it denote the graph. 16:15:16 … Maybe something at the top of the TRIG document that says how this works. 16:15:35 I'm not hearing anyone complain about "It is intended that triple statements made using that IRI are being made about the graph." 16:15:52 guus: You could give guidance on what the relationship would identify. Can we reach consensus? 16:16:27 AndyS: The IRI denotes the graph is what I get out of the statement. 16:16:36 Not AndyS 16:16:48 q+ 16:16:58 PatH: The IRI denotes the graph is what I get out of the statement. 16:17:02 swh: What Gavin said doesn't ring any alarm bells. 16:17:04 tnx 16:17:24 ack AndyS 16:17:29 sandro: Maybe we can try to flush out the implications of gavin's text. 16:18:02 q- 16:18:40 Need to use IRI in triple and IRI-graph association consistently but it's easy to get it wrong. 16:19:16 AndyS: The chief problem is this is global. Don't say denote if that isn't' what's meant. 16:19:23 Sure, PatH but people can something else about that IRI and the world will be happily inconsistent 16:19:39 s/AndyS/PatH/ 16:19:47 e.g. : size "57 bytes" using to denote the g-snap 16:19:52 bye 16:19:54 -Souri 16:19:55 bye 16:19:56 -PatH 16:19:57 -[Garlik] 16:19:58 -cygri 16:19:59 bye 16:19:59 -Sandro.a 16:19:59 -MacTed 16:19:59 -davidwood 16:20:02 -AlexHall 16:20:03 -ericP.a 16:20:05 -Ivan 16:20:05 -gavinc 16:20:06 AlexHall has left #rdf-wg 16:20:16 - +1.707.318.aaaa 16:20:42 -AndyS 16:20:43 Me too 16:20:50