IRC log of rdf-wg on 2013-03-20

Timestamps are in UTC.

14:55:41 [RRSAgent]: RRSAgent has joined #rdf-wg
14:55:41 [RRSAgent]: logging to http://www.w3.org/2013/03/20-rdf-wg-irc
14:55:43 [trackbot]: RRSAgent, make logs world
14:55:43 [Zakim]: Zakim has joined #rdf-wg
14:55:45 [trackbot]: Zakim, this will be 73394
14:55:45 [Zakim]: ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 5 minutes
14:55:46 [trackbot]: Meeting: RDF Working Group Teleconference
14:55:46 [trackbot]: Date: 20 March 2013
14:56:44 [Guus]: Guus has joined #rdf-wg
14:57:38 [Guus]: zakim, this will be rdf
14:57:38 [Zakim]: ok, Guus, I see SW_RDFWG()11:00AM already started
14:57:46 [yvesr]: Zakim, who is on th ephone
14:57:47 [Zakim]: I don't understand 'who is on th ephone', yvesr
14:57:50 [yvesr]: Zakim, who is on the phone
14:57:50 [Zakim]: I don't understand 'who is on the phone', yvesr
14:57:54 [yvesr]: Zakim, who is on the phone?
14:57:54 [Zakim]: On the phone I see ??P9, Guus
14:58:03 [yvesr]: Zakim, ??P9 is me
14:58:03 [Zakim]: +yvesr; got it
14:58:05 [Guus]: trackbot, start meeting
14:58:08 [trackbot]: RRSAgent, make logs world
14:58:10 [trackbot]: Zakim, this will be 73394
14:58:10 [Zakim]: ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 2 minutes
14:58:11 [trackbot]: Meeting: RDF Working Group Teleconference
14:58:11 [trackbot]: Date: 20 March 2013
14:58:40 [Guus]: chair: David
15:00:28 [AZ]: AZ has joined #rdf-wg
15:00:28 [gavinc]: Zakim, who is on the phone?
15:00:28 [Zakim]: I notice SW_RDFWG()11:00AM has restarted
15:00:30 [Zakim]: On the phone I see yvesr, Guus, Sandro, GavinC, OpenLink_Software, bhyland
15:00:33 [Guus]: zakim, who is here?
15:00:33 [Zakim]: On the phone I see yvesr, Guus, Sandro, GavinC, OpenLink_Software, bhyland
15:00:35 [Zakim]: On IRC I see AZ, Guus, Zakim, RRSAgent, ScottB, gavinc, TallTed, davidwood, mischat, manu1, yvesr, manu, sandro, ericP, trackbot
15:00:38 [davidwood]: Zakim, bhyland is me
15:00:38 [Zakim]: +davidwood; got it
15:00:38 [markus]: markus has joined #rdf-wg
15:01:34 [Arnaud]: Arnaud has joined #rdf-wg
15:01:37 [gkellogg]: gkellogg has joined #rdf-wg
15:02:09 [TallTed]: Zakim, code?
15:02:10 [Zakim]: the conference code is 73394 (tel:+1.617.761.6200 sip:zakim@voip.w3.org), TallTed
15:02:14 [davidwood]: Zakim, who is here?
15:02:14 [Zakim]: On the phone I see yvesr, Guus, Sandro, GavinC, TallTed (muted), davidwood
15:02:16 [Zakim]: On IRC I see gkellogg, Arnaud, markus, AZ, Guus, Zakim, RRSAgent, ScottB, gavinc, TallTed, davidwood, mischat, manu1, yvesr, manu, sandro, ericP, trackbot
15:02:16 [Zakim]: +??P30
15:02:38 [Zakim]: +??P31
15:02:41 [AZ]: Zakim, ??P30 is me
15:02:42 [Zakim]: +AZ; got it
15:02:42 [gkellogg]: zakim, I am ??P31
15:02:43 [Zakim]: +gkellogg; got it
15:03:53 [davidwood]: Zakim, pick a victim
15:03:53 [Zakim]: Not knowing who is chairing or who scribed recently, I propose gkellogg
15:03:56 [yvesr]: i can do it - but phone quality is very bad
15:03:59 [ericP]: Zakim, please dial ericP-office
15:03:59 [Zakim]: ok, ericP; the call is being made
15:04:00 [Zakim]: +EricP
15:04:02 [gavinc]: -q sits
15:04:05 [gkellogg]: scribe: gkellogg
15:04:05 [Zakim]: +??P28
15:04:13 [markus]: zakim, ??P28 is me
15:04:13 [Zakim]: +markus; got it
15:04:40 [Guus]: zakim, who is here?
15:04:40 [Zakim]: On the phone I see yvesr, Guus, Sandro, GavinC, TallTed (muted), davidwood, AZ, gkellogg, EricP, markus
15:04:42 [Zakim]: On IRC I see gkellogg, Arnaud, markus, AZ, Guus, Zakim, RRSAgent, ScottB, gavinc, TallTed, davidwood, mischat, manu1, yvesr, manu, sandro, ericP, trackbot
15:05:22 [davidwood]: PROPOSED to accept the minutes of the 13 March telecon:
15:05:22 [davidwood]
15:05:22 [davidwood]: http://www.w3.org/2011/rdf-wg/meeting/2013-03-13
15:05:23 [Zakim]: +Tony
15:05:26 [zwu2]: zwu2 has joined #rdf-wg
15:05:27 [gkellogg]: davidwood: back to normal DST alignment next week.
15:05:43 [ScottB]: Zakim, Tony is temporarily me
15:05:43 [Zakim]: +ScottB; got it
15:05:58 [gkellogg]: gavinc: some edits required to minutes from last week.
15:06:05 [davidwood]: Review of action items
15:06:05 [davidwood]: http://www.w3.org/2011/rdf-wg/track/actions/pendingreview
15:06:05 [davidwood]: http://www.w3.org/2011/rdf-wg/track/actions/open
15:06:12 [AZ]: q+
15:06:26 [gavinc]: +q
15:06:49 [Zakim]: +Arnaud
15:06:51 [gkellogg]: ericp: I moved over and consolidated Turtle tests.
15:06:52 [davidwood]: ACTION-233?
15:06:52 [trackbot]: ACTION-233 -- Gavin Carothers to publish the consolidated test suite -- due 2013-03-06 -- OPEN
15:06:52 [trackbot]: http://www.w3.org/2011/rdf-wg/track/actions/233
15:07:19 [gkellogg]: gavinc: Now that they're consolidated, we need to fix license and links and tag them so they don't change.
15:07:36 [Zakim]: +zwu2
15:07:41 [gkellogg]: … Try to finish test suite issues this week.
15:07:54 [davidwood]: ACTION-231?
15:07:54 [trackbot]: ACTION-231 -- David Wood to create a wiki page to track Turtle CR comments and notify the editors. -- due 2013-02-27 -- OPEN
15:07:54 [trackbot]: http://www.w3.org/2011/rdf-wg/track/actions/231
15:07:59 [PatH]: PatH has joined #rdf-wg
15:08:42 [gkellogg]: ericp: Created a page to cover comments on Turtle and implementation reports.
15:08:48 [Zakim]: +PatH
15:09:10 [davidwood]: http://www.w3.org/2011/rdf-wg/wiki/Turtle_Candidate_Recommendation_Comments
15:09:20 [davidwood]: CLOSE ACTION-231
15:09:20 [trackbot]: Closed ACTION-231 Create a wiki page to track Turtle CR comments and notify the editors..
15:09:22 [Zakim]: +Souri
15:09:55 [cgreer]: cgreer has joined #rdf-wg
15:09:55 [Souri]: Souri has joined #rdf-wg
15:10:07 [davidwood]: PROPOSED to accept the minutes of the 13 March telecon: http://www.w3.org/2011/rdf-wg/meeting/2013-03-13
15:10:28 [ericP]: -> http://www.w3.org/mid/51246FAC.6080807@dajobe.org turtle test case license
15:10:34 [gkellogg]: gavinc: W3C has note on how to publish test cases that we need to conform to.
15:10:42 [Zakim]: +cgreer
15:10:43 [gkellogg]: … Everything needs to be dual-licensed.
15:10:52 [gavinc]: http://www.w3.org/2011/rdf-wg/meeting/2013-03-13
15:11:19 [markus]: looks good to me
15:11:24 [AZ]: looks fine
15:11:33 [davidwood]: RESOLVED to accept the minutes of the 13 March telecon: http://www.w3.org/2011/rdf-wg/meeting/2013-03-13
15:12:00 [gkellogg]: AZ: I reviewed RDF Semantics.
15:12:07 [davidwood]: ACTION-235?
15:12:07 [trackbot]: ACTION-235 -- Antoine Zimmermann to review RDF 1.1 Semantics -- due 2013-03-06 -- OPEN
15:12:07 [trackbot]: http://www.w3.org/2011/rdf-wg/track/actions/235
15:12:42 [AZ]: Review: http://lists.w3.org/Archives/Public/public-rdf-wg/2013Mar/0131.html
15:12:43 [davidwood]: CLOSE ACTION-235
15:12:43 [trackbot]: Closed ACTION-235 Review RDF 1.1 Semantics.
15:13:06 [gkellogg]: Topic: Semantics
15:13:11 [davidwood]: Three alternative approaches for fixing the blank node scope problem: http://lists.w3.org/Archives/Public/public-rdf-wg/2013Mar/0143.html
15:13:30 [davidwood]: Zakim, who is here?
15:13:30 [Zakim]: On the phone I see yvesr, Guus, Sandro, GavinC, TallTed (muted), davidwood, AZ, gkellogg, EricP, markus, ScottB, Arnaud (muted), zwu2, PatH, Souri, cgreer
15:13:34 [Zakim]: On IRC I see Souri, cgreer, PatH, zwu2, gkellogg, Arnaud, markus, AZ, Guus, Zakim, RRSAgent, ScottB, gavinc, TallTed, davidwood, mischat, manu1, yvesr, manu, sandro, ericP,
15:13:34 [Zakim]: ... trackbot
15:13:48 [gkellogg]: davidwood: Richard was planning on coming up with a proposal, but hasn't yet.
15:14:13 [PatH]: +q
15:14:18 [gkellogg]: … Pat proposed accepting suggestion #2
15:14:21 [gavinc]: -q
15:14:22 [davidwood]: ack AZ
15:15:07 [gkellogg]: AZ: I sent a proposal that added to Pat's proposals, but adds things I find more satisfying.
15:15:22 [gkellogg]: … A link to the comments is on the Wiki.
15:15:26 [davidwood]: ack PatH
15:15:28 [AZ]: http://www.w3.org/2011/rdf-wg/wiki/User:Azimmerm/Blank-node-scope-again
15:15:44 [gkellogg]: path: AZ's message raises a bunch of issues which require thought.
15:16:22 [gkellogg]: … pfps put together a proposal suggesting that we don't need to go into scopes in semantics at all.
15:16:42 [gkellogg]: … I think he's right and we can simplify a lot of this, taking us back closer to the 2004 document.
15:17:14 [gkellogg]: … Perhaps we should take the idea offline and carry on in email, but keep it off the document critical path.
15:17:37 [gkellogg]: … I also pointed out a few editorial changes I'd like done to the other documents to take up the slack.
15:17:53 [gkellogg]: … BNode IDs in different documents don't identify the same blank node.
15:18:21 [gkellogg]: … There's an issue on datasets: can a BNode ID in two different datasets identify the same node? I hope no, but if yes, there's some work to be done.
15:18:25 [sandro]: q?
15:18:30 [sandro]: q+ to answer Pat
15:19:25 [gkellogg]: … I think BNode scoping is closed as far as semantics is concerned. It presumes that other documents will make clear exactly when BNode identifiers identify the same node, It should be clear in NT/NQ that they don't identify nodes in other graphs.
15:19:25 [sandro]: +1 Pat: blank node identifiers in different documents never identify the same blank node.
15:19:47 [sandro]: for splitting, use skolemization.
15:19:55 [gkellogg]: … If it's possible to split a document and retain scope among them, we need to come up with something.
15:20:16 [gkellogg]: AZ: I agree with Pat's analysis.
15:20:22 [davidwood]: ack sandro
15:20:22 [Zakim]: sandro, you wanted to answer Pat
15:20:49 [gkellogg]: sandro: I agree with Pat. Two dataset documents can't, but can't to datasets?
15:21:00 [gavinc]: Huh, so what happens if I download Turtle1 from IRI1... when I read it back from the file system it no longer is the same graph as if I get Turtle1 again from IRI1? :\
15:21:08 [gkellogg]: … You could have a dataset which is a subset of another one, so that they could share nodes.
15:21:19 [pfps]: pfps has joined #rdf-wg
15:21:26 [TallTed]: TallTed has joined #rdf-wg
15:21:38 [gavinc]: Do we define a way to say these two documents represent the same graph?
15:21:41 [gkellogg]: PatH: yes, it may be required. Semantics needs a clear story about shared bnode scopes.
15:21:58 [gkellogg]: sandro: absolutely, BNode identifiers in two documents identify different nodes.
15:22:08 [tbaker]: tbaker has joined #rdf-wg
15:22:20 [gkellogg]: … The need to separate BNodes among different documents is part of why we added skolumization.
15:22:40 [gkellogg]: PatH: Someone could invent a serialization that had split scopes.
15:22:50 [davidwood]: q?
15:22:51 [AZ]: In my bscope proposal, I distinguish between arbitrary RDF triples in the abstract syntax, and "concrete triples" which are the triples that have a concrete realisation in a file or system
15:23:22 [AZ]: q+
15:24:12 [sandro]: PROPOSED: In all current known RDF syntaxes, blank node identifiers in two different documents always identify different blank nodes
15:24:33 [yvesr]: +1 - the inverse would break too many things
15:24:44 [gkellogg]: davidwood: perhaps "in all W3C syntaxes"
15:24:47 [sandro]: PROPOSED: In all current W3C (draft or REC) RDF syntaxes, blank node identifiers in two different documents always identify different blank nodes
15:25:25 [gkellogg]: AZ: I think there can be useful use-cases when someone want's to split a graph into several files. For that, there should be a format that allows us to do this.
15:25:43 [sandro]: q+
15:25:48 [davidwood]: ack AZ
15:25:54 [gkellogg]: … I thought that N-Triples could be a good choice for this. If you want to transmit a huge graph, it may be better to use several files.
15:26:11 [gkellogg]: … I would not like to say that all formats standardized necessarily have one scope.
15:26:20 [gkellogg]: … It also depends on what is meant by "one document".
15:26:24 [zwu2]: +1 to AZ, I was wondering what is a document
15:26:33 [davidwood]: ack sandro
15:26:56 [gkellogg]: sandro: I think that would allow a lot of accidental ways to share things that weren't intended.
15:27:20 [zwu2]: q+
15:27:20 [gkellogg]: … You probably would need to use UUIDs so you don't conflict, so why not us GenIDs? They're like nodes.
15:27:35 [davidwood]: ack zwu
15:27:40 [gkellogg]: davidwood: If I wanted to break a graph into multiple files, I just wouldn't use BNodes.
15:28:00 [gkellogg]: zwu: what does document mean?
15:28:09 [sandro]: I think "document" should mean Web Page.
15:28:09 [gkellogg]: sandro: I think it has to mean "web page".
15:28:13 [AZ]: q+
15:28:20 [gkellogg]: zwu: a single page identified by a URL?
15:28:20 [gavinc]: -0, totally not true but don't care as avoiding blank nodes in publication avoids the whole issue
15:28:37 [gkellogg]: davidwood: If you POST/PUT or PATCH it doesn't necessarily get a single URL.
15:28:44 [sandro]: +1 davidwood yes, POST complicates this.
15:28:48 [Souri]: q+
15:28:54 [gavinc]: The word DOCUMENT sucks :P
15:29:02 [gkellogg]: ericp: I wonder if we should keep the abstraction, so we can talk about documents that aren't necessarily RDF databases.
15:29:05 [yvesr]: AZ, for this use-case you'd probably compress the data and during that same step split it in several files
15:29:23 [davidwood]: ack AZ
15:29:27 [gkellogg]: zwu: I think "document" does not have a clear definition. It's important to clearly define what we mean by "document".
15:29:33 [yvesr]: AZ, so that could be handled by a post-processing/packaging job - doesn't need to be handled by the RDF format itself
15:29:41 [gavinc]: How about? In all current W3C (draft or REC) RDF syntaxes, blank node identifiers in two different representations always identify different blank nodes
15:29:59 [gkellogg]: AZ: I was wondering about when you have a steam of RDF data, which may not terminate. At some point in time, you'd like to say it is finished, and we're starting a new graph.
15:30:20 [gkellogg]: … It may be that two separate packets have the same BNode, because of an arbitrary split. How to handle this?
15:30:23 [PatH]: q+
15:30:25 [gavinc]: Document sucks, avoid the word document.
15:30:32 [sandro]: PROPOSED: Two different RDF graph serializations (g-texts) can not have any blank nodes in common, using any current W3C syntax (draft or REC)
15:30:34 [markus]: I would say this is application-dependent.. per default blank nodes are scoped per file.. if you need something else it's application-specific
15:31:01 [gkellogg]: davidwood: It seems to me, when we say "document" we mean two different things: sometime's their the same, sometimes different.
15:31:17 [gkellogg]: … It may refer to a document in a particular format (eg Turtle, RDF/XML).
15:31:41 [gkellogg]: … Other times, it might mean an input which may be discrete, but could be a stream.
15:31:43 [zwu2]: I like "serialization"
15:31:52 [gkellogg]: … A serialization ends when it has an EOF.
15:31:55 [sandro]: PROPOSED: Two different RDF graph serializations (g-texts) cannot parse to RDF Graphs that have any blank nodes in common, using any current W3C syntax (draft or REC)
15:32:15 [gkellogg]: … I think the source of confusion is that we sometimes think of it as bits in a ...
15:32:19 [sandro]: q?
15:32:21 [davidwood]: ack Souri
15:32:59 [sandro]: PROPOSED: Two different RDF graph serializations (g-texts) cannot parse to RDF Graphs that have any blank nodes in common, using any current W3C syntax (draft or REC). For use cases which want something like that, try genid.
15:33:11 [gkellogg]: souri: I agree with AZ, data does not come in one shot. The data could be billion's of triples that I'm not loading at the same time. There is a need to say that the BNode scopes are the same.
15:33:40 [gkellogg]: … Form a systems point of view, there should be an option to allow this from a practical point of view.
15:33:46 [gkellogg]: s/Form/From/
15:33:49 [AZ]: +1 souri
15:34:05 [sandro]: PROPOSED: Two different RDF graph serializations (g-texts) do not parse to RDF Graphs that have any blank nodes in common, using any current W3C syntax (draft or REC). For use cases which want something like that, try genid, or some non-standard local override.
15:34:08 [Arnaud]: +q
15:34:12 [gkellogg]: davidwood: too bad the use case wasn't identified earlier, but there seems to be strong support.
15:34:22 [AZ]: I mentioned this use case several times in the past in emails
15:34:28 [gkellogg]: … Could you not solve this problem using some unique identifier?
15:34:43 [Arnaud]: -q david just made my point :)
15:34:48 [Arnaud]: -q
15:34:59 [gkellogg]: souri: we could always solve it that way, but the dataset we're receiving is not something we created. It could be a 10-billion triple dataset we don't want to modify.
15:35:28 [cgreer]: sounds like scope to me
15:35:28 [gkellogg]: … It may not be advisable to replace BNodes with identifiers. It's like we need some notion of continuity.
15:35:40 [davidwood]: ack PatH
15:35:43 [sandro]: +1 souri some kind of "continuity" of blank nodes, some "namespace identifier" would be very useful/practical given other people's data feeds.
15:36:20 [pchampin]: pchampin has joined #rdf-wg
15:36:24 [gkellogg]: PatH: rather than use terminology like "documents", we can say that any method of transmitting RDF via a surface syntax, must record what the intended scope of the identifiers is.
15:36:42 [AZ]: +1 PatH
15:36:48 [gkellogg]: … For N-Triples, it's the document, for TriG, it's the dataset document. For other things, some sort of "End of Graph" signal.
15:36:58 [sandro]: pat: Any method of transmitting RDF content in a surface syntax must indicate/record what the intended scope of the blank node identifiers in that syntax is. For Turtle and TriG ... etc. it's the document/serialization. The scope of the blank nodes has to be indicated.
15:37:01 [sandro]: +1 pat
15:37:03 [ericP]: i think the aspect that we care about in defining a document is the bnode scope
15:37:07 [gkellogg]: … The scope of identifiers needs to be indicated, because they're local identifiers. Just what are they local to?
15:37:12 [ericP]: there are several specs which define the bnode scope
15:37:17 [Zakim]: +??P6
15:37:23 [gkellogg]: … I think talking about BNode scopes is the natural way.
15:37:34 [davidwood]: q?
15:37:34 [sandro]: q?
15:37:46 [ericP]: souri's case appears to me to move the bnode scope out of RDF standards, but still perfectly doable from a systems perspective
15:37:53 [zwu2]: or this can be handled by vendors
15:38:01 [gkellogg]: davidwood: I'm concerned that we're talking about other people's data that we don't want to modify.
15:38:21 [gkellogg]: … One idea would be to have some sort of serialization for this purpose.
15:38:27 [Souri]: There is a physical separation by space (files) and time for the same (logical) document: is there a way to identify the logical identity of the document
15:38:28 [sandro]: PROPOSED: Two different RDF graph serializations (g-texts) do not parse to RDF Graphs that have any blank nodes in common, using any current W3C syntax (draft or REC). For use cases which want something like that, try genid, or some non-standard local override, or some possible new syntax that allows specifying bnode scope.
15:38:35 [ericP]: +1 to RDF doesn't support this use case
15:38:37 [sandro]: q+
15:38:37 [ericP]: q?
15:38:37 [cgreer]: +1 david
15:38:44 [Arnaud]: +1 david
15:38:52 [gkellogg]: … In the general case, I have to say that RDF does not support this use case. Any individual implementation may, but I'm nervous about reverse-engineering the semantics for this purpose.
15:38:57 [gavinc]: "Merging two graphs treats the blank nodes in each graph as being existentially quantified in that graph, so that no blank node from one graph is allowed to stray into the scope of the other graph's surrounding quantifier. This is appropriate when the graphs come from different sources and there is no justification for assuming that a blank node in one refers to the same entity as any blank node in the other."
15:39:10 [gkellogg]: PatH: semantics just refers to the BNode scope, semantics comes out of that.
15:39:24 [zwu2]: one serialization can be materialized into multiple files, there is no conflict
15:39:25 [sandro]: PROPOSED: Two different RDF graph serializations (g-texts) do not parse to RDF Graphs that have any blank nodes in common, using any current W3C syntax (draft or REC). For use cases which want something like that, try genid, or some non-standard local override, or some possible new syntax that allows specifying bnode scope.
15:39:48 [ericP]: +1
15:39:51 [cgreer]: that's good +1
15:39:53 [gkellogg]: +1
15:39:53 [yvesr]: +1
15:39:54 [davidwood]: +1
15:39:55 [Arnaud]: +1
15:39:56 [zwu2]: +1
15:39:57 [AZ]: +0
15:39:57 [sandro]: +1
15:40:01 [Souri]: +1
15:40:03 [markus]: +1
15:40:11 [Guus]: +1
15:40:13 [pchampin]: +1 (modulo some minor rephrasing)
15:40:22 [PatH]: +1
15:40:27 [gavinc]: -1, TriG does ;) It produces RDF Graphs with blank nodes in common
15:40:27 [TallTed]: +1
15:40:51 [PatH]: it reads a bit brutal.
15:40:53 [Arnaud]: I think the standard way of addressing this is NOT to use blank nodes :)
15:41:16 [gkellogg]: davidwood: good catch. Clearly what was written wasn't what was intended.
15:41:17 [TallTed]: +1 Arnaud
15:41:21 [davidwood]: +1 Arnaud
15:41:27 [cgreer]: "do not parse to RDF datasets that have any blank nodes in common" ?
15:41:35 [gkellogg]: gavinc: I agree with the spirit of the proposal.
15:42:01 [gkellogg]: PatH: the problem is, it is the graphs which matter.
15:42:42 [gkellogg]: sandro: how about "graph or dataset serializations?"
15:42:51 [markus]: the proposal is right, it talks about graph serialization formats
15:42:53 [pchampin]: PROPOSED: bnodes generated from parsing two different RDF serializations (g-texts) are distinct for any current W3C syntax (draft or RDF).
15:43:09 [pchampin]: ... (followed by the "to be rephrased" explaination that Sandro added)
15:43:33 [gkellogg]: ericp: I think Oracle needs a non-standard override. "use scope 1 or scope 2"
15:43:51 [gavinc]: +1 prefer representation to serialization but editorial
15:44:54 [davidwood]: PROPOSED: bnodes generated from parsing two different RDF serializations (g-texts) are distinct for any current W3C syntax (draft or REC). For use cases which want something like that, try genid, or some non-standard local override, or some possible new syntax that allows specifying bnode scope.
15:45:11 [ericP]: +1
15:45:15 [gkellogg]: +1
15:45:16 [PatH]: +1
15:45:18 [zwu2]: +1
15:45:18 [yvesr]: +1
15:45:19 [AZ]: +1
15:45:27 [gavinc]: +1 prefer representation to serialization but editorial
15:45:29 [sandro]: +1
15:45:31 [Souri]: +1
15:45:32 [Arnaud]: +1
15:45:38 [sandro]: (without "g-texts")
15:45:43 [TallTed]: +1
15:45:44 [davidwood]: +1 without g-text
15:46:12 [davidwood]: RESOLVED: bnodes generated from parsing two different RDF serializations are distinct for any current W3C syntax (draft or REC). For use cases which want something like that, try genid, or some non-standard local override, or some possible new syntax that allows specifying bnode scope.
15:46:28 [gkellogg]: gavinc: requires updating 5 different syntaxes :(
15:46:47 [gkellogg]: … I think it affects all documents, but it is how things are already done.
15:47:16 [AZ]: excellent, we've made progress on bnode scope issues!
15:47:28 [gavinc]: "Given two blank nodes, it is possible to determine whether or not they are the same."
15:47:47 [gkellogg]: gavinc: does this text just go away from RDF concepts?
15:48:12 [gkellogg]: PatH: I don't think it matters, so leave it alone.
15:48:39 [Guus]: for AOB: admin issues wrt short names reports
15:48:51 [TallTed]: change "it is possible" to "it may be possible"?
15:48:52 [gkellogg]: ericp: is the concern if it needs to be scoped? (in a graph, in a dataset, …)
15:48:55 [AZ]: If two bnodes are the same, then they are not two bnodes!
15:49:20 [gkellogg]: Topic: JSON-LD
15:49:56 [gkellogg]: markus: basically, we had @base at the beginning, but dropped and and recently re-introduced it (@vocab too).
15:50:20 [markus]: {
15:50:21 [markus]: "@context": {
15:50:23 [markus]: "@base": "http://example.com/",
15:50:24 [markus]: "@vocab": "http://schema.org/"
15:50:26 [markus]: },
15:50:28 [markus]: "@id": "/people/markus",
15:50:28 [gkellogg]: … This adds some complexity, as there can be multiple base IRIs, which could confuse developers, knowing what is in scope at a given point in the document.
15:50:29 [markus]: "@type": "Person",
15:50:31 [markus]: "name": "Markus Lanthaler",
15:50:32 [markus]: "gender": "male"
15:50:34 [markus]: }
15:50:44 [markus]: {
15:50:46 [markus]: "@id": "http://example.com/people/markus",
15:50:47 [gkellogg]: … @base is used for things which are not properties, or values of @type.
15:50:47 [markus]: "@type": "http://schema.org/Person",
15:50:49 [markus]: "http://schema.org/name": "Markus Lanthaler",
15:50:50 [markus]: "http://schema.org/gender": "male"
15:50:52 [markus]: }
15:50:56 [gkellogg]: … @vocab is used for properties and types.
15:51:15 [gkellogg]: … @vocab allows us to use smaller contexts.
15:51:39 [gkellogg]: … @base is a bit less clear, but allows the use of relative IRIs but preserve the document base in spite of location.
15:51:56 [gkellogg]: … The question is, is it to confusing, or does the value out weight the potential confusion.
15:52:00 [pchampin]: q+
15:52:13 [gkellogg]: … Of course, most every other serialization supports @base. And @vocab is similar to an empty prefix in Turtle.
15:52:38 [davidwood]: ack sandro
15:52:42 [gkellogg]: … It was mentioned in recent reviews, and I've made some updates to clarify.
15:53:01 [Souri]: s/out weight/outweigh/
15:53:07 [ericP]: gavinc, i think turtle's fine WRT bnodes:
15:53:07 [gkellogg]: sandro: I'm okay either way; I think they should be there and they're not that confusing.
15:53:16 [ericP]: informative: "A fresh RDF blank node is allocated for each unique blank node label in a document. Repeated use of the same blank node label identifies the same RDF blank node."
15:53:19 [gkellogg]: … @vocab could be renamed, but it's not obvious what a better name would be.
15:53:19 [ericP]: normative: "BLANK_NODE_LABEL : The string matching the second argument, PN_LOCAL, is a key in bnodeLabels. If there is no corresponding blank node in the map, one is allocated."
15:53:22 [davidwood]: ack pchampin
15:54:13 [gkellogg]: pchampin: although confusing at first site, it's easily explained. The argument for schema.org has two arguments: it only needs to be created once. The other is, why not use an empty prefix?
15:54:24 [sandro]: "evolutation issues" are serious here, with schema.org.
15:54:25 [gavinc]: ericp, 'fresh' isn't defined ;) but yes
15:54:27 [ericP]: gavinc, i guess PatH could look at "fresh RDF blank node" and "one is allocated" to see if it aligns with semantics
15:54:30 [gkellogg]: … Is that not an acceptable trade-off considered the complexity @vocab introduces?
15:54:49 [gkellogg]: markus: Some JSON parsers don't support properties which are empty strings.
15:55:03 [gkellogg]: … You could say ":" is used to specify the empty prefix, but it's not clear.
15:55:12 [sandro]: maybe the keyword @default for the empty string?
15:55:26 [gkellogg]: davidwood: Is any action necessary by this WG?
15:55:43 [gkellogg]: markus: I don't think so, unless there are major disagreements.
15:56:25 [gkellogg]: … They are currently planned to be in the spec. @base is marked at-risk, as it was added late. There are some complexities for which the details could vary.
15:57:03 [gkellogg]: sandro: I like the idea of using ":" as a hack to specify that there is an empty prefix.
15:57:11 [gkellogg]: q+
15:57:38 [gkellogg]: markus: we distinguier between CURIEs and Terms because of the presence of a ":".
15:57:45 [sandro]: *shrug* I guess.
15:57:47 [davidwood]: ack gkellogg
15:58:40 [gkellogg]: gkellogg: @vocab and @base is highly symetric with RDFa.
15:58:54 [gkellogg]: davidwood: they need to get into LC, and stabilize.
15:59:03 [sandro]: q+
15:59:16 [gavinc]: The syntax has added how many keywords since last publication?
15:59:17 [gkellogg]: markus: the API spec needs some minor fixes, but the syntax spec is pretty stable.
15:59:34 [gkellogg]: … There are a few minor things to change in the API spec.
15:59:46 [davidwood]: ack sandro
15:59:49 [gkellogg]: davidwood: there was a deadline put on with the working group extension.
16:00:01 [gkellogg]: sandro: are we taking the API spec to REC in parallel.
16:00:04 [gkellogg]: markus: yes.
16:00:43 [gkellogg]: davidwood: in the extension request, we said that we would go to LC in march.
16:00:54 [Zakim]: -Arnaud
16:01:11 [gkellogg]: … Can you make editorial changes so that the WC can promote the doc to LC next week?
16:01:16 [gkellogg]: … That's what we agreed to do.
16:01:35 [gkellogg]: markus: We could do that, but don't we need an Algorithms review first?
16:02:04 [gkellogg]: davidwood: it looks like we won't make the schedule.
16:02:19 [gkellogg]: sandro: we could have the reviews done and decide to publish in a week.
16:02:28 [Zakim]: -ScottB
16:02:39 [davidwood]: Zakim, who is here?
16:02:39 [Zakim]: On the phone I see yvesr, Guus, Sandro, GavinC, TallTed (muted), davidwood, AZ, gkellogg, EricP, markus, zwu2, PatH, Souri, cgreer, pchampin
16:02:41 [Zakim]: On IRC I see pchampin, tbaker, TallTed, pfps, Souri, cgreer, PatH, zwu2, gkellogg, Arnaud, markus, AZ, Guus, Zakim, RRSAgent, ScottB, gavinc, davidwood, mischat, manu1, yvesr,
16:02:41 [Zakim]: ... manu, sandro, ericP, trackbot
16:02:52 [TallTed]: Zakim, unmute me
16:02:53 [Zakim]: TallTed should no longer be muted
16:03:06 [gkellogg]: davidwood: searching for another reviewer of the JSON-LD algorithms doc.
16:03:37 [markus]: json-ld.org/spec/latest/json-ld-api/
16:03:45 [markus]: http://json-ld.org/spec/latest/json-ld-api/
16:03:51 [gkellogg]: markus: we'll need to mirror from the json-ld.org doc to the W3C mirror.
16:04:05 [gkellogg]: … We sync manually from time to time, but the other is automatically updated.
16:04:25 [markus]: http://json-ld.org/spec/latest/json-ld-syntax/
16:05:16 [sandro]: http://json-ld.org/spec/latest/json-ld-api/ and http://json-ld.org/spec/latest/json-ld-syntax/
16:05:46 [gkellogg]: ACTION: zwu to review JSON-LD API document
16:05:46 [trackbot]: Error finding 'zwu'. You can review and register nicknames at <http://www.w3.org/2011/rdf-wg/track/users>.
16:05:53 [gkellogg]: ACTION: sandro to review JSON-LD API document
16:05:53 [trackbot]: Created ACTION-240 - Review JSON-LD API document [on Sandro Hawke - due 2013-03-27].
16:06:09 [gkellogg]: s/zwu/zhe/
16:06:21 [gkellogg]: ACTION: she to review JSON-LD API document
16:06:21 [trackbot]: Error finding 'she'. You can review and register nicknames at <http://www.w3.org/2011/rdf-wg/track/users>.
16:06:36 [davidwood]: ACTION: zhe to review JSON-LD API document
16:06:36 [trackbot]: Created ACTION-241 - Review JSON-LD API document [on Zhe Wu - due 2013-03-27].
16:06:45 [gkellogg]: ACTION: zhe to review JSON-LD API doc
16:06:45 [trackbot]: Created ACTION-242 - Review JSON-LD API doc [on Zhe Wu - due 2013-03-27].
16:07:25 [Guus]: q+
16:07:29 [davidwood]: ack Guus
16:08:03 [gkellogg]: guus: we have an issue to change short-names for documents.
16:08:14 [gkellogg]: … We have rdf11-semantics, and rdf-turtle.
16:08:23 [gavinc]: http://www.w3.org/TR/turtle/
16:08:28 [Zakim]: +MHausenblas
16:08:33 [cygri]: cygri has joined #rdf-wg
16:08:34 [gkellogg]: sandro: I think we only do this when names collide.
16:08:51 [sandro]: sandro: I would only put the 11 it's the "same" document.
16:08:55 [gkellogg]: guus: use 11 for updates, not for new docs.
16:09:21 [gkellogg]: sandro: does trig need rdf-? It's unambiguous.
16:11:11 [PatH]: I have to leave. Thanks for all the bnodes.
16:11:27 [Zakim]: -PatH
16:11:27 [sandro]: http://www.w3.org/TR/json-ld-syntax/">http://www.w3.org/TR/json-ld-syntax/ will redirect to http://www.w3.org/TR/json-ld
16:11:32 [gkellogg]: gkellogg: we'd like to rename the JSON-LD docs too to json-ld and json-ld-algorithms (or processing)
16:11:51 [zwu2]: thanks & bye
16:11:55 [Zakim]: -davidwood
16:11:55 [Zakim]: -Guus
16:11:56 [Zakim]: -cgreer
16:11:56 [AZ]: bye
16:11:57 [Zakim]: -zwu2
16:11:58 [Zakim]: -Souri
16:12:04 [Zakim]: -AZ
16:12:25 [gkellogg]: sandro: when it comes time to publish, we'll change the short-names.
16:12:50 [ericP]: heya
16:13:03 [Guus]: trackbot, end meeting
16:13:03 [trackbot]: Zakim, list attendees
16:13:03 [Zakim]: As of this point the attendees have been Guus, yvesr, Sandro, GavinC, davidwood, TallTed, AZ, gkellogg, EricP, markus, ScottB, Arnaud, zwu2, PatH, Souri, cgreer, pchampin,
16:13:07 [Zakim]: ... MHausenblas
16:13:11 [trackbot]: RRSAgent, please draft minutes
16:13:11 [RRSAgent]: I have made the request to generate http://www.w3.org/2013/03/20-rdf-wg-minutes.html trackbot
16:13:12 [trackbot]: RRSAgent, bye
16:13:12 [RRSAgent]: I see 5 open action items saved in http://www.w3.org/2013/03/20-rdf-wg-actions.rdf :
16:13:12 [RRSAgent]: ACTION: zwu to review JSON-LD API document [1]
16:13:12 [RRSAgent]: recorded in http://www.w3.org/2013/03/20-rdf-wg-irc#T16-05-46
16:13:12 [RRSAgent]: ACTION: sandro to review JSON-LD API document [2]
16:13:12 [RRSAgent]: recorded in http://www.w3.org/2013/03/20-rdf-wg-irc#T16-05-53
16:13:12 [RRSAgent]: ACTION: she to review JSON-LD API document [3]
16:13:12 [RRSAgent]: recorded in http://www.w3.org/2013/03/20-rdf-wg-irc#T16-06-21
16:13:12 [RRSAgent]: ACTION: zhe to review JSON-LD API document [4]
16:13:12 [RRSAgent]: recorded in http://www.w3.org/2013/03/20-rdf-wg-irc#T16-06-36
16:13:12 [RRSAgent]: ACTION: zhe to review JSON-LD API doc [5]
16:13:12 [RRSAgent]: recorded in http://www.w3.org/2013/03/20-rdf-wg-irc#T16-06-45
16:13:19 [markus]: I think we decided to keep json-ld-api and not change it to json-ld-algorithms, gkellogg, didn't we?