Chatlog 2012-10-17

From RDF Working Group Wiki
Jump to: navigation, search

See panel, original RRSAgent log or preview nicely formatted version.

Please justify/explain non-obvious edits to this page, in your "edit summary" text.

14:57:00 <RRSAgent> RRSAgent has joined #rdf-wg
14:57:00 <RRSAgent> logging to http://www.w3.org/2012/10/17-rdf-wg-irc
14:57:02 <trackbot> RRSAgent, make logs world
14:57:02 <Zakim> Zakim has joined #rdf-wg
14:57:04 <trackbot> Zakim, this will be 73394
14:57:04 <Zakim> ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 3 minutes
14:57:05 <trackbot> Meeting: RDF Working Group Teleconference
14:57:05 <trackbot> Date: 17 October 2012
15:01:19 <AlexHall> AlexHall has joined #rdf-wg
15:01:40 <davidwood> Zakim, this is rdfwg
15:01:40 <Zakim> ok, davidwood; that matches SW_RDFWG()11:00AM
15:01:56 <Zakim> + +1.443.212.aacc
15:02:10 <Zakim> -??P8
15:02:18 <Guus> zakim, who is here?
15:02:18 <Zakim> On the phone I see [IPcaller], ??P6, GavinC, Arnaud, +1.540.538.aaaa, +31.20.598.aabb, OpenLink_Software, AlexHall
15:02:22 <Zakim> On IRC I see AlexHall, Zakim, RRSAgent, davidwood, pchampin, Arnaud, tbaker, AndyS, Guus, gkellogg, MacTed, gavinc, yvesr, manu, manu1, mischat, trackbot, ericP
15:02:30 <davidwood> Zakim, aaaa is me
15:02:30 <Zakim> +davidwood; got it
15:02:31 <tbaker> zakim, IPcaller is probably me
15:02:31 <Zakim> +tbaker?; got it
15:02:34 <Guus> zakim, +31.20 is me
15:02:34 <Zakim> +Guus; got it
15:02:35 <Zakim> +??P8
15:02:39 <gkellogg> zakim, I am ??P8
15:02:39 <Zakim> +gkellogg; got it
15:02:44 <PatH> PatH has joined #rdf-wg
15:02:46 <pchampin> Zakim, ??P6 is me
15:02:47 <Zakim> +pchampin; got it
15:02:57 <MacTed> Zakim, OpenLink_Software is temporarily me
15:02:57 <Zakim> +MacTed; got it
15:02:58 <MacTed> Zakim, mute me
15:02:58 <Zakim> MacTed should now be muted
15:03:20 <davidwood> Chair: David Wood
15:03:30 <Zakim> +Sandro
15:03:54 <Zakim> + +1.603.897.aadd
15:03:56 <sandro> sandro has joined #rdf-wg
15:03:59 <Zakim> +PatH
15:04:26 <AndyS> AndyS has joined #rdf-wg
15:04:27 <Souri> Souri has joined #rdf-wg
15:04:48 <Souri> Souri has joined #rdf-wg
15:04:49 <davidwood> Scribe: Thomas Baker
15:04:58 <davidwood> scribenick: tbaker
15:05:10 <Zakim> +[GVoice]
15:05:13 <davidwood> PROPOSED to accept the minutes of the 10 Oct telecon:
15:05:13 <davidwood> 
15:05:13 <davidwood>    http://www.w3.org/2011/rdf-wg/meeting/2012-10-10
15:05:24 <davidwood> RESOLVED to accept the minutes of the 10 Oct telecon:
15:05:24 <davidwood> 
15:05:24 <davidwood>    http://www.w3.org/2011/rdf-wg/meeting/2012-10-10
15:05:37 <Zakim> +[IPcaller]
15:05:44 <AndyS> zakim, IPCaller is me
15:05:44 <Zakim> +AndyS; got it
15:06:08 <tbaker> Topic:Review of action items
15:06:38 <davidwood> Topic: FTF3
15:06:38 <tbaker> Topic: Next meeting
15:06:49 <davidwood> Agenda: http://www.w3.org/2011/rdf-wg/wiki/FTF3#Agenda
15:06:59 <davidwood> Objectives - http://www.w3.org/2011/rdf-wg/wiki/F2F3-objectives
15:07:39 <tbaker> David: Feel free to object - a work in progress.  Concept: focus on the documents we want to produce, bring them closer to publication.
15:08:38 <tbaker> ...JSON and Turtle.  Intention: go through documents, identify where we have immediate actions, what Notes we might want to produce.  Formalize that process. Comments?
15:09:02 <AndyS> Looked OK to me (but I'm not at F2F3!)
15:09:10 <gavinc> heh, it's still 1:15am to 9am :D
15:09:30 <mox601> mox601 has joined #rdf-wg
15:09:35 <tbaker> ... To people who intend to attend remotely from West Coast: trying to keep the schedule reasonable.
15:10:12 <davidwood> Topic: Provenance Constraints Review
15:10:14 <tbaker> Topic: Provenance Constraints
15:10:31 <MacTed> MacTed has changed the topic to: RDF-WG -- http://www.w3.org/2011/rdf-wg/ -- current agenda http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2012.10.17
15:10:43 <PatH> thanks to tbaker, im having trouble hearing.
15:10:46 <tbaker> David: We need to get back to the Prov WG - have to tell the WG something.  I'd like to circulate a proposal and see if we can agree.
15:10:54 <tbaker> ...Volunteers to create such a proposal?
15:11:22 <ericP> q+ to ask our current geek confidence on the mapping of bundles back and forth to named graphs
15:11:38 <davidwood> ack ericP
15:11:38 <Zakim> ericP, you wanted to ask our current geek confidence on the mapping of bundles back and forth to named graphs
15:11:49 <tbaker> ... Hoping to see something with a bit more guidance than [X].
15:12:14 <davidwood> s/[X]/a negative response/
15:12:25 <tbaker> Eric: What is our current confidence on how well bundles can be represented in NGs and how well NGs support the semantics that bundles would benefit from?
15:12:52 <tbaker> ... No particular guidance to represent NGs as things that can be re-used as bundles.
15:12:59 <sandro> q+ to address gbox/gsnap
15:13:13 <tbaker> David: GBox versus GSnap.  Because they capture a state in time: snaps.
15:13:23 <davidwood> ack sandro
15:13:23 <Zakim> sandro, you wanted to address gbox/gsnap
15:13:30 <tbaker> Eric: Librarians think of Snaps and invite other identifiers for Boxes.
15:14:00 <tbaker> Sandro: They think as static Boxes - distinction doesn't matter.
15:14:33 <tbaker> ... What they do in their examples is make it a Web page - technically mutable, but "don't change it".  GBox.
15:15:02 <tbaker> ... Don't think the GBox/Gsnap distinction matters in Provenance context.  They are thinking about as immutable.
15:15:11 <ericP> +1 to "they're thinkg of it as immutable"
15:15:29 <tbaker> David: In agenda, limited guidance on serializing RDF databasets, serializing as Trig.
15:15:35 <sandro> q?
15:15:56 <tbaker> Pat: My reading of Box - their chief interest is being able to refer to them in RDF.
15:16:09 <tbaker> Sandro: Disagree.  We're not endorsing, but not forbidding.
15:16:21 <tbaker> Pat: But we're not saying we're imposing that interpretation.
15:16:23 <MacTed> Zakim, unmute me
15:16:23 <Zakim> MacTed should no longer be muted
15:16:56 <tbaker> Ted: Been in both groups.  Qst is more: is Prov in conflict with what we are doing?  I think not - mostly it complements and overlaps.
15:17:27 <sandro> q?
15:17:45 <tbaker> David: ...which is why we can get back with a statement - guidance that it is complementary as long as they are thinking of their immutable bundles as RDF datasets that could be serailized with one of our forthcoming mechanisms.
15:17:56 <tbaker> Sandro: But they use plain Turtle or RDFa.
15:18:03 <tbaker> David: They may not know about new ones.
15:18:32 <tbaker> Eric?: If someone publishes a bunch of graphs, without re-writing, can be consider it to be a bundle?
15:18:43 <tbaker> Sandro: Described inside?
15:18:47 <davidwood> s/Eric?/Eric/
15:19:20 <sandro> s/Described inside?/do you mean the name graphs inside the dataset as being bundles?/
15:19:30 <tbaker> Eric: Probably. Sure - you can write down whatever you want in graphs probably not as useful as ... [...]
15:20:03 <tbaker> Sandro: Any guidance we give must be crisp.  We are muddy enough.  Suggest specific wording changes.
15:21:06 <tbaker> ... I'm not ready to suggest such a change.  Maybe suggest that a tutorial or note show examples.  We could commit to that.  Just an idea.
15:21:12 <davidwood> q?
15:21:15 <sandro> sandro: We could offer to include a prov/dataset example in some future document we do.
15:22:05 <tbaker> Sandro: More concrete: that we will include Prov dataset examples in one of our documents - one that talks about datasets.
15:22:07 <tbaker> David: Primer?
15:22:09 <Guus> yes, should be Primer
15:22:13 <tbaker> Guus: Definitely.
15:22:19 <sandro> sandro: I suggest that at least one of our examples of datasets use prov.
15:23:03 <Guus> q+
15:23:07 <ericP> PROPOSAL: all datasets (e.g. Trig documents) can be described as PROV Bundles
15:23:10 <tbaker> Pat: Prov world is strange.  Examples could be misleading.
15:23:12 <sandro> +1 PatH there's a danger that Prov's complexity confuses people trying to use datasets....
15:23:32 <davidwood> ack Guus
15:23:45 <tbaker> Guus: I thought that, but starting to appreciate.  So as long as we give context on purpose of vocabulary, okay to use.
15:24:00 <sandro> ericP, Bundle==graph; set of Bundles== dataset, I *think(
15:24:01 <AndyS> Agree with Pat :: -1 to prov as examples -- stick to something more universally understood (one example, yes; not the main theme)
15:24:26 <tbaker> David: Would hurt community to ignore Prov.  In implementer world, everyone maintains provenance info for things they call NGs.  Guidance we provide will affect implementations.
15:24:44 <tbaker> ... Interested why Andy agrees with Pat on this.
15:24:54 <PatH> prov =/= Prov.
15:25:31 <tbaker> Andy: Prov used, but WG has taken a particular slice of provenance space - could confuse people who see "provenance" differently.
15:25:50 <tbaker> ... Simple: "this was created by Bob".
15:25:59 <sandro> sandro: let's get behind Prov for provenance.
15:26:22 <PatH> No, lets read Prov first and see if it makes sense.
15:26:33 <tbaker> Eric: They came up with "simplistic provenance".  You can say as much or little as you know, that's fine.  Do not need to use the most complex example.
15:27:01 <tbaker> David: If we cannot provide guidance, we could hasten Prov demise (see Powder).
15:27:27 <Zakim> + +1.408.992.aaee
15:27:50 <tbaker> ... Suppose people were to implement Prov independently of what we do.  Does this implement things in the wild that would make Pat uncomfortable?
15:27:52 <AndyS> q+
15:28:18 <tbaker> ... We seem to have widening gulf between RDF/LD/OWL... implementations and the model theory.  Wonder if Prov's adoption forces that.
15:29:22 <tbaker> Pat: No. My problem with Prov, personally: littered with philosophical claims off-base.  Example of car disappearing here, appearing there. Prov has thrown away Snap/Box.
15:29:42 <pfps> pfps has joined #rdf-wg
15:30:43 <ericP> i think PatH's issue is that they specifically address static states. we in RDF land have identifiers for dynamic objects (like the state of a car).
15:30:58 <tbaker> Eric: Either you misread, or didn't read far enough.  One thing we struggled with: start with physical example before getting to ethereal stuff of Web.  Thing / Representation / Alternatie Repns / Serializations... scary.  Tree -> Chair -> Auction - a clear track.
15:31:18 <AndyS> s.Eric.MacTed.
15:31:26 <tbaker> ... Provenance of that stuff is the focus.  That model can be applied to ethereal, but unclear.
15:32:24 <tbaker> Pat: Main critique, backing off a bit: result is complicated and will therefore quietly die.  Will be in philosophical debates until end of time.
15:33:07 <tbaker> Eric: You can do scruffy instead of complex.  If you only know X, say that.  But if you have the details, can be used.
15:33:10 <davidwood> ack AndyS
15:33:16 <tbaker> Pat: Still think it will not be used in practice.
15:33:36 <ericP> s/Eric: You can do scruffy/MacTed: You can do scruffy/
15:33:46 <tbaker> Andy: Putting Prov on the critical path?  Would reopen all the NG things.
15:34:12 <tbaker> Sandro: Proposal only to use Prov vocab for Prov examples.
15:34:27 <tbaker> Andy: All datasets can be described as bundles [??].
15:35:04 <tbaker> Eric: Datasets can be expressed via Prov - wanted to test.  They invent snaps for the car in its intermediate states.
15:35:32 <tbaker> Andy: Keep them simple. If we get into discussing cars, rather have nothing.
15:35:52 <tbaker> Eric: Net determination: context-dependent - sounds like our...
15:35:57 <sandro> PROPOSED: We reply to the Prov WG that we don't see any technical incompatibilities between their work and RDF 1 or RDF 1.1 as we imagine it; we wish them well, but don't see much we can practically do to help.
15:36:02 <tbaker> Andy: On issue of examples....
15:36:10 <ericP> s/Eric: Net determination:/MacTed: Net determination:/
15:36:15 <davidwood> It seems to me that the PROV docs suggest something more akin to periodic or occasional state measurements.  You make a snap when you want one.
15:36:28 <tbaker> Sandro: Do we have consensus on this?  We do not see conflict with RDF 1.1 but do not see much to do to help.
15:36:28 <sandro> q?
15:36:39 <PatH> +1 sandro
15:37:12 <tbaker> David: Really like to see some kind of suggestion from us that we understand bundles in relation to Datasets and GSnaps.
15:37:18 <Guus> +1 in spirit, but phrasing pls a bit more neutral
15:37:31 <tbaker> Sandro: Do not see us getting closer.
15:37:54 <tbaker> Guus: I'll try to rephrase.
15:38:07 <AndyS> We can help with one or two simple examples but are then asking them to write them? write any example they want?
15:38:17 <tbaker> Scribe: Thomas Baker
15:38:23 <tbaker> Scribenick: tbaker
15:38:24 <MacTed> PROPOSED: We reply to the Prov WG that we don't see any technical incompatibilities between their work and RDF 1 or RDF 1.1 as we currently imagine it.  If a suitable example is eventually included in our docs, we will try to use PROV output therein.
15:38:42 <sandro> +1
15:38:52 <Guus> +1
15:38:55 <davidwood> +1
15:38:55 <tbaker> tbaker: +1
15:39:04 <pchampin> +1
15:39:05 <Arnaud> +1
15:39:11 <pfps> +0.5
15:39:12 <gkellogg> +0
15:39:14 <PatH> tbaker, many thnx. I have rock/hardplace: typing requires I use speaker, can't hear words on speaker :-(
15:39:30 <tbaker> ACTION: Guus to send resolved text to Prov WG.
15:39:30 <trackbot> Created ACTION-192 - Send resolved text to Prov WG. [on Guus Schreiber - due 2012-10-24].
15:39:33 <PatH> +1
15:39:50 <davidwood> RESOLVED: We reply to the Prov WG that we don't see any technical incompatibilities between their work and RDF 1 or RDF 1.1 as we currently imagine it.  If a suitable example is eventually included in our docs, we will try to use PROV output therein.
15:40:08 <davidwood> Topic: Turtle Tests
15:40:21 <davidwood> https://svn.apache.org/repos/asf/jena/Experimental/riot-reader/testing/RIOT/Lang/
15:40:30 <tbaker> David: Andy did good work on Turtle tests.
15:40:45 <gavinc> Yay, thanks AndyS! Tests are awesome (even if it involves Perl)... lets move on!
15:41:48 <PatH> Impressive :-)
15:41:49 <tbaker> Andy: Done as much as I planned.  They are all syntax-only tests.  Keeping the structure of... produce Ntriples.  Need to be checked. Used vocab, pre-SPARQL.
15:42:30 <tbaker> David: Gavin, you have outstanding actions re: internationalization.
15:42:44 <tbaker> Gavin: Editors agree on language tag stuff.
15:42:55 <tbaker> Eric: Am writing it up as we speak.
15:43:13 <tbaker> ... I'm free to send to X group as long as Gavin and I agree, right?
15:43:23 <tbaker> ... will report back what I sent.
15:44:01 <tbaker> Guus: Responses from commenters?
15:44:06 <tbaker> Gavin: None.
15:44:22 <tbaker> Guus: Prod - silence is consent.
15:44:30 <tbaker> Gavin: Happy to do.
15:44:51 <davidwood> Topic: TriG Syntax
15:44:52 <tbaker> Guus: Re: timing: have something by....
15:45:21 <davidwood> Discuss whether in our dataset syntax, triples of the dataset's default graph MUST or MAY be surrounded by curly braces.
15:45:22 <davidwood> Related question: How can a human tell a Turtle document from a TriG document?
15:45:31 <tbaker> David: Leftover from several weeks ago.  Curly braces in default graphs.  Did we cover?
15:45:38 <tbaker> Sandro: Not covered yet.
15:45:51 <tbaker> Unknown person XXX says 'but emerging consensus on "MUST"'.
15:45:55 <tbaker> David: Propose this?
15:46:01 <gavinc> PROPOSED: In TriG, triples of the dataset's default graph MUST be surrounded by curly braces.
15:46:05 <sandro> +1
15:46:11 <sandro> +0
15:46:20 <PatH> +/-0
15:46:25 <sandro> (oops!   meant 0)
15:46:31 <gavinc> +1
15:46:33 <AndyS> +0.9
15:46:36 <tbaker> tbaker: 0
15:46:39 <Arnaud> 0
15:46:41 <davidwood> +0
15:46:41 <gkellogg> +0
15:46:42 <Souri> +1
15:46:46 <ericP> +.5
15:47:05 <MacTed> {+1}
15:47:10 <AndyS> zakim, it's between zero and one
15:47:10 <Zakim> I don't understand 'it's between zero and one', AndyS
15:47:12 <PatH> Zakim has a sense of humor?
15:47:38 <davidwood> RESOLVED: In TriG, triples of the dataset's default graph MUST be surrounded by curly braces.
15:48:13 <davidwood> PROPOSED: Implementations that parse and store information from TriG documents MAY turn the TriG default graph into a named graph with a name chosen in an implementation-dependent way.
15:48:37 <PatH> q+
15:48:40 <tbaker> David: Next proposal.... Came from discussion with implementers re: how they wanted to handle things.
15:48:43 <davidwood> ack PatH
15:49:05 <AndyS> q+ to ask why is this is decision we need to make
15:49:18 <tbaker> Pat: Can hear the worms in the can. When you change default graph to named graph, do you have a new dataset?
15:50:00 <tbaker> Sandro: Pragmatic thing: if I send you a dataset, reasonable to load one that differs in this way.  If we had a tool to load a dataset from a Trig file, would be part of conformance.
15:50:15 <tbaker> Pat: To me, sounds like implmementation weeds.  Should not even be discussing.
15:50:17 <gavinc> +1 to not talking about this
15:50:40 <tbaker> Sandro: Like RDF Contexts - when you skolemize... But agree it is about impelmentn details.
15:50:50 <sandro> s/Contexts/Concepts 1.1/
15:51:02 <davidwood> ack AndyS
15:51:02 <Zakim> AndyS, you wanted to ask why is this is decision we need to make
15:51:07 <tbaker> Pat: Easy to read as saying that it is same dataset when you make this change, or that they are operationally indistinguishable...
15:52:06 <tbaker> Andy: Don't understand why we need to mention this possible.  When you send data to someone, they may do something with it.  Add extra triples.  An expectation.  Trig is about moving data from A to B.
15:52:34 <tbaker> David: Can we turn on head and say that when a serialin is parsed, it becomes a new dataset?  Saying this?
15:52:39 <tbaker> Sandro: Don't like.
15:53:05 <davidwood> Isn't that what PatH and AndyS just proposed?
15:53:10 <gavinc> +q to ask Sandro or whoever added this proposal WHY we need to say this?
15:53:27 <tbaker> ... Let's look at what test suite would say about Trig.  Conformant Trig parser - turns default graph into NG.  Will that pass test suite??
15:53:40 <tbaker> ... Close enough?
15:53:57 <tbaker> Andy: Define what we have done by [] a document from one place to another.
15:54:15 <tbaker> Sandro: Not advocating, just trying to understand.
15:54:22 <sandro> sandro: I'm not actually advocating this, just trying to understand it.
15:54:46 <PatH> we defined graph equivalence so we could talk about that re-identifying blank nodes. I would like us to define something like that for this, if we decide it.
15:54:56 <sandro> +1 PatH
15:55:01 <tbaker> David: Original proposal came from Ivan.  Stand by my comment.  Didn't appear out of nowhere.  Not sure we need to make a decision about this.  "Minus-zero-ish" about this.
15:55:10 <davidwood> q?
15:55:20 <davidwood> ack gavinc
15:55:20 <Zakim> gavinc, you wanted to ask Sandro or whoever added this proposal WHY we need to say this?
15:55:25 <gavinc> ack
15:56:16 <sandro> manu, manu1, gkellogg, are you up for discussing how JSON-LD might handle or not handle the idea of a default graph?
15:56:36 <gkellogg> sure
15:56:49 <tbaker> Pat: [see comment above) - If we are going to allow a modification to a dataset to "not count" under certain circumstances, then we should define what we mean by equivalence.
15:56:52 <sandro> q+
15:56:56 <gavinc> eh, that's not part of TriG
15:57:26 <tbaker> ... Not sure why anyone would want to do this to graphs.  We as spec writers should define equivalence relation btw datasets and clean up.
15:58:12 <tbaker> Sandro: Idea is to try to not force everyone to implement full datasets with datasets, but also quad stores (without default graphs) - though they can easily be added.
15:58:53 <tbaker> Unknown person says "Algorithm places triples into default graph, but language in terms of quads where fourth position is empty if you are not in the context of a NG."
15:58:57 <davidwood> ack sandro
15:59:11 <gavinc> s/XXX/gkellogg
15:59:19 <tbaker> ... Round trip triples out to an unnamed context is JSON LD.
15:59:38 <tbaker> Sandro: Faithfully?  Then I see no problem.
15:59:45 <gavinc> Yes! Lets do nothing
15:59:45 <tbaker> David: Agree to do nothing and move on?
15:59:51 <sandro> sandro: Can you round trip Trig and JSON LD with default + NGs?
15:59:58 <sandro> manu: Yes
16:00:13 <tbaker> XXX: Ivan should have a chance to comment.
16:00:16 <gavinc> move on without prejudice
16:00:17 <tbaker> David: Not closing door.
16:00:22 <sandro> s/XXX/MacTed/
16:00:52 <tbaker> PROPOSED: The WG suggests it's a good practice to put metadata about a TriG document in the document's default graph.
16:00:52 <Zakim> - +1.408.992.aaee
16:00:55 <davidwood> PROPOSED: The WG suggests it's a good practice to put metadata about a TriG document in the document's default graph.
16:01:11 <gavinc> -1 :D
16:01:22 <tbaker> Sandro: Don't care. Gave up with formal semantics for Trig.  That's okay.
16:01:48 <sandro> sandro: If you don't KNOW the default graph is metadata, then it doesn't help you.
16:01:53 <Zakim> - +1.603.897.aadd
16:02:05 <tbaker> David: Conceivably, syntactical way of X whether it contains contextual metadata.
16:02:27 <PatH> Sandro: since datasets have no semantics, no real metadata anyway, so doenst matter.
16:02:30 <tbaker> ... You are minus one re: best practice?
16:02:34 <tbaker> Sandro: Yes.
16:02:51 <AndyS> "suggest good practice" +1 (all about if anyone wants to write the text and see it through the process -- make not critical path)
16:02:51 <tbaker> David: We can say something/nothing about default graph.
16:03:29 <tbaker> XXX: Gets very strange when you merge datasets.  Reality: if you want to merge, doesn't help at all.
16:03:37 <tbaker> s/XXX/Gavin/
16:04:01 <PatH> Not clear to me where we can put a best practice recommendation in the specs. (?)
16:04:24 <PatH> q+
16:04:26 <tbaker> Sandro: Disagree. [Describes options for renaming NGs}.  Maybe cleanest to have way of flagging a particular graph as asserted (i.e., with metadata).
16:04:41 <tbaker> Unknown perhaps XXX says "Does it have to be flagged in syntax?"
16:04:51 <davidwood> ack PatH
16:05:25 <manu> zakim, code?
16:05:25 <Zakim> the conference code is 73394 (tel:+1.617.761.6200 sip:zakim@voip.w3.org), manu
16:05:38 <tbaker> Pat: We should not make recommendations about metadata.  Implies that X can be used to refer to the graph, which it cannot.  Our decision re: datasets has rendered all talk of metadata moot.
16:05:42 <sandro> +1 Pat
16:05:47 <Zakim> +??P89
16:05:54 <manu> zakim, I am ??P89
16:05:54 <Zakim> +manu; got it
16:06:11 <tbaker> David: You're saying that implementations that do this are outside the law...
16:06:54 <tbaker> Sandro: Breaking constraints vs providing addl constraints.  You would have to be hold out of band that metadata is asserted.
16:07:37 <tbaker> Scribe lost thread.
16:07:45 <gavinc> editor lost thread
16:07:50 <tbaker> Pat: We proposed syntaxes, but...
16:08:22 <sandro> I don't think so, manu
16:08:27 <tbaker> David: We'll take this off agenda.  Not clear whether this leaves Trig syntax in a good place.
16:08:29 <Guus> we have no normative support for it; Primer seems like obvious place to include metadata examples and explain what little/no support we provide
16:08:32 <tbaker> AOB?
16:08:42 <manu> Keep in mind that JSON-LD normalization will specify a quad syntax.
16:08:48 <tbaker> David: meeting adjourned.
16:08:48 <Zakim> -gkellogg
16:08:52 <Zakim> -AlexHall
16:08:52 <PatH> pat thanks tbasker
16:08:54 <Zakim> -manu
16:09:08 <AndyS> ADJOURNED
16:09:15 <PatH> s/tbasker/tbaker
16:09:17 <sandro> Zakim, who is on the call?
16:09:17 <Zakim> On the phone I see tbaker?, pchampin, GavinC, Arnaud (muted), davidwood, Guus, MacTed, Sandro, PatH, [GVoice], AndyS
16:09:22 <AlexHall> AlexHall has left #rdf-wg
16:09:28 <sandro> Zakim, who is on the call?
16:09:28 <Zakim> On the phone I see tbaker?, pchampin, GavinC, Arnaud (muted), davidwood, Guus, MacTed, Sandro, PatH, [GVoice], AndyS
16:09:40 <manu> We had hoped to make it aligned with a TRiG (or Quad) REC of some kind... but it looks like we'll have to define our own (or take TRiG/Quads to REC)
16:10:06 <davidwood> http://www.w3.org/2011/rdf-wg/wiki/Scribes
16:10:19 <gavinc> http://dvcs.w3.org/hg/rdf/raw-file/default/trig/index.html#sec-graph-statements
16:10:24 <sandro> manu, huh?
16:10:28 <AndyS> http://www.w3.org/2009/CommonScribe/manual.html
16:11:00 <pchampin> providence ???
16:12:07 <manu> sandro: For the RDF Graph Normalization algorithm (that is used in the Web Payments work... which takes a JSON-LD document, normalizes the graph, serializes it, and then digitally signs it), we serialize to a set of quads.
16:12:19 <manu> (since JSON-LD has the concept of a named graph, we have to have quads).
16:12:41 <manu> I was just raising the point that I was hoping that the RDF WG would formalize a Quad format as a REC, so we could refer to it.
16:12:50 <AndyS> For the minutes: this is useful: http://lists.w3.org/Archives/Public/public-rdf-dawg/2009AprJun/0406.html
16:12:57 <manu> but it looks like we'll have to do that work when we formalize the RDF Graph Normalization algorithm now.
16:13:23 <gavinc> ... where is RDF Graph Normalization being formalized?
16:13:44 <AndyS> manu - I would like to see N-Quads (but guess you could use a restrict (and ugly?) TriG)
16:13:44 <gkellogg> Sadly, N-Quads doesn't seem to be on the table, and can't serialize an empty-graph anyway, which seems to be part of TriG now.
16:14:02 <gavinc> manu... what obvious reasons?
16:14:05 <manu> time
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000335