15:02:58 <RRSAgent> RRSAgent has joined #rdf-wg
15:02:58 <RRSAgent> logging to http://www.w3.org/2012/11/21-rdf-wg-irc
15:03:00 <trackbot> RRSAgent, make logs world
15:03:00 <Zakim> Zakim has joined #rdf-wg
15:03:02 <trackbot> Zakim, this will be 73394
15:03:02 <Zakim> ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 57 minutes
15:03:03 <trackbot> Meeting: RDF Working Group Teleconference
15:03:03 <trackbot> Date: 21 November 2012
15:48:42 <rdf-wg> rdf-wg has joined #rdf-wg
15:50:00 <rdf-wg> zakim, who is on the call?
15:50:00 <Zakim> SW_RDFWG()11:00AM has not yet started, rdf-wg
15:50:01 <Zakim> On IRC I see rdf-wg, Zakim, RRSAgent, MacTed, ivan, mischat, AndyS, davidwood, gavinc, Arnaud, trackbot, manu1, manu, yvesr, sandro, ericP
15:50:53 <tbaker> tbaker has joined #rdf-wg
15:51:30 <tbaker> zakim, who is on the call?
15:51:30 <Zakim> SW_RDFWG()11:00AM has not yet started, tbaker
15:51:31 <Zakim> On IRC I see tbaker, Zakim, RRSAgent, MacTed, ivan, mischat, AndyS, davidwood, gavinc, Arnaud, trackbot, manu1, manu, yvesr, sandro, ericP
15:51:37 <Zakim> SW_RDFWG()11:00AM has now started
15:51:38 <Zakim> +EricP
15:52:07 <ericP> tbaker, are you trying to debug something?
15:52:20 <ericP> i've joined (started) the call in case that helps
15:53:18 <tbaker> Yes  an irc client on my iPhone
15:53:44 <Zakim> +??P2
15:53:52 <ericP> ahh, you're just asking questions of Zakim 'cause he's likely to answer
15:54:11 <tbaker> Will be only on irc today - in car.
15:54:13 <tbaker> Yes
15:56:35 <pfps> pfps has joined #rdf-wg
15:57:04 <Zakim> +pfps
15:58:27 <pchampin_> pchampin_ has joined #rdf-wg
16:00:38 <Zakim> + +1.540.898.aaaa
16:00:39 <sandro> RRSAgent, pointer?
16:00:39 <RRSAgent> See http://www.w3.org/2012/11/21-rdf-wg-irc#T16-00-39
16:00:53 <davidwood> Zakim, aaaa is mw
16:00:53 <Zakim> +mw; got it
16:01:01 <Zakim> +Sandro
16:01:05 <davidwood> Zakim, aaaa is me
16:01:05 <Zakim> sorry, davidwood, I do not recognize a party named 'aaaa'
16:01:13 <davidwood> Zakim, mw is me
16:01:13 <Zakim> +davidwood; got it
16:01:48 <Zakim> +??P14
16:01:50 <Zakim> +[IPcaller]
16:01:54 <AndyS> zakim, P14 is me
16:01:54 <Zakim> sorry, AndyS, I do not recognize a party named 'P14'
16:01:58 <ivan> zakim, dial ivan-voip
16:01:58 <Zakim> ok, ivan; the call is being made
16:02:00 <Zakim> +Ivan
16:02:00 <AndyS> zakim, IPCaller is me
16:02:00 <Zakim> +AndyS; got it
16:02:12 <pchampin_> zakim, ??P14 is me
16:02:12 <Zakim> +pchampin_; got it
16:02:40 <yvesr> scribenick: yvesr
16:02:40 <davidwood> Chair: David Wood
16:02:50 <davidwood> Zakim, who is here?
16:02:50 <Zakim> On the phone I see EricP, ??P2, pfps, davidwood, Sandro, pchampin_, AndyS, Ivan
16:02:52 <Zakim> On IRC I see pchampin_, pfps, tbaker, Zakim, RRSAgent, MacTed, ivan, mischat, AndyS, davidwood, gavinc, Arnaud, trackbot, manu1, manu, yvesr, sandro, ericP
16:02:52 <Zakim> +MHausenblas
16:02:52 <Zakim> +Arnaud
16:03:36 <Zakim> + +1.603.897.aabb
16:03:38 <Zakim> - +1.603.897.aabb
16:03:52 <Zakim> + +1.603.897.aacc
16:03:55 <AZ> AZ has joined #rdf-wg
16:04:01 <Souri> Souri has joined #rdf-wg
16:04:14 <sandro> sandro has changed the topic to: RDF-WG -- http://www.w3.org/2011/rdf-wg/ -- agenda will be http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2012.11.21
16:04:20 <markus> markus has joined #rdf-wg
16:04:29 <cygri> cygri has joined #rdf-wg
16:04:45 <cygri> zakim, mhausenblas is temporarily me
16:04:45 <Zakim> +cygri; got it
16:04:48 <AndyS> zakim, who is making noise?
16:04:50 <davidwood> PROPOSED to accept the minutes of the 7 Nov telecon:
16:04:50 <davidwood> 
16:04:50 <davidwood>    http://www.w3.org/2011/rdf-wg/meeting/2012-11-07
16:04:59 <Zakim> AndyS, listening for 10 seconds I heard sound from the following: davidwood (39%), Ivan (60%)
16:05:12 <yvesr> RESOLVED  to accept the minutes of the 7 Nov telecon
16:05:28 <ivan> zakim, mute me
16:05:28 <Zakim> Ivan should now be muted
16:05:36 <davidwood> Review of action items
16:05:36 <davidwood> http://www.w3.org/2011/rdf-wg/track/actions/pendingreview
16:05:36 <davidwood> http://www.w3.org/2011/rdf-wg/track/actions/open
16:05:58 <yvesr> davidwood: andys completed two actions
16:06:01 <Arnaud> zakim, unmute me
16:06:01 <Zakim> Arnaud was not muted, Arnaud
16:06:08 <Zakim> +??P17
16:06:20 <markus> zakim, ??P17 is me
16:06:20 <Zakim> +markus; got it
16:06:32 <yvesr> Arnaud: my action is about the rdf schema document, we were unsure danbri would be able to edit it
16:06:38 <yvesr> Arnaud: danbri said he will be able to do it
16:06:39 <davidwood> CLOSE ACTION-198
16:06:39 <trackbot> ACTION-198 Check with Dan what he wants to do with regard to editorship (remain editor, leave it to Arnaud, remain co-editor?) closed
16:06:50 <davidwood> CLOSE ACTION-206
16:06:50 <trackbot> ACTION-206 Put Turtle tests into W3C space. closed
16:06:56 <davidwood> CLOSE ACTION-207
16:06:56 <trackbot> ACTION-207 Do some documentation/README for the tests. closed
16:06:57 <yvesr> Arnaud: but is there much work to do, now we resolved on rdf:Seq?
16:07:01 <AndyS> AndyS actions - http://lists.w3.org/Archives/Public/public-rdf-wg/2012Nov/0223.html
16:07:05 <yvesr> Arnaud: So I think that takes care of my action
16:07:19 <yvesr> davidwood: Moving on to open actions
16:07:30 <Zakim> + +081165aadd
16:07:38 <yvesr> sandro: I want to talk about my IETF action
16:07:39 <sandro> http://www.iana.org/assignments/well-known-uris/well-known-uris.xml
16:07:39 <AZ> Zakim, aadd is me
16:07:39 <Zakim> +AZ; got it
16:07:56 <davidwood> CLOSE ACTION-82
16:07:56 <trackbot> ACTION-82 Draft well-known URI template and propose WG resolution that it is "stable" enough for IETF. closed
16:08:01 <yvesr> sandro: genid is now registered
16:08:10 <cygri> excellent!
16:08:27 <yvesr> davidwood: we'll move on to RDF Concepts
16:08:31 <yvesr> topic: RDF Concepts
16:08:54 <cygri> ISSUE-104?
16:08:54 <trackbot> ISSUE-104 -- Too many informative Notes in RDF Concepts -- open
16:08:54 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/104
16:09:00 <davidwood> Check for consensus on http://www.w3.org/2011/rdf-wg/track/issues/104 (Too many Notes)
16:09:00 <davidwood> 	1.	PROPOSAL: Delete four informative Notes from Concepts as described in http://lists.w3.org/Archives/Public/public-rdf-wg/2012Nov/0131.html
16:09:07 <yvesr> davidwood: let's move on to check on consensus on ISSUE-104
16:09:38 <yvesr> cygri: The background for the issue is that there is a lot of informative text in RDF Concepts
16:09:56 <yvesr> ... It was pointed out that this can be overwhelming
16:10:11 <yvesr> ... The places I identified were based on my own bias
16:10:33 <yvesr> ... The four that are listed in this email are: 1) Concerns with XML 1.1
16:10:47 <yvesr> ... 2) Confusing language tags with locales
16:10:57 <yvesr> ... 3) Some details about how to use schema assets
16:11:29 <yvesr> ... 4) Section regarding language tags
16:11:45 <yvesr> ... If no one wants to speak up for any of them, then I'll drop them
16:11:47 <pfps> These notes were added in response to external comments, I believe.   However, I still support removing them.  Just don't lose them, because there may be cries to reinstate them.
16:12:15 <yvesr> davidwood: There has been a discussion on the mailing list, this should be relatively uncontroversial
16:12:24 <yvesr> cygri: I agree with what pfps said on IRC
16:12:47 <yvesr> ... But I think 8 years later we might try again to remove them
16:12:57 <yvesr> ... If people are concerned about it then we can reinstate them
16:12:59 <zwu2> zwu2 has joined #rdf-wg
16:13:01 <yvesr> ... They don't add much value
16:13:24 <yvesr> davidwood: We could make a WG resolution that we are going to resolve ISSUE-104
16:13:51 <Zakim> + +1.603.438.aaee
16:14:03 <zwu2> zakim, aaee is me
16:14:03 <Zakim> +zwu2; got it
16:14:04 <yvesr> sandro: davidwood is saying the document should provide a pointer
16:14:27 <yvesr> davidwood: We should vote on this
16:14:37 <cygri> +1
16:14:39 <yvesr> +1
16:14:46 <ivan> +1
16:14:46 <pchampin_> +1
16:14:47 <davidwood> +1
16:14:47 <sandro> s/document/resolution/
16:14:47 <Arnaud> +1
16:14:52 <Souri> +1
16:14:55 <markus> +1
16:14:59 <sandro> +0.5  (I like more info....)
16:15:03 <pfps> +1
16:15:08 <tbaker> +1
16:15:10 <AndyS> +1
16:15:21 <zwu2> +1
16:15:44 <sandro> http://lists.w3.org/Archives/Public/public-rdf-wg/2012Nov/0131.html
16:15:51 <AZ> ok +1
16:15:53 <sandro> (that's the proposal we're voting on)
16:16:13 <yvesr> RESOLVED Delete four informative Notes from Concepts as described in http://lists.w3.org/Archives/Public/public-rdf-wg/2012Nov/0131.html
16:16:17 <davidwood> Check for consensus on ISSUE-110 (Term for g-box)
16:16:17 <davidwood> PROPOSAL 1: Informally call g-boxes “RDF sources” in Concepts
16:16:52 <davidwood> PROPOSAL 2: Keep the informal term “g-box” in Concepts
16:17:12 <yvesr> davidwood: This is an editorial issue in Concepts
16:17:31 <yvesr> ... There has been a discussion on the mailing list
16:17:39 <Zakim> +OpenLink_Software
16:17:44 <yvesr> ... Unless there are any objections, we should leave that as an editorial issue
16:17:44 <MacTed> Zakim, OpenLink_Software is temporarily me
16:17:44 <Zakim> +MacTed; got it
16:17:51 <yvesr> sandro: I'd like a WG decision on this
16:17:54 <AndyS> RDF sources
16:17:59 <AZ> 1
16:18:00 <sandro> 1
16:18:00 <yvesr> RDF sources
16:18:01 <pfps> RDF sources
16:18:04 <zwu2> RDF sources sounds better
16:18:07 <markus> RDF sources
16:18:11 <cygri> fine with either
16:18:17 <davidwood> RDF Sources (1)
16:18:18 <pfps> \me down with g-men and g-boxes!
16:18:20 <sandro> 1 for "rdf sources" 2 for "g-box"
16:18:25 <pchampin_> no preference for me
16:18:30 <ericP> 2
16:18:33 <FabGandon> FabGandon has joined #rdf-wg
16:18:40 <tbaker> +1 rdf sources
16:18:48 <Souri> 1
16:18:53 <Arnaud> 0, no real preference, both proposals have their pros and cons
16:19:13 <yvesr> davidwood: I believe the vote has come down on the side of RDF sources
16:19:21 <yvesr> ... If anyone objects, could they please speak up
16:19:23 <Arnaud> rdf sources sound better
16:19:44 <yvesr> ... Let's resolve ISSUE-110 by calling them RDF Sources in concepts
16:19:48 <Arnaud> g-box seems more precise/technical
16:19:53 <yvesr> RESOLVED:  Informally call g-boxes “RDF sources” in Concepts
16:20:32 <sandro> (closes issue-110)
16:20:35 <yvesr> pfps: g-box is RDF sources, g-snaps are RDF graphs
16:20:41 <yvesr> s/pfps/ericp
16:20:55 <sandro> RRSAgent, pointer?
16:20:55 <RRSAgent> See http://www.w3.org/2012/11/21-rdf-wg-irc#T16-20-55
16:21:23 <yvesr> davidwood: We need to close ISSUE-104 and ISSUE-110
16:21:30 <yvesr> ... We need to assign reviewers for RDF concepts
16:21:46 <AZ> I can review Concepts
16:22:15 <yvesr> cygri: Some background, I'll implement those changes and we still have time today we might talk about two outstanding issues, by the end of the day we should have the next Working Draft
16:22:24 <yvesr> ... That's why we need some reviewers for
16:22:55 <yvesr> cygri: There's only one obstacle in the way to last call, there are a couple of oustanding issues marked throughout the document
16:23:18 <yvesr> ... There is at least one more around fragment identifiers and dataset syntaxes
16:23:24 <yvesr> ... There might be some other issues as well
16:23:49 <yvesr> ... The main issue with going to Last Call while we still don't have first Working Drafts on closely related documents
16:24:12 <davidwood> Open issues related to Concepts: http://www.w3.org/2011/rdf-wg/track/products/8
16:24:16 <yvesr> ... As soon as we have those, I don't see any big problems
16:24:29 <yvesr> davidwood: In the tracker we still have 7 issues opened
16:24:34 <yvesr> ... Before we get to Last Call
16:24:47 <yvesr> cygri: An interesting one is the dataset transformation one (isomorphism etc.)
16:24:53 <yvesr> ... I'll bring them up on the mailing list
16:25:01 <yvesr> davidwood: We have AZ volunteering to review
16:25:13 <davidwood> Zakim, who is here?
16:25:13 <Zakim> On the phone I see EricP, ??P2, pfps, davidwood, Sandro, pchampin_, AndyS, Ivan (muted), Arnaud (muted), cygri, +1.603.897.aacc, markus, AZ, zwu2, MacTed
16:25:16 <Zakim> On IRC I see FabGandon, zwu2, cygri, markus, Souri, AZ, pchampin_, pfps, tbaker, Zakim, RRSAgent, MacTed, ivan, mischat, AndyS, davidwood, gavinc, Arnaud, trackbot, manu1, manu,
16:25:16 <Zakim> ... yvesr, sandro, ericP
16:25:33 <yvesr> davidwood: pfps would you be interested in reviewiewing concepts?
16:25:55 <yvesr> pfps: Yes, I can do it
16:26:03 <yvesr> ... It'll probably get done soon
16:26:49 <cygri> ACTION: pfps to review RDF-Concepts ED http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html
16:26:49 <trackbot> Created ACTION-210 - Review RDF-Concepts ED http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html [on Peter Patel-Schneider - due 2012-11-28].
16:27:02 <cygri> ACTION: AZ to review RDF-Concepts ED http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html
16:27:02 <trackbot> Created ACTION-211 - Review RDF-Concepts ED http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html [on Antoine Zimmermann - due 2012-11-28].
16:27:14 <davidwood> ACTION: AZ to review RDF-Concepts ED http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html
16:27:14 <trackbot> Created ACTION-212 - Review RDF-Concepts ED http://dvcs.w3.org/hg/rdf/raw-file/default/rdf-concepts/index.html [on Antoine Zimmermann - due 2012-11-28].
16:27:48 <yvesr> gavinc: On the Turtle document, the main work is to review the tests
16:27:52 <davidwood> Topic: Turtle LC
16:27:56 <yvesr> ... We also noticed some bug in the grammar
16:28:05 <cygri> AZ, i'll send you an email when today's resolutions are implemented, should be later today
16:28:09 <cygri> pfps, i'll send you an email when today's resolutions are implemented, should be later today
16:28:28 <yvesr> ... There are some machine ways of generating coverage of language features
16:28:33 <yvesr> ... We did that in SPARQL 1.1
16:28:39 <yvesr> ... It did help us identify issues
16:28:45 <AndyS> s/1.1/1.0/
16:28:47 <yvesr> ... I am working on it now for Turtle
16:29:11 <AndyS> +1 to eric checking coverage
16:29:53 <yvesr> gavinc: If we do the coverage tests, we can identify that we don't have any tests for example using a SPARQL-like PREFIX
16:30:19 <AndyS> s/gavinc/EricP/
16:30:43 <yvesr> ... How do I take an XML document and test a set of XML paths? There are multiple ways to do it - does anyone have a preferred way?
16:31:07 <yvesr> AndyS: The important bit is the report of the coverage
16:31:28 <yvesr> ericP: If people are happy that I run it on my laptop, then I'll go ahead
16:31:41 <yvesr> ... But please tell me if you want to be able to run it elsewhere
16:31:54 <yvesr> ericP: If we have a lot of tests that test the same features
16:32:01 <yvesr> ... It makes it harder for us to manage
16:32:24 <yvesr> AndyS: There will be duplication
16:32:48 <yvesr> ... Turtle is simpler, so managing the tests shouldn't be too bad
16:33:27 <yvesr> ericP: Whether we like redundancy is a question of taste
16:33:51 <yvesr> AndyS: Should we beginning to ask to use the test suite outside of the working group?
16:34:02 <yvesr> davidwood: I don't think it's too early to start engaging with people
16:34:22 <yvesr> ericP: It would be nice to have the test suite nailed down before going to last call
16:34:32 <yvesr> davidwood: Who would like to take an action to ask dajobe?
16:34:56 <yvesr> ACTION: davidwood to contact dajobe about Turtle and test suites
16:34:56 <trackbot> Created ACTION-213 - Contact dajobe about Turtle and test suites [on David Wood - due 2012-11-28].
16:35:02 <Zakim> + +33.4.92.96.aaff
16:35:27 <yvesr> davidwood: do we have anyone from JSON-LD today?
16:35:30 <yvesr> markus: I'm here
16:35:32 <davidwood> Topic: JSON-LD
16:35:34 <FabGandon> Zakim, aaff is me
16:35:34 <Zakim> +FabGandon; got it
16:35:34 <yvesr> topic: JSON-LD
16:35:53 <yvesr> markus: Yesterday we were able to resolve the last remaining issues for the syntax specification
16:36:01 <yvesr> ... The issues are resolved, the spec is being updated
16:36:09 <yvesr> ... There are a few minor open issues for the APIs
16:36:25 <yvesr> ... But mostly details around the algorithms
16:36:42 <yvesr> ... The spec has to be updated, and then it will be ready for another round of reviews
16:37:08 <yvesr> davidwood: cygri made some suggestions around the language about RDF in the JSON-LD specs
16:37:16 <yvesr> cygri: I had an action to provide some input there
16:37:37 <yvesr> ... One of the thing that happened in the mean time is that the relationship between the syntax and the api has been discussed
16:37:58 <yvesr> ... The data model will not be part of the syntax document, but part of what used to be the API Document (to be renamed)
16:38:14 <yvesr> ... What was an issue around the syntax document will be an issue around that other document
16:38:30 <yvesr> ... Within the JSON-LD community group there is a pretty good idea on how to resolve this
16:38:46 <yvesr> cygri: By next week we should get a more subtstantial update
16:39:35 <yvesr> markus: There has been some discussions around JSON-LD as a graph syntax or as a dataset syntax
16:39:37 <cygri> ISSUE-105?
16:39:37 <trackbot> ISSUE-105 -- Graphs, datasets, authoritative representations, and content negotiation -- open
16:39:37 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/105
16:39:50 <yvesr> ... We are not very sure where the RDF WG is headed in terms of dataset syntaxes
16:39:56 <yvesr> ... We're waiting on that decision
16:40:24 <yvesr> sandro: It will be difficult as it will be difficult to align
16:40:49 <yvesr> ... JSON-LD seems like it is a dataset syntax, but on the other hand datasets are not assertive in any sense
16:41:01 <yvesr> ... They don't convey information about the world the way a graph syntax does
16:41:40 <yvesr> sandro: JSON-LD will have to drive the response to that issue
16:42:14 <yvesr> cygri: What the WG decided not to define how an RDF dataset can be treated as a logical expression
16:42:31 <yvesr> ... Graphs can be treated as logical expressions, but not datasets
16:42:42 <yvesr> ... Whether that's a problem remains to be seen
16:42:57 <yvesr> sandro: I think JSON-LD needs to be treated as a logical expression
16:43:03 <ivan> json-ld is 'just' a syntax
16:43:18 <yvesr> AndyS: Where does JSON-LD needs logical assertions?
16:43:26 <yvesr> sandro: I think the use-case is data merging
16:43:35 <yvesr> ... You can't merge JSON
16:43:52 <yvesr> ... And you can't merge JSON-LD if you don't treat them as conveying RDF
16:44:12 <yvesr> AndyS: I am trying to find something that's a technical aspect of JSON-LD that this impacts
16:44:36 <yvesr> sandro: The default graph has special standing as a logical expression
16:44:42 <yvesr> ... where it doesn't in datasets in general
16:44:57 <sandro> Sandro: I think we need the default graph in a JSON-LD dataset to be a logical express -- like an RDF graph -- and unlike Datasets in general.
16:44:59 <yvesr> cygri: A question here would be around use-case
16:45:09 <yvesr> ... When would that be a problem?
16:45:30 <yvesr> sandro: The case is when you get data from a bunch of different JSON-LD sources and want to merge it
16:45:43 <yvesr> ... How do you know that the default graph is actually the contet
16:45:47 <yvesr> s/contet/content
16:45:49 <pchampin_> q+
16:45:58 <yvesr> AndyS: You could argue it's up to the application
16:46:00 <pfps> +1 to applications being in control
16:46:17 <yvesr> sandro: I guess my point is that iif it's up to the application, you could just use JSON
16:46:19 <ivan> +1 to applications, too
16:46:29 <pfps> Huh?  JSON-LD isn't any different from Trig, so there shouldn't be anything more there.
16:46:30 <pchampin_> q-
16:46:32 <yvesr> AndyS: The application should be in control of which graph to merge and which graph not to merge
16:46:38 <ivan> +1 to pfps
16:47:03 <yvesr> sandro: You need to be able to a data source using just the URI and know you're getting a graph from it
16:47:08 <yvesr> ... You don't get that from JSON-LD
16:47:26 <yvesr> ... The intention is that if the provider wants to give you a graph, they can
16:47:31 <pfps> After all, I might use JSON-LD in a context where the unnamed graph doesn't have any special precedence.
16:47:46 <yvesr> AndyS: We are talking in the absence of the key developers
16:47:57 <pfps> However, there is something here.  Suppose that you are looking for a graph, and you get a dataset.  What do you do then?
16:48:02 <yvesr> sandro: I just wanted to point out they shouldn't wait on us to resolve that issue
16:48:22 <pchampin_> :-D
16:48:25 <pchampin_> q+
16:48:51 <markus> could we specify in JSON-LD that we treat the default graph in that specific way or would that be at odds with something?
16:49:09 <markus> sandro: that might resolve the issue
16:49:13 <sandro> sandro: I think it would solve the problem for JSON-LD to say: you can treat this is as graph source, if you want, and when you do, you get the default graph.
16:49:23 <yvesr> cygri: What exactly is going on when you publish datasets on the web and how that compares with publishing graphs? It is a big issue, and should be high priority
16:49:35 <pchampin_> q-
16:50:11 <yvesr> davidwood: Let's move on to ISSUE-107 and ISSUE-109
16:50:21 <yvesr> topic: RDF Concepts
16:50:33 <cygri> ISSUE-107?
16:50:33 <trackbot> ISSUE-107 -- Definition of blank nodes (editorial-ish) -- open
16:50:33 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/107
16:50:38 <yvesr> davidwood: Let's look first at ISSUE-107
16:50:40 <davidwood> ISSUE-107 (Blank Nodes definition) has turned out to be more “interesting” than expected, so I propose to move it towards the end of the agenda as:
16:50:40 <davidwood> PROPOSAL: Change RDF Concepts section 3.4 and 3.5 with updated text proposed on http://www.w3.org/2011/rdf-wg/wiki/User:Rcygania2/B-Scopes
16:51:10 <yvesr> cygri: Issue 107 started out as some editorial comments on how blank nodes are worded in the 2004 spec
16:51:19 <yvesr> ... It has grown a bit beyong an editorial issue
16:51:31 <yvesr> ... As we have come to terms with that issue around scopes of blank nodes
16:51:40 <yvesr> ... It is not spelled out properly in the 2004 spec
16:51:53 <yvesr> ... The b-scopes proposal is one attempt to make that notion of scope explicit
16:52:07 <yvesr> ... It shouldn't change anything to any of the previous resolutions
16:52:25 <yvesr> ... What it does is that it defines the term 'scope' that other specifications can use
16:52:43 <yvesr> ... When do blank nodes need to be relabeled?
16:52:59 <yvesr> ... RDF documents (be a dataset or a graph) are their own scope
16:53:05 <yvesr> ... Beyond that, it's up to the implementation
16:53:18 <yvesr> ... I am not sure we have consensus this is what we should do
16:53:36 <ericP> q+ to ask if the issue text is still the proposal
16:53:43 <davidwood> ack ericP
16:53:43 <Zakim> ericP, you wanted to ask if the issue text is still the proposal
16:54:05 <cygri> http://www.w3.org/2011/rdf-wg/wiki/User:Rcygania2/B-Scopes
16:54:13 <yvesr> ericP: The text in the proposal - is that the final text? Or was it refined in the email chain?
16:54:28 <yvesr> cygri: Yes, it is - it's not what is in ISSUE-107 though
16:54:57 <yvesr> ... But the discussion on ISSUE-107 led to that proposal
16:55:09 <yvesr> ... This proposal would close 107 as well
16:55:29 <yvesr> ericP: I would suggest changing the first sentence - a blank node is a node without a label
16:55:56 <yvesr> ... Sorry, was looking at the wrong text
16:56:01 <AZ> q+
16:56:34 <yvesr> davidwood: My understanding is that blank nodes never need to be relabelled
16:56:45 <yvesr> ... Is that correct or has the discussion moved on?
16:57:09 <yvesr> cygri: Blank nodes need to be relabeled when moving to a different scopes and the blank node identifier is already in use in the new scope
16:57:28 <sandro> q+
16:57:37 <yvesr> ... This should be addressed in the proposal
16:58:02 <davidwood> ack AZ
16:58:23 <yvesr> AZ: I think the notion of scope is not something that should be in the abstract syntax
16:58:32 <yvesr> ... It should be addressed in Concepts
16:58:44 <yvesr> ... The notion of scope is inherent to the notion of blank nodes
16:58:50 <yvesr> ... They 'cary' their scope
16:59:15 <yvesr> ... The blank nodes get a scope whenever they are put in a document
16:59:35 <yvesr> ... You don't copy the abstract syntax - you copy documents
17:00:06 <yvesr> ... You can take the same bnode, put it in another document, it will be the same bnode but with a different scope
17:00:17 <yvesr> cygri: I don't think that works
17:00:41 <yvesr> ... For two reasons, 1) Yes it is correct scopes are important for documents but that's by far not the only place
17:00:48 <yvesr> ... e.g. in the RDB2RDF working group
17:00:59 <yvesr> ... One of the most difficult issue we had to solve was around blank nodes
17:01:33 <yvesr> ... Implementations can define their own scope, and some do
17:01:53 <yvesr> ... I don't think it's possible to solve the issue by just considering documents
17:02:17 <yvesr> ... 2) Mathematical structures are not copied around, but blank nodes are copied around, which is addressed by this proposal
17:02:50 <yvesr> AZ: I don't think you copy whatever is in the abstract syntax, you copy a representation of it
17:03:12 <yvesr> ... I meant documents in a very broad sense, which would include how RDF graphs are represented in memory for example
17:03:35 <yvesr> cygri: The reason why we have the abstract structure is to specify this broad concept of documents
17:03:40 <sandro> I wonder about the first sentence being more like "A blank node is an abstract syntax entity which corresponds one-to-one to pairs of blank node identifiers (which are Unicode strings) and scopes."  Perhaps semantically that's slightly cleaner.
17:03:46 <yvesr> ... What you're suggesting is to introduce a new layer
17:04:13 <yvesr> AZ: There is the abstract syntax, and all the concrete forms of represneting this abstract syntax, in-memory, serialisations, etc.
17:04:35 <sandro> q?
17:04:47 <yvesr> cygri: The reason this abstract syntax exist is to identify a generic way to talk about RDF graphs, across all representations
17:04:55 <AndyS> sandro's approach is interesting
17:04:56 <yvesr> ... The notion of scope of blank nodes fits exactly into that
17:05:12 <sandro> davidwood?
17:05:12 <yvesr> ... That's exactly why we have it
17:05:25 <yvesr> AZ: I don't think it should be attached to a bnode
17:05:39 <pchampin_> +1 cygri: as scope is a common feature of all concrete syntaxes, it has its place in the abstract syntax
17:05:39 <davidwood> yeah - 1 min
17:05:43 <yvesr> ... If you put a bnode in a graph, which is represented somewhere, then the bnode really gains its scope
17:05:59 <sandro> also maybe s/scope/blank node scope/
17:06:05 <yvesr> ... When you're talking just about triples, then I don't see why there would be a scope
17:06:28 <yvesr> davidwood: I think you both laid out your positions, there is disagreement on this issue
17:06:32 <davidwood> ack sandro
17:06:38 <yvesr> ... I would like to move the more detailed disccusion to the mailing list
17:07:01 <yvesr> sandro: I really like this approach in general - I just wonder a little bit about fresh blank nodes being defined
17:07:17 <yvesr> ... Maybe we can say something about a blank node being reused?
17:07:28 <cygri> sandro, define "not used" :-)
17:07:45 <yvesr> sandro: Maybe AZ can propose a change that would address that?
17:08:05 <yvesr> sandro: A blank node is a pair of bnode identifier and scope
17:08:20 <yvesr> ... I think weneed to address this question around reuse of bnode identifiers
17:08:57 <yvesr> cygri: I don't have a feeling of how much support this idea has
17:09:00 <AndyS> q+
17:09:03 <sandro> PROPOSED: We like http://www.w3.org/2011/rdf-wg/wiki/User:Rcygania2/B-Scopes
17:09:06 <sandro> +1
17:09:09 <pchampin_> +1
17:09:16 <yvesr> ... Some indication of how much support it has would be helpful
17:09:18 <ivan> +1
17:09:21 <markus> +1
17:09:22 <Arnaud> +1
17:09:24 <AZ> -1
17:09:27 <zwu2> +1
17:09:32 <yvesr> +0.5
17:09:33 <MacTed> +1 to the general idea.  not sure whether it's better to introduce "scope" in context of "bnode" or vice versa ... or put them parallel
17:09:47 <AndyS> 0
17:09:56 <Souri> +1
17:10:24 <yvesr> AndyS: I have some issues with the text as well
17:10:43 <yvesr> ... It starts to talk about scopes of identifiers, then goes on to talk about scopes of blank nodes
17:11:26 <yvesr> sandro: Could this be cleaned up with some tweaks?
17:11:33 <yvesr> AndyS: Yes, I think so
17:11:56 <yvesr> ... But I am still a bit uneasy there is something deeper going on
17:12:42 <yvesr> davidwood: Can you suggest changes to the text that address that issue around identifiers?
17:12:50 <yvesr> AndyS: No, because I think something deeper is going on
17:13:23 <sandro> has Pat weighed in on this, in email?
17:13:30 <yvesr> AndyS: A database is also a kind of document, it only matters when things go in and out of it
17:13:37 <davidwood> q?
17:13:41 <davidwood> ack AndyS
17:14:00 <cygri> sandro, it's basically PatH's “surfaces” proposal, repackaged
17:14:09 <yvesr> ... A way would be to make bnode have global scopes
17:14:11 <davidwood> sandro, yes, PatH has
17:14:11 <sandro> but has he commented on this text?
17:14:32 <yvesr> AndyS: I think we share similar concerns with AZ, maybe not to the same degree
17:14:52 <yvesr> davidwood: We'll have to end it here
17:15:15 <yvesr> davidwood: AOB, and we'll adjourn
17:15:21 <ivan> zakim, drop me
17:15:21 <Zakim> Ivan is being disconnected
17:15:22 <Zakim> -Ivan
17:15:24 <zwu2> bye & happy thanksgiving!
17:15:28 <Zakim> -FabGandon
17:15:31 <Zakim> -AZ
17:15:32 <Zakim> -EricP
17:15:32 <Zakim> - +1.603.897.aacc
17:15:33 <Zakim> -AndyS
17:15:33 <Zakim> -davidwood
17:15:34 <Zakim> -MacTed
17:15:34 <Zakim> -Sandro
17:15:35 <Zakim> -Arnaud
17:15:35 <Zakim> -cygri
17:15:37 <Zakim> -pfps
17:15:37 <Zakim> -markus
17:15:40 <Zakim> -??P2
17:15:46 <Zakim> -pchampin_
17:15:55 <MacTed> yvesr - http://www.w3.org/2011/rdf-wg/wiki/Scribes
17:17:08 <pchampin_> richard, got a minute?
17:17:55 <pchampin_> about your b-scope proposal, again
17:18:04 <pchampin_> something that occured to me while re-reading it:
17:18:24 <pchampin_> do you (intent to) constrain all the b-nodes in a graph to belong to the same scope??
17:18:39 <pchampin_> it seems to me that it is not the case currently
17:18:43 <pchampin_> and it would be odd...
17:18:51 <pchampin_> (if they didn't, I mean)
17:19:30 <cygri> pchampin_, there's no such constraint in the current proposal. a graph could contain bnodes from several scopes
17:19:40 <pchampin_> that's how I read it
17:19:52 <cygri> ... but it says that most systems are expected to manage a single scope
17:19:57 <cygri> ... and all documents are a single scope
17:20:05 <pchampin_> right
17:20:27 <pchampin_> so in any RDF file or "regular" system, that would be the case
17:20:29 <cygri> ... therefore, it's unlikely that this ever matters. i don't think adding the constraint would really change anything
17:20:37 <Zakim> -zwu2
17:20:38 <Zakim> SW_RDFWG()11:00AM has ended
17:20:38 <Zakim> Attendees were EricP, pfps, +1.540.898.aaaa, Sandro, davidwood, Ivan, AndyS, pchampin_, Arnaud, +1.603.897.aabb, +1.603.897.aacc, cygri, markus, +081165aadd, AZ, +1.603.438.aaee,
17:20:38 <Zakim> ... zwu2, MacTed, +33.4.92.96.aaff, FabGandon
17:20:40 <pchampin_> I agree
17:20:59 <AndyS> What about "local labels" to clearly separate from bnode globality.
17:21:26 <cygri> AndyS, i'm not sure what that means. aren't blank node identifiers local labels?
17:21:47 <cygri> and why should blank nodes even be global?
17:21:59 <sandro> we called them NodeID in RDF/XML....
17:22:40 <pchampin_> well, they *are* global in a way: the pair (b-node-id, b-scope) is itself globally scoped
17:23:16 <AndyS> Bnodes are global (see the discussion on the mailing list).  It's that representation that has a label in a doc.  "local label" is an attempt to stress that, Just "id" or "label" has confused people for ages.
17:24:10 <AndyS> NodeID was (hindsight!) a less than ideal name.  XML ids are doc scoped.
17:24:14 <cygri> AndyS, thinking of bnodes as global is a mistake, IMO. they are bound to a scope.
17:24:49 <cygri> ... same label in a different scope is a different blank node, always
17:25:17 <cygri> ... IOW, no two scopes ever share the same blank node (although they may share the same label)
17:25:46 <AndyS> Agree the last part but that's about labels , not the bnode
17:25:51 <FabGandon> FabGandon has left #rdf-wg
17:27:22 <AndyS> When I merge graphs and I have |bnode| :p 123 in one then the new graph has |bnode| : p 123  -- |bnode| denotes what it did before.   we relabel for serialziation reasons to stop local names clashing, (c.f. relative URIs)
17:29:25 <sandro> (I sometimes wish we defined n-triples as the "abstract" syntax of RDF.    Going more abstract than strings was probably a mistake.)
17:29:41 <cygri_> cygri_ has joined #rdf-wg
17:30:04 <pchampin_> Andy: that's a way to look at it; but what does it change if you replace it with another blank node?
17:30:05 <AndyS> sandro - yep - could define RDF as a symbol processor and keep everything.
17:30:07 <cygri> AndyS, pchampin_, sorry had a brief network problem
17:30:16 <pchampin_> Sandro: well, n-triples has order, which would be a pain
17:30:22 <cygri> RRSAgent, pointer?
17:30:22 <RRSAgent> See http://www.w3.org/2012/11/21-rdf-wg-irc#T17-30-22
17:30:41 <pchampin_> ... or would have solved the list issue :-D
17:31:01 <AndyS> pchampin_ , that's an entailment. Same *meaning*.
17:31:10 <pchampin_> well...
17:31:41 <cygri> AndyS, why do you say that the new label in the merged graph denotes the same blank node? i don't see what you gain by insisting on that
17:31:55 <pchampin_> I wouldn't
17:32:08 <pchampin_> On the contrary, I absolutely don't care which bnode it denotes
17:32:13 <pchampin_> whether it is the same or not
17:33:06 <cygri> AdnyS, doesn't that just perpetuate the problem that I can copy a blank node all the way to afghanistan and back, and then don't know whether the incoming blank node is the same or not?
17:33:14 <cygri> s/AdnyS/AndyS/
17:33:49 <pchampin_> he're is another story with bnodes, which may explain my position:
17:33:51 <AndyS> cygri - in the local context, may be handles on bNode.  (hence I prefer to talk about docs exchanged on web - but you argue for a defn of labels ouside of docs exhanged)
17:34:11 <AndyS> "An RDF graph is copied into a scope" mixes levels to my reading.
17:34:32 <pchampin_> I have this piece of code that queries a graph with SPARQL, then builds a 2nd SPARQL query with the results of the 1st one
17:34:39 <cygri> AndyS, how about "A blank node is copied into a scope"?
17:34:53 <pchampin_> querying the same graph, only it was easier that way with SPARQL 1.0
17:35:17 <pchampin_> it woreked perfectly as long as I was getting URIs, but it broke when the 1rst query returned a b-node
17:35:36 <AndyS> "The merge of two RDF graphs is the result of copying both graphs into a target scope." does not deal with merge into existing graph (typically usage?)
17:35:36 <pchampin_> because I can not use that b-node in my SPARQL query (that's just an anonymous variable)
17:36:26 <cygri> AndyS, everything can happen in the same scope and the definition still works.
17:36:33 <AndyS> pchampin_: in local APIs, that works just fine.  It's the serialization step that messes it up.
17:37:09 <pchampin_> yep; that's the annoying stuff: SPARQL is not equivalent to API calls
17:37:28 <cygri> AndyS, pchampin_, yup, you crossed a boundary so you got a new blank node. it has nothing to do with the old one in the store, as far as you can tell from the result graph
17:37:31 <pchampin_> bottom line: I'm used to think of bnodes as volatile things
17:38:03 <AndyS> cygri - globally unique ids is well understood technology hence I don't see gloabl as a problem - indeed a benefit.
17:38:04 <cygri> pchampin_, and the solution there is of course skolemization, because IRIs survive those boundary crossings
17:38:06 <pchampin_> so, ok, theoretically, two graphs can share bnodes, but really, who cares
17:38:23 <pchampin_> as long as you can't (always) rely on that
17:38:31 <AndyS> graph / Subgraph share bnodes and it matetr.
17:38:35 <AndyS> graph / Subgraph share bnodes and it matters.
17:38:47 <pchampin_> how?
17:39:09 <AndyS> cygri is right that a graph store is a scope - not the each graph of the store.
17:39:39 <cygri> AndyS, global bnode labels are an implementation choice. All installations of your software may share a single global scope, nothing wrong with that.
17:39:44 <AndyS> graph = set => def subgraph as subset of the triples.
17:40:44 <pchampin_> AndyS: again, from a theoretical point of view, that's right, and that would remain so; and I think Richard's proposal does not remove this
17:40:55 <AndyS> cygri - scopes and labels are all implementation. (the clue is in the "blank" :-) identity isn't name)
17:40:58 <pchampin_> on the other hand, I find it useful to make the notion of scope explicit
17:42:39 <cygri> AndyS, sure, but you keep saying that blank nodes are global. i'm saying that this is an artificial complication. they don't need to be global. you made an implementation choice that supports global blank nodes. that choice doesn't mean it's needed in the abstract syntax
17:43:03 <pchampin_> gotta go, guys
17:43:08 <pchampin_> thanks you for the interesting discussion
17:43:09 <AndyS> see email with pat.
17:43:23 <cygri> bye pchampin_
17:43:24 <AndyS> GTG - as I said I was 0.  The wording for me does not reflect my concern.  I made my point but I don't care enough and can't find the time for a mega thread.
17:43:24 <pchampin_> I'll try to write a proposal for copy and merge that (I hope) would please Antoine more
17:44:32 <cygri> ok, AndyS. consensus still seems to be far off on this one anyway
17:44:47 <AndyS> I think on-global is the complication.  Requires extra concept.
17:45:19 <AndyS> bye
17:47:02 <ivan_> ivan_ has joined #rdf-wg
20:45:24 <ivan> ivan has joined #rdf-wg
20:57:09 <Zakim> Zakim has left #rdf-wg
21:50:39 <gkellogg> gkellogg has joined #rdf-wg
23:59:44 <ivan> ivan has joined #rdf-wg