IRC log of rdf-wg on 2013-06-12

Timestamps are in UTC.

14:51:58 [RRSAgent]
RRSAgent has joined #rdf-wg
14:51:58 [RRSAgent]
logging to http://www.w3.org/2013/06/12-rdf-wg-irc
14:52:07 [davidwood]
Zakim, this will be rdf
14:52:07 [Zakim]
ok, davidwood; I see SW_RDFWG()11:00AM scheduled to start in 8 minutes
14:52:18 [davidwood]
chair: David Wood
14:53:57 [Guus]
Guus has joined #rdf-wg
14:54:38 [Guus]
trackbot, start meeting
14:54:40 [trackbot]
RRSAgent, make logs world
14:54:42 [trackbot]
Zakim, this will be 73394
14:54:42 [Zakim]
ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 6 minutes
14:54:43 [trackbot]
Meeting: RDF Working Group Teleconference
14:54:43 [trackbot]
Date: 12 June 2013
14:55:11 [Guus]
chair: davidwood
14:55:23 [pchampin]
pchampin has joined #rdf-wg
14:56:13 [Zakim]
SW_RDFWG()11:00AM has now started
14:57:06 [Guus]
zakim, mute me
14:57:06 [Zakim]
sorry, Guus, I do not know which phone connection belongs to you
14:57:17 [Guus]
zakim, who is here?
14:57:17 [Zakim]
On the phone I see no one
14:57:18 [Zakim]
On IRC I see pchampin, Guus, RRSAgent, Zakim, gavinc, ivan, AndyS, TallTed, SteveH, manu1, Arnaud, davidwood, manu, sandro, trackbot, yvesr, ericP, mischat
14:58:54 [yvesr]
Zakim, who is on the phone?
14:58:54 [Zakim]
On the phone I see no one
14:58:56 [gkellogg]
gkellogg has joined #rdf-wg
14:59:28 [gavinc]
Zakim, this is rdfwg
14:59:28 [Zakim]
gavinc, this was already SW_RDFWG()11:00AM
14:59:29 [Zakim]
ok, gavinc; that matches SW_RDFWG()11:00AM
14:59:54 [gavinc]
Zakim, who is here?
14:59:54 [Zakim]
On the phone I see no one
14:59:55 [Zakim]
On IRC I see gkellogg, pchampin, Guus, RRSAgent, Zakim, gavinc, ivan, AndyS, TallTed, SteveH, manu1, Arnaud, davidwood, manu, sandro, trackbot, yvesr, ericP, mischat
15:00:46 [ScottB]
ScottB has joined #rdf-wg
15:00:49 [ivan]
zakim, dial ivan-voip
15:00:49 [Zakim]
ok, ivan; the call is being made
15:00:56 [cgreer]
cgreer has joined #rdf-wg
15:01:08 [ivan]
hi
15:01:08 [Guus]
hi
15:01:19 [pfps]
pfps has joined #rdf-wg
15:01:20 [ivan]
I am (barely) here
15:02:01 [pfps]
zakim, who is on the phone?
15:02:01 [Zakim]
On the phone I see no one
15:02:21 [AndyS]
zakim, IPcaller is me
15:02:21 [Zakim]
sorry, AndyS, I do not recognize a party named 'IPcaller'
15:03:03 [Arnaud]
Arnaud has changed the topic to: RDF WG -- current agenda: http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2013.06.12
15:04:16 [markus]
markus has joined #rdf-wg
15:04:35 [Souri]
Souri has joined #rdf-wg
15:05:27 [davidwood]
Zakim, pick a scribe
15:05:27 [Zakim]
I don't see anyone present, davidwood
15:07:54 [AndyS]
scribenick: AndyS
15:08:00 [AndyS]
scribe: Andy Seaborne
15:08:08 [davidwood]
PROPOSED to accept the minutes of the 5 June telecon: https://www.w3.org/2013/meeting/rdf-wg/2013-06-05
15:08:12 [AndyS]
Agenda: http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2013.06.12
15:08:22 [pfps]
minutes look fine
15:08:25 [ivan]
zakim, mute me
15:08:26 [Zakim]
sorry, ivan, I do not know which phone connection belongs to you
15:08:35 [gavinc]
topics are not nested correctly, however I don't -really- care
15:08:35 [davidwood]
RESOLVED to accept the minutes of the 5 June telecon: https://www.w3.org/2013/meeting/rdf-wg/2013-06-05
15:08:46 [davidwood]
http://www.w3.org/2011/rdf-wg/track/actions/pendingreview
15:09:06 [pfps]
actually I added that reference
15:09:31 [sandro]
action-245?
15:09:31 [trackbot]
ACTION-245 -- Eric Prud'hommeaux to (with Sandro) to copy or proxy Turtletests2013 to http://www.w3.org/2013/Turtletests/...,">http://www.w3.org/2013/Turtletests/..., updating all base or ttl references to http://example/base/ to be http://www.w3.org/2013/Turtletests/ -- due 2013-04-03 -- PENDINGREVIEW
15:09:32 [gavinc]
http://www.w3.org/2013/TurtleTests/
15:09:32 [trackbot]
http://www.w3.org/2011/rdf-wg/track/actions/245
15:09:43 [AndyS]
action-265?
15:09:43 [trackbot]
ACTION-265 -- David Wood to implement the langString resolution in rdf-concepts AND ENJOY IT -- due 2013-05-29 -- PENDINGREVIEW
15:09:43 [trackbot]
http://www.w3.org/2011/rdf-wg/track/actions/265
15:09:55 [davidwood]
ACTION-219?
15:09:55 [trackbot]
ACTION-219 -- Patrick Hayes to informatively reference "XML Schema Datatypes in RDF and OWL" -- due 2012-12-26 -- CLOSED
15:09:55 [trackbot]
http://www.w3.org/2011/rdf-wg/track/actions/219
15:10:09 [davidwood]
ACTION-267?
15:10:09 [trackbot]
ACTION-267 -- Eric Prud'hommeaux to update http://www.w3.org/2011/rdf-wg/wiki/Turtle_Test_Suite http://www.w3.org/2011/rdf-wg/wiki/Turtle-CR-Request#Implementation_Information http://www.w3.org/2011/rdf-wg/wiki/Main_Page to point to http://www.w3.org/2013/TurtleTests/ -- due 2013-06-05 -- PENDINGREVIEW
15:10:10 [trackbot]
http://www.w3.org/2011/rdf-wg/track/actions/267
15:10:35 [pfps]
fine by me
15:12:01 [gavinc]
http://www.w3.org/2013/TurtleTests/README
15:12:18 [AndyS]
(I see wiki pt to hg (after refresh))
15:12:30 [gavinc]
Isn't there apache magic to include the readme in the directory page?
15:13:07 [gavinc]
Yeah, it's being served as text/plain ASCII and is UTF-8 ;)
15:13:15 [davidwood]
http://www.w3.org/2013/TurtleTests/README has some unresolvable characters
15:13:16 [zwu2]
zwu2 has joined #rdf-wg
15:15:00 [AndyS]
davidwood: open action items
15:15:19 [AndyS]
guus: drop poll action ... overtaken by events
15:15:25 [gavinc]
Sandro did mine
15:16:19 [davidwood]
ACTION-203?
15:16:20 [trackbot]
ACTION-203 -- Eric Prud'hommeaux to identify an editor for a NOTE: Practical Use Cases of RDF Datasets -- due 2012-11-06 -- OPEN
15:16:20 [trackbot]
http://www.w3.org/2011/rdf-wg/track/actions/203
15:16:25 [davidwood]
^^ Sandro
15:17:01 [davidwood]
Topic: LC Drafts of Concepts and Semantics
15:17:19 [davidwood]
PROPOSED: We'll add some text to rdf-concepts saying systems MAY support blank node graph names and that communication between systems that do and don't can be aided by using Skolemization (and this closes ISSUE-131).
15:17:21 [AndyS]
davidwood: sandro proposal on issue 131:
15:17:22 [PatH]
PatH has joined #rdf-wg
15:17:57 [AndyS]
davidwood: invitation to discuss
15:18:27 [AndyS]
sandro: have not let cygri know
15:18:46 [PatH]
ignore this test line
15:18:53 [AndyS]
... SteveH may also have an option
15:19:14 [zwu2]
MAY sounds acceptable
15:19:14 [AndyS]
davidwood: systems != syntax formats
15:19:37 [AndyS]
ivan: this postpones discussion on formats
15:19:37 [pfps]
q+
15:19:56 [AndyS]
sandro: JSON-LD - no change.
15:20:09 [davidwood]
ack pfps
15:20:10 [zwu2]
q+
15:20:38 [AndyS]
pfps: are we backing away from blank nodes for graph names?
15:21:15 [AndyS]
sandro: we are backing way from the current doc defn of RDF dataset.
15:21:52 [AndyS]
pfps: if we can not requiring bnode for graphs, concepts should not mention it.
15:22:15 [AndyS]
... if defn is IRIs, concepts, semantics should stick just to that.
15:23:06 [AndyS]
... middle ground of "MAY" is distasteful.
15:23:06 [zwu2]
q-
15:23:33 [AndyS]
... because it is inappropriate for concepts to go beyond defn.
15:23:35 [gavinc]
Violate the spec! It's a classic way of getting what you want ;)
15:23:43 [SteveH]
davidwood, I don't think I have anything to add - seems unnecessary, but also mostly harmless
15:23:56 [AndyS]
... prefer particular syntaxes to go "beyond RDF"
15:23:59 [davidwood]
SteveH, thanks.
15:24:22 [PatH]
Can some one repost the link to the text please?
15:24:44 [AndyS]
sandro: systesm have done this for a long time (script -- which??)
15:24:47 [gavinc]
PROPOSED: We'll add some text to rdf-concepts saying systems MAY support blank node graph names and that communication between systems that do and don't can be aided by using Skolemization (and this closes ISSUE-131).
15:24:49 [PatH]
noisy keyboard
15:25:01 [davidwood]
PatH, http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2013.06.12#LC_Drafts_of_Concepts_and_Semantics
15:25:31 [AndyS]
pfps: acceptable to me is bNodes for graph names and then say systems may not implement that feature.
15:25:33 [sandro]
peter: Graph names can be blank nodes, and we put in concepts that surface syntaxes and systems MAY disallow blank node graph names (and use skolemization)
15:26:06 [AndyS]
zwu2: Would Oracle's current be compliant or not?
15:26:19 [AndyS]
pfps: yes
15:26:24 [Guus]
yes
15:27:14 [AndyS]
pfps: already have RDF/XML can't represent all RDF graphs.
15:27:33 [AndyS]
davidwood: is skolemization an acceptable approach?
15:27:37 [sandro]
david: A system that uses skolemization is still conformant
15:28:04 [AndyS]
zwu2: so use generated URIs for graph names.
15:28:31 [gavinc]
Eh? Web Browsers can create URLs.
15:28:37 [AndyS]
q+ to ask about skolemization assumption on existing systems
15:28:53 [pfps]
clients *can* skolemize all they want, they just can't make the skolems dereference
15:29:05 [AndyS]
sandro: theoretical impossible to generate ?? on browsers.
15:29:11 [gavinc]
pfps, yes, that
15:29:20 [davidwood]
q?
15:29:33 [PatH]
q+ to say that skolemize is defined for graphs, not datasets.
15:29:44 [AndyS]
andys: (raises eyebrow at that claim)
15:30:20 [pfps]
it's theoretically impossible to generate deferenceable IRIs if you don't have control of some dereferencing process
15:30:34 [gavinc]
New syntax example, JSON-LD
15:30:52 [AndyS]
sandro: current syntaxes don't allow bnodes for graph names - want to change TriG to allow that - JSOn-LD already does.
15:30:53 [davidwood]
ack AndyS
15:30:53 [Zakim]
AndyS, you wanted to ask about skolemization assumption on existing systems
15:31:17 [sandro]
sandro: eg Oracle will have to Skolemize somewhere between JSON-LD parsing and the Database
15:32:23 [sandro]
andy: or through, EG, the Jena API
15:33:04 [AndyS]
zwu2; why does a web client not able to generate an IRI? eg UUID.
15:33:38 [AndyS]
sandro: web browsers do not provide a robust random number
15:34:13 [AndyS]
... current problem with browser - may change in the future
15:34:25 [Arnaud1]
Arnaud1 has joined #rdf-wg
15:34:33 [davidwood]
ack PatH
15:34:33 [Zakim]
PatH, you wanted to say that skolemize is defined for graphs, not datasets.
15:35:43 [AndyS]
PatH: strictly skolemization of datasets is different and needs defining.
15:35:58 [AndyS]
sandro: UUIDs are not good for RDF.
15:36:25 [AndyS]
sandro: good looking IRI (scribe --> tag: ?)
15:36:46 [AndyS]
sandro: pfps proposal is OK and I prefer.
15:37:06 [pfps]
graph names can be IRIs, surface syntaxes and systems *MAY* not handle them
15:38:13 [pfps]
graph names can be IRIs, surface syntaxes may require IRIs, systems may skolemize them
15:38:14 [PatH]
Big echo suddsuddensuddenly
15:38:43 [davidwood]
q?
15:38:50 [pfps]
oops, wrong way around
15:39:05 [gkellogg]
Question is, is it appropriate to use a Skolum ID where a BNode is otherwise inappropriate?
15:39:30 [pfps]
graph names can be blank nodes, surface syntaxes may require IRIs for graph names, systems may skolemize blank nodes on input or output
15:40:10 [sandro]
PROPOSED: RDF Datasets include blank nodes for graph names; include a note that systems MAY use Skolemization to provide this functionality -- they don't need to STORE blank nodes in this role, just consume them, and they MAY produce them
15:41:28 [markus]
q+
15:41:36 [gkellogg]
There are no existing serializations which can represent BNodes in these positions
15:41:48 [gkellogg]
(Other than JSON-LD)
15:41:53 [markus]
q-
15:42:45 [markus]
was about to say the same that gregg just said. how can existing systems be "conformant" to something that didn't exist before?
15:42:48 [pfps]
the original proposal had dataset not having blank nodes as graph names, my counter is to have datasets allow blank nodes as graph names
15:43:07 [PatH]
Becasue the new stuff is optional?
15:43:33 [davidwood]
q?
15:43:36 [pfps]
this has the benefit that JSON-LD *is* RDF, which means that they JSON-LD documents *must* define in terms of RDF!
15:44:24 [PatH]
No, json also allows bnodes in property position.
15:44:40 [sandro]
Zero or more named graphs. Each named graph is a pair consisting of an IRI (the graph name), and an RDF graph. Graph names are unique within an RDF dataset.
15:44:41 [PatH]
But we *could* go there...
15:45:16 [sandro]
Change to: Zero or more named graphs. Each named graph is a pair consisting of a the graph name (and IRI or a blank node) , and an RDF graph. Graph names are unique within an RDF dataset.
15:45:24 [pfps]
i'm getting less and less enamored of JSON-LD
15:46:41 [sandro]
andy: Allow systems that dont do this to be conformant by invoking the general rule that you can reject anything
15:49:02 [AndyS]
andys: state that the definition has changed and note systems may provide the old form.
15:49:11 [sandro]
PROPOSED: We change the definition of RDF Datasets to allow blank node graph names. We note that earlier definitions of datasets did not include blank node graph names, and Skolemizaiton may be useful in providing compatiblity
15:49:22 [PatH]
We need to say that two dofferent datasets cannot share a blank node. Bnodes are unique to the dataset.
15:49:49 [pfps]
at some point there will have to be some WG discussion of Antoine's comments on Semantics
15:50:10 [sandro]
sandro: in a pure mathematical sense, I don't tink that's true Pat.
15:50:39 [PatH]
OK, withdraw that.
15:50:39 [sandro]
PROPOSED: We change the definition of RDF Datasets to allow blank node graph names. We note (in rdf-concepts) that earlier definitions of datasets did not include blank node graph names, and Skolemizaiton may be useful in providing compatiblity. This closes issue-131
15:50:44 [AndyS]
issue-131?
15:50:44 [trackbot]
ISSUE-131 -- How can one create an RDF dataset without being a web server? -- open
15:50:44 [trackbot]
http://www.w3.org/2011/rdf-wg/track/issues/131
15:50:50 [sandro]
+1
15:50:54 [markus]
+1
15:50:54 [gkellogg]
+1
15:50:57 [davidwood]
+1
15:50:59 [pfps]
+1, but the previous objectors may still object
15:51:00 [zwu2]
0
15:51:01 [AndyS]
0
15:51:02 [cgreer]
+1
15:51:10 [gavinc]
0
15:51:16 [Souri]
0
15:51:20 [Arnaud1]
+1
15:51:28 [PatH]
I just now read Antoine's email, will respond to it later today. It needs a detailed response.
15:51:35 [PatH]
+1
15:51:40 [yvesr]
0
15:52:00 [ivan]
+1
15:52:08 [gavinc]
In 6 months there won't be a working group :P
15:52:12 [PatH]
consensus by attenuation.
15:52:15 [yvesr]
actually, revising to +0.5
15:52:19 [pfps]
the only changes to Semantics are localized in the section on RDF Datasets
15:52:38 [sandro]
eric: +1
15:52:43 [PatH]
NOt clear we need any changes to Sematnics.
15:52:52 [sandro]
RESOLVED: We change the definition of RDF Datasets to allow blank node graph names. We note (in rdf-concepts) that earlier definitions of datasets did not include blank node graph names, and Skolemizaiton may be useful in providing compatiblity. This closes issue-131
15:53:08 [davidwood]
On ISSUE-136 Sandro suggests: PROPOSED: The formal meaning of an RDF Dataset is no less than the formal meaning of its default graph. This revises an earlier decision that datasets in general have no formal semantics, in order to allow for the use of specific dataset semantics to be signalled within a dataset. This feature to be added to rdf-concepts and rdf-mt, marked AT RISK for LC, since it hasn't been discussed much yet.
15:53:16 [sandro]
topic: issue-136
15:53:29 [AndyS]
issue-136?
15:53:29 [trackbot]
ISSUE-136 -- How can one indicate which semantics are intended for a Dataset? -- raised
15:53:29 [trackbot]
http://www.w3.org/2011/rdf-wg/track/issues/136
15:54:56 [AndyS]
This was only raised less than 24hrs ago.
15:55:17 [AndyS]
sandro: default graph is the truth condition of the dataset
15:56:50 [AndyS]
ivan: don't understand the terminology
15:57:15 [AndyS]
andys: anyone can object at any time - W3C process does not place a limit.
15:57:18 [pfps]
Option 2 amounts to saying that the formal interpretation of a dataset is the interpretation of its default graph
15:57:36 [sandro]
PROPOSED: Close ISSUE-137 adopting Option 2; the truth condition of datasets are the truth conditions of its default graph
15:57:42 [davidwood]
q?
15:57:55 [markus]
gotta run.. I'm late already
15:57:59 [PatH]
Yes, exactly. So datasets actually do have a semantics now, and we say what it is.
15:58:09 [markus]
I'm +1 to the proposal that has just been posted by Sandro
15:58:12 [gavinc]
The truth condition of a dataset who's default graph is the union/merge/whatever of it's named graph is now amazing ;)
15:58:26 [davidwood]
Thanks, markus
15:58:30 [pfps]
this is the cleanest way of proceeding - Option 3 requires looking inside the graph and pulling out parts, which we have never done
15:59:31 [AndyS]
q+
15:59:44 [sandro]
pfps: so it's a semantic-preserving operation, when you do the json-ld thing of using th defaule graph
16:00:00 [ivan]
+1
16:00:06 [davidwood]
ack AndyS
16:01:36 [sandro]
andy: what about simple entailment turning IRIs into blank nodes in a dataset?
16:01:36 [AndyS]
ack me
16:01:37 [sandro]
pat: I think it's harmless here.
16:01:52 [sandro]
AndyS: I'm not convinced. There may need to be stronger conditions.
16:02:05 [sandro]
PROPOSED: Close ISSUE-137 adopting Option 2; the formal interpretation of a datasets is the formal interpretation of its default graph
16:02:24 [pfps]
+1
16:02:25 [sandro]
PROPOSED: Close ISSUE-137 adopting Option 2; the formal interpretation (truth conditions) of a datasets is the formal interpretation (truth conditions) of its default graph
16:02:28 [ivan]
+1
16:02:30 [gavinc]
0
16:02:31 [gkellogg]
+1
16:02:36 [pfps]
Ivan: this is editorial
16:02:38 [yvesr]
+1
16:02:40 [zwu2]
+1
16:02:41 [sandro]
+1
16:02:44 [sandro]
issue-137?
16:02:44 [trackbot]
ISSUE-137 does not exist.
16:02:51 [PatH]
+1
16:02:52 [sandro]
we mean 136
16:03:00 [davidwood]
0
16:03:26 [Arnaud1]
+1
16:03:27 [sandro]
RESOLVED: Close ISSUE-136 adopting Option 2; the formal interpretation (truth conditions) of a datasets is the formal interpretation (truth conditions) of its default graph
16:03:27 [Guus]
+0
16:03:30 [pchampin]
+1
16:03:35 [sandro]
eric: +1
16:03:40 [Souri]
0
16:04:06 [sandro]
scribe: sandro
16:04:10 [sandro]
topic: Publications
16:04:32 [sandro]
davidwood: We can't proceed on rdf-mt because of Antoine's comments, and rdf-concepts needs editorial changes I haven't done yet
16:04:42 [sandro]
davidwood: I propose we advance them next week
16:04:54 [sandro]
pfps: Antoine's changes are too big for that.
16:05:23 [sandro]
pfps: Then we need to say over-ride Antoine's
16:06:02 [sandro]
sandro: I'm waiting to see a rebuttal to Antoine's persepctive
16:06:59 [sandro]
ACTION: pfps to send email describing the differences between Antoine's view and the current draft (eg you don't know what the datatype interpretation is)
16:06:59 [trackbot]
Created ACTION-272 - Send email describing the differences between Antoine's view and the current draft (eg you don't know what the datatype interpretation is) [on Peter Patel-Schneider - due 2013-06-19].
16:07:32 [sandro]
pfps: I sympathize with his position, where you want everything to refer, or you don't know what things mean.
16:07:50 [sandro]
davidwood: Isn't this the big philosophical leap of RDF in general?
16:08:20 [sandro]
pfps: Not quite. THere's a difference between not exactly knowing, and not knowing what something refers to
16:08:54 [sandro]
pfps: eg we know who brad pitt is, without really *knowing* brad pitt
16:09:03 [sandro]
pat: I'll also reply to Antoine
16:09:21 [sandro]
davidwood: cf the Polymorphic Nature of Null
16:09:28 [sandro]
pfps: Null is even worse.
16:09:41 [sandro]
pfps: You could be talking about THREE different things with Null.
16:10:14 [sandro]
topic: ISSUE-23
16:10:17 [davidwood]
On ISSUE-23, Sandro suggests: PROPOSED: Close ISSUE-23 without general comment. We have been deciding on a case-by-case basis (YES for Turtle/TriG, NO for JSON-LD) and we expect to continue to do so.
16:10:32 [pfps]
+2
16:10:32 [davidwood]
ISSUE-23?
16:10:32 [trackbot]
ISSUE-23 -- Does going from single-graph to multi-graph require new format and new media types? -- open
16:10:32 [trackbot]
http://www.w3.org/2011/rdf-wg/track/issues/23
16:10:36 [sandro]
PROPOSED: Close ISSUE-23 without general comment. We have been deciding on a case-by-case basis (YES for Turtle/TriG, NO for JSON-LD) and we expect to continue to do so.
16:10:45 [pfps]
+2
16:10:58 [sandro]
+1
16:10:59 [davidwood]
+1
16:10:59 [cgreer]
+1
16:11:01 [yvesr]
+1
16:11:01 [gkellogg]
+1
16:11:03 [gavinc]
+∞
16:11:04 [pchampin]
+1
16:11:05 [ivan]
+1
16:11:09 [PatH]
0
16:11:11 [zwu2]
+1
16:11:16 [Souri]
+1
16:11:31 [Arnaud]
+1
16:11:40 [sandro]
RESOLVED: Close ISSUE-23 without general comment. We have been deciding on a case-by-case basis (YES for Turtle/TriG, NO for JSON-LD) and we expect to continue to do so.
16:12:38 [sandro]
topic: Rec Track Status of TriG, N-Quads, N-Triples
16:13:30 [sandro]
sandro: WG decide rec-track, then leadership figured we didn't have time
16:13:50 [sandro]
davidwood: Sorry I failed to close the loop on these
16:14:01 [sandro]
gavin: We've had some comments on this.
16:14:26 [sandro]
davidwood: We should probably make this a WG decision
16:14:55 [sandro]
s/TriG, /
16:14:57 [sandro]
s/TriG, //
16:15:06 [Guus]
q+
16:15:11 [sandro]
we're just talking about N-Quads and N-Triples
16:16:00 [sandro]
gavin: So N-Triples was in a Recommendation, but it wasn't itself Recommended.
16:16:04 [davidwood]
http://www.w3.org/2011/rdf-wg/wiki/Rdf-extension.html says that TriG will be a Rec, but NT and NQ will be notes. That's what we want to do.
16:16:16 [sandro]
guus: Now it's upgraded, even though it's not in a Recommendation any more.
16:16:41 [sandro]
guus: It used to be an Appendix that said don't use this.
16:16:44 [sandro]
gavin: We've had comments that it's weird.
16:17:36 [sandro]
gavin: ... from when N-Triples was part of Turtle Doc. Greg Williams and David Robillard say 'it's weird'
16:18:04 [sandro]
davidwood: We're very close to out of time
16:18:27 [sandro]
guus: We COULD bring this back to the rec track. The question is whether we should take on this work now.
16:19:26 [gkellogg]
We have a test suite for N-Triples, we don't for TriG!
16:19:32 [sandro]
sandro: My take is this is mostly up to Gavin, if he's really up to doing all the work.
16:20:04 [Arnaud]
while I understand Gavin's frustration I think it is unwise to underestimate what it would take to add any spec to the Rec track at this point
16:20:10 [sandro]
gavin: TriG is late because we told it to wait for Datasets; that's not my fault.
16:20:14 [gkellogg]
Pointer to test suite for TriG?
16:20:38 [sandro]
gavin: It's hard to take these things to Rec Track. but N-Triples is very widely used. Not having it be a Rec feels very strange.
16:20:55 [sandro]
gavin: No test suite for TriG yet.
16:21:26 [sandro]
davidwood: Gavin, how would you like to proceed.
16:21:52 [sandro]
gavin: I'd like to take the N-* syntaxes to Rec. It will be easier than TriG.
16:22:21 [sandro]
gavin: There's a very real chance that TriG wont make it, but N-Quads will.
16:22:37 [sandro]
davidwood: Last Call by end of june, Gavin?
16:22:56 [sandro]
Guus: TriG is our first priority
16:23:41 [sandro]
gavin: I see no way that TriG will make it out of LC and to PR in six months. That's a fantasy at this point.
16:24:05 [sandro]
gavin: No TriG implementations work the same as they used to; no TriG implementations work today -- we can't get this done in six months
16:24:36 [gkellogg]
I have one I think is reasonably complete.
16:24:56 [sandro]
sandro: We *could* do two implemantions in a week
16:25:37 [sandro]
ivan: Let's go on optimistically. We can drop TriG to note later.
16:25:46 [sandro]
sandro: Right -- so don't drop N-Quad yet.
16:26:03 [sandro]
gavin: Yes, -- which ever goes to PR first, or both (or JSON-LD) wins.
16:26:15 [sandro]
davidwood: We're over time.
16:26:18 [zwu2]
bye
16:26:31 [sandro]
davidwood: We'll try to get rdf-concepts and rdf-mt to last call by next week.
16:26:35 [Arnaud]
regrets for next week (LDP F2F)
16:26:55 [sandro]
ADJOURNED
16:27:11 [ScottB]
ScottB has left #rdf-wg
17:42:32 [Zakim]
SW_RDFWG()11:00AM has ended
17:42:33 [Zakim]
Attendees were
18:35:22 [Zakim]
Zakim has left #rdf-wg
19:28:22 [swh]
swh has joined #rdf-wg
19:32:10 [swh_]
swh_ has joined #rdf-wg
19:44:05 [swh_]
swh_ has joined #rdf-wg
19:45:56 [swh_]
swh_ has joined #rdf-wg
21:18:16 [Arnaud]
Arnaud has joined #rdf-wg
22:24:26 [yvesr]
yvesr has joined #rdf-wg
22:40:56 [yvesr]
yvesr has joined #rdf-wg
22:49:32 [Arnaud1]
Arnaud1 has joined #rdf-wg
23:16:27 [Arnaud]
Arnaud has joined #rdf-wg