IRC log of rdf-wg on 2013-05-15

Timestamps are in UTC.

14:59:46 [RRSAgent]
RRSAgent has joined #rdf-wg
14:59:46 [RRSAgent]
logging to
14:59:48 [trackbot]
RRSAgent, make logs world
14:59:48 [Zakim]
Zakim has joined #rdf-wg
14:59:50 [trackbot]
Zakim, this will be 73394
14:59:50 [Zakim]
ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 1 minute
14:59:51 [trackbot]
Meeting: RDF Working Group Teleconference
14:59:51 [trackbot]
Date: 15 May 2013
15:00:05 [AndyS]
zakim, this is 73394
15:00:05 [Zakim]
ok, AndyS; that matches SW_RDFWG()11:00AM
15:00:19 [Zakim]
15:00:31 [Zakim]
15:00:31 [Zakim]
15:00:36 [yvesr]
Zakim, ??P19 is me
15:00:37 [Zakim]
+yvesr; got it
15:00:39 [AndyS]
zakim, IPCaller is me
15:00:40 [Zakim]
+AndyS; got it
15:01:06 [pfps]
zakim, who is on the phone?
15:01:06 [Zakim]
On the phone I see +1.408.992.aaaa, yvesr, Guus_Schreiber, AndyS
15:01:09 [Zakim]
15:01:14 [pfps]
zakim, aaaa is me
15:01:14 [Zakim]
+pfps; got it
15:01:17 [Zakim]
15:01:20 [TallTed]
Zakim, OpenLink_Software is temporarily me
15:01:21 [Zakim]
+TallTed; got it
15:01:23 [TallTed]
Zakim, mute me
15:01:23 [Zakim]
TallTed should now be muted
15:01:37 [TallTed]
TallTed has changed the topic to: RDF WG -- current agenda:
15:01:52 [AZ]
AZ has joined #rdf-wg
15:02:17 [Zakim]
+ +1.415.686.aabb
15:02:25 [gkellogg]
zakim, I am aabb
15:02:25 [Zakim]
+gkellogg; got it
15:02:33 [Zakim]
+ +1.540.538.aacc
15:02:34 [Zakim]
15:02:42 [AZ]
Zakim, aacc is me
15:02:42 [Zakim]
+AZ; got it
15:03:07 [davidwood]
Zakim, aacc is me
15:03:08 [Zakim]
sorry, davidwood, I do not recognize a party named 'aacc'
15:03:21 [AZ]
Zakim, ??P33 is me
15:03:21 [Zakim]
+AZ; got it
15:03:30 [davidwood]
Zakim, aacc is me
15:03:30 [Zakim]
sorry, davidwood, I do not recognize a party named 'aacc'
15:03:36 [davidwood]
Zakim, who is here?
15:03:37 [Zakim]
On the phone I see pfps, yvesr, Guus_Schreiber, AndyS, TallTed (muted), Sandro, gkellogg, AZ, AZ.a
15:03:37 [Zakim]
On IRC I see AZ, Zakim, RRSAgent, gkellogg, pfps, AndyS, Guus, davidwood, TallTed, manu1, sandro, trackbot, yvesr, manu, ericP, mischat
15:03:56 [davidwood]
Zakim, AZ.a is me
15:03:56 [Zakim]
+davidwood; got it
15:04:03 [davidwood]
Zakim, who is here?
15:04:03 [Zakim]
On the phone I see pfps, yvesr, Guus_Schreiber, AndyS, TallTed (muted), Sandro, gkellogg, AZ, davidwood
15:04:06 [Zakim]
On IRC I see AZ, Zakim, RRSAgent, gkellogg, pfps, AndyS, Guus, davidwood, TallTed, manu1, sandro, trackbot, yvesr, manu, ericP, mischat
15:04:15 [pfps]
15:04:18 [pfps]
15:04:29 [AZ]
Zakim, mute me
15:04:29 [Zakim]
AZ should now be muted
15:05:00 [TallTed]
15:05:22 [TallTed]
Zakim, unmute me
15:05:22 [Zakim]
TallTed should no longer be muted
15:05:23 [davidwood]
ack AZ, davidwood
15:05:39 [sandro]
zakim, who is muted?
15:05:40 [Zakim]
I see AZ muted
15:05:42 [davidwood]
Zakim, unmute me
15:05:42 [Zakim]
davidwood was not muted, davidwood
15:05:47 [AZ]
Zakim, unmute me
15:05:47 [Zakim]
AZ should no longer be muted
15:05:48 [Zakim]
15:05:50 [TallTed]
Zakim, AZ is davidwood
15:05:50 [Zakim]
+davidwood; got it
15:06:01 [TallTed]
Zakim, who's here?
15:06:01 [Zakim]
On the phone I see pfps, yvesr, Guus_Schreiber, AndyS, TallTed, Sandro, gkellogg, davidwood.a, davidwood, Souri
15:06:03 [Zakim]
On IRC I see AZ, Zakim, RRSAgent, gkellogg, pfps, AndyS, Guus, davidwood, TallTed, manu1, sandro, trackbot, yvesr, manu, ericP, mischat
15:06:11 [AZ]
david, sorry to have used your identity
15:06:15 [Guus]
zakim, mute me
15:06:15 [Zakim]
Guus_Schreiber should now be muted
15:06:49 [AZ]
Zakim, davidwood is me
15:06:49 [Zakim]
+AZ; got it
15:07:08 [davidwood]
Zakim, davidwood.a is me
15:07:08 [Zakim]
+davidwood; got it
15:07:11 [sandro]
zakim, who is on the call?
15:07:11 [Zakim]
On the phone I see pfps, yvesr, Guus_Schreiber (muted), AndyS, TallTed, Sandro, gkellogg, davidwood, AZ, Souri
15:08:16 [Zakim]
15:08:18 [davidwood]
Zakim, pick a victim
15:08:18 [Zakim]
Not knowing who is chairing or who scribed recently, I propose TallTed
15:08:38 [Zakim]
15:08:39 [TallTed]
Zakim, mute me
15:08:39 [Zakim]
TallTed should now be muted
15:08:41 [sandro]
scribe: sandro
15:08:49 [davidwood]
PROPOSED to accept the minutes of the 08 May telecon:
15:08:49 [davidwood]
15:08:49 [davidwood]
15:08:51 [pchampin]
pchampin has joined #rdf-wg
15:09:00 [Souri]
Souri has joined #rdf-wg
15:09:11 [davidwood]
RESOLVED to accept the minutes of the 08 May telecon:
15:09:19 [davidwood]
Review of action items
15:09:19 [davidwood]
15:09:19 [davidwood]
15:09:21 [pfps]
Minutes are fine
15:09:25 [sandro]
RESOLVED: to accept the minutes of the 08 May telecon:
15:09:31 [pchampin]
zakim, ??P37 is me
15:09:31 [Zakim]
+pchampin; got it
15:09:53 [davidwood]
close ACTION-215
15:09:53 [trackbot]
Closed ACTION-215 Provide update to RDF semantics within two weeks..
15:09:56 [pfps]
That's an old aciton that was done quite some time ago (action 215)
15:10:21 [davidwood]
close ACTION-262
15:10:21 [trackbot]
Closed ACTION-262 Add text to Concepts that bad language tag is a syntax error.
15:10:50 [davidwood]
15:10:50 [trackbot]
ACTION-261 -- Patrick Hayes to add wording to Semantics to again define the merge of two RDF graphs -- due 2013-05-08 -- OPEN
15:10:50 [trackbot]
15:11:28 [davidwood]
Topic: LC Drafts of Concepts and Semantics
15:11:29 [sandro]
davidwood: I think action-261 is done. can you confirm, pfps?
15:11:29 [sandro]
pfps: not at the moment
15:11:45 [davidwood]
Sandro's suggestion: blank nodes as named graph labels
15:11:45 [davidwood]
15:12:53 [pfps]
action 261 is done - I'm marking it pending
15:12:53 [trackbot]
Error finding '261'. You can review and register nicknames at <>.
15:13:46 [davidwood]
Sandro: The RDF WG decided that the thing in the forth column in a quad can only be a URI (not a literal or a blank node). However, the LDP WG highlighted that blank nodes could be useful to define a PATCH.
15:13:54 [sandro]
{ :insert <#i>; :delete <#d> } <#i> { .... } <#d> {.... }
15:15:35 [pfps]
does SPARQL consume datasets? If not, there is no problem at all
15:15:42 [sandro]
15:15:53 [sandro]
15:16:16 [pfps]
15:16:19 [yvesr]
that answers it, thank you :)
15:16:49 [pfps]
15:16:53 [AndyS]
Clarification -- PATCH /url does give a base
15:17:31 [sandro]
davidwood: Andy, thoughts?
15:18:13 [davidwood]
Andy: A PATCH must be defined against a base URL.
15:20:14 [sandro]
AndyS: I don't think this is a great way to do patch, as seen in my example
15:21:39 [sandro]
sandro: I'm not saying this a complete design for patch, just that we should enable it.
15:22:19 [sandro]
davidwood: this isn't about LDP -- it's about whethe we are over-constraining RDF by forcing the 4th column of a quad to be a URI
15:22:20 [yvesr]
davidwood, +1 - it's not really about PATCH, but about overconstraining RDF
15:22:44 [sandro]
davidwood:So what are your thoughts on that, Andy?
15:23:15 [sandro]
andy: I don't have the energy to go through all this again
15:23:38 [sandro]
andy: I want this WG to finish. I'd rather not re-open this issue, because it could take a lot of time.
15:23:49 [sandro]
eric: Let's just nail it today?
15:24:30 [sandro]
andy: Lack of practical experience.
15:24:48 [sandro]
andy: I'm trying not to re-open the whole debate
15:25:24 [sandro]
davidwood: Can I please have a list of all resolutions, Sandro?
15:26:28 [sandro]
davidwood: I think this would help JSON-LD. it might help LDP. I take Andy's point.
15:27:21 [ericP]
_:statement1 dc:author "Bob" . _:statement1 { :TheMoon :madeOf :greenCheese } .
15:27:21 [sandro]
greg: There's no open issue. It was about use of JSON-LD for transient messages.
15:27:31 [sandro]
greg: which is much like this.
15:27:41 [davidwood]
15:27:43 [gkellogg]
15:27:56 [sandro]
davidwood: It's an At Risk issue, but not recorded in issue tracker.
15:28:43 [sandro]
greg: this would be a great way to resolve the issue. and I think there is implementation experience.
15:28:48 [sandro]
andy: who?
15:29:02 [sandro]
greg: python,
15:29:05 [sandro]
greg: ruby
15:29:07 [davidwood]
15:29:27 [gkellogg]
Python librdf, Ruby rdf.rb allow bnodes as graph names.
15:29:31 [sandro]
AndyS: Internally, Jena doesn't care about this. It doesn't even mind bnodes as predicates. It's just in publishing data that it gets restricted.
15:29:34 [davidwood]
ack pfps
15:29:37 [sandro]
eric: same with SWObjects
15:29:51 [sandro]
PROPOSED: Datasets can use blank nodes as graph names, not just IRIs.
15:30:04 [gkellogg]
15:30:04 [sandro]
15:30:06 [ericP]
15:30:08 [pfps]
15:30:09 [TallTed]
15:30:10 [yvesr]
15:30:10 [AndyS]
15:30:16 [Souri]
15:30:18 [AZ]
15:30:30 [sandro]
davidwood: Souri? Explain?
15:30:58 [pfps]
15:31:33 [pfps]
15:31:42 [pfps]
15:32:01 [sandro]
Souri: The way way interepreted balnk nodes earlier, we used rraph names to distinguish blank nodes. I understand we're thinking of sharing blank nodes between graphs. But earlier, they were separate. That seems to be going away here, and that's going to cause issues.
15:32:11 [sandro]
souri: I could put minus .5
15:32:22 [sandro]
souri: I do not fully understand all the implications.
15:32:24 [davidwood]
ack pfps
15:32:59 [sandro]
pfps: My reading is that for months in the WG, it's been assumed that graphs can share blank nodes.
15:33:34 [sandro]
sandro: agreed. I'll make that list of resolutions later today.
15:33:34 [gkellogg]
LDP PATCH semantics would probably depend on BNodes being shared between graphs.
15:33:54 [sandro]
souri: We are okay with that. Graph names themselves being blank nodes --- that's not something we had assumed.
15:34:01 [ericP]
15:34:23 [sandro]
souri: so _:g1 { ... _:b1 ... } I can interpret it as _:g1__b1
15:34:33 [sandro]
souri: I'll have to make something unique
15:34:36 [sandro]
15:34:36 [ericP]
15:35:10 [AZ]
RDF datasets do not have a formal semantics, so graph IRI, graph bnode, graph literal or graph schmuck, it's all equally meaningless...
15:35:18 [pfps]
I'm still confused. Is the issue *just* consing up unique IDs for blank nodes within datasets?
15:35:29 [sandro]
souri: It's almost like graph is a name for a set of triple, ... and then that itself becoming a blank node, it causes some difficulty in implement. I'm fine with it -- I just wanted to let you know it's not the most desirable thing.
15:35:49 [davidwood]
ack sandro
15:35:49 [yvesr]
one of the main issue i can see is around scoping - would probably need some semantics to back it up
15:36:25 [davidwood]
ack ericP
15:36:26 [sandro]
sandro: you can skolemize
15:36:31 [sandro]
sandro: I'm hearing consensus.
15:36:43 [sandro]
eric: if souri has new information, he could bring that forward.
15:36:49 [TallTed]
there was a -1
15:37:40 [Souri]
15:37:45 [sandro]
RESOLVED: Datasets can use blank nodes as graph names, not just IRIs.
15:38:06 [pfps]
actions on concepts and semantics?
15:38:06 [davidwood]
Language tags
15:38:06 [davidwood]
15:38:11 [sandro]
davidwood: Souri (or anyone else) if you find a new problem with this, please say so.
15:38:23 [sandro]
topic: Language Tags
15:38:25 [ericP]
YES!!!! a thousand n3 use cases are now available to RDF
15:38:30 [AndyS]
q+ to ask we do this properly - change SPARQL
15:38:38 [davidwood]
ack AndyS
15:38:38 [Zakim]
AndyS, you wanted to ask we do this properly - change SPARQL
15:39:22 [gkellogg]
TriG, NQuads
15:39:30 [yvesr]
ericP, now to re-vote on literal as subjects :)
15:39:57 [sandro]
eric: Andy, do you want the action?
15:40:20 [sandro]
I don't see how it's an erratum.
15:41:29 [sandro]
sandro: oh, okay, if it's called "RDF Dataset" in SPARQL, then yes, we should bring them into line some day.
15:41:31 [pfps]
So the solution is to add a note in Concepts indicating that there is a difference.
15:41:32 [sandro]
sandro: so errata is fine.
15:42:22 [AndyS]
ACTION Andy Raise an erratum against SPARQL for the resolution that graph names can be blank nodes.
15:42:22 [trackbot]
Created ACTION-263 - Raise an erratum against SPARQL for the resolution that graph names can be blank nodes. [on Andy Seaborne - due 2013-05-22].
15:42:36 [sandro]
action: sandro find the history and suggest phrasing for Concepts
15:42:36 [trackbot]
Created ACTION-264 - Find the history and suggest phrasing for Concepts [on Sandro Hawke - due 2013-05-22].
15:42:37 [Zakim]
15:43:21 [davidwood]
Language tags as values
15:43:21 [davidwood]
15:43:48 [PatH]
PatH has joined #rdf-wg
15:44:25 [sandro]
andy: There are couple of ways to treat language tags. We could take it out of the syntax issues.
15:44:45 [sandro]
andy: that would license systems to decide how much of langstring to implement
15:45:12 [sandro]
andy: I think the same can be done for xsd:string -- it doesnt have to be required for implementation
15:45:19 [sandro]
15:46:14 [sandro]
davidwood: People seem to like this proposal
15:46:29 [sandro]
davidwood: Need a WG resolution on this
15:46:34 [davidwood]
15:46:39 [davidwood]
ack sandro
15:47:35 [sandro]
sandro: I THINK this just enables a new kind of implentaiton, not changing any existing ones
15:47:42 [pfps]
q+ to disagree with Sandro
15:47:48 [sandro]
davidwood: I don't think this changes any existing implementations
15:47:57 [pfps]
q+ ... in a certain way
15:48:31 [ericP]
-> a proposal for language tag case-insensitivity which does not lean on a value mapping
15:48:31 [davidwood]
ack pfps
15:48:31 [Zakim]
pfps, you wanted to disagree with Sandro
15:48:36 [PatH]
15:48:38 [sandro]
AndyS: I don't think it would license some behaviors that are currently not licensed, that's all.
15:49:05 [sandro]
pfps: So suppose I want to write a new impl, and I don't lile lang tag string, so I'm going to ignreo @
15:49:12 [sandro]
sandro: that wouldnt be a turtle parser
15:50:04 [AndyS]
ericP -- it's ambiguous as to one or two triples in the graph -- depends when tests are done.
15:50:11 [sandro]
s/AndyS: I don't think/AndyS: I think/
15:50:42 [ericP]
15:51:17 [sandro]
sandro: I agree, using rdf:PlainLiteral would be simpler here, but that would involve some backtracking.
15:51:17 [ericP]
q+ to ask how the plain literal solution is a solution
15:51:40 [sandro]
pat: It would get rid of the strange characters.
15:52:05 [davidwood]
ack PatH
15:52:16 [davidwood]
ack ericP
15:52:16 [Zakim]
ericP, you wanted to ask how the plain literal solution is a solution
15:52:21 [PatH]
15:52:21 [sandro]
david: We might run a few minutes late that noon ET.
15:52:47 [AndyS]
LANG("abc") = xsd:string and LANG("abc"@en) = rdf:langString.
15:52:48 [pfps]
my proposal is that Concepts doesn't have plain literals any more, instead it uses rdf:plainLiteral
15:52:52 [davidwood]
s/late that/later than/
15:53:29 [sandro]
eric: The problem we're trying to address is that historically, if you give triples with "foo"@EN, rdf said that was the same literal as "foo"@en. And BCP-47 says DONT case-fold.
15:54:38 [sandro]
pfps: rdf:PlainLiteral says the lexical space of the language tag does not have to be lower case, but the value space has it lower case.
15:54:45 [Zakim]
15:54:51 [sandro]
we lost AndyS
15:55:08 [AndyS]
zakim cut me off!
15:55:13 [sandro]
pat: the std rules of lang tags would only apply to the value space
15:55:33 [Zakim]
15:55:39 [AndyS]
zakim, IPCaller is me
15:55:39 [Zakim]
+AndyS; got it
15:56:51 [sandro]
eric: Proposal that the value space have case folded version.
15:57:01 [sandro]
eric: Proposal that these are compared case-insentively
15:57:20 [sandro]
sandro: I don't think we understand this well enough right now.
15:57:34 [sandro]
pat: look at rdf:PlainLiteral spec carefully during the week.
15:57:40 [davidwood]
15:57:56 [sandro]
AndyS: SPARQL already committed to something, because we understood there would not be rdf:LangString any more.
15:58:06 [sandro]
15:58:10 [ericP]
2013-05-15T15:52:47Z <AndyS> LANG("abc") = xsd:string and LANG("abc"@en) = rdf:langString.
15:58:15 [sandro]
s/any more//
15:58:33 [pfps]
what is rdf:langString??
15:59:03 [sandro]
AndyS: rdf:langString is in the docs right now.
15:59:09 [davidwood]
Topic: LC for JSON-LD API
15:59:13 [sandro]
Pat: It's the ugly thing I want to get rid of. :-)
15:59:19 [AZ]
Andy, you mean DATATYPE("abc"@en) = rdf:langString ?
15:59:35 [ericP]
i was wondering that
15:59:58 [sandro]
greg: Briefly, the LC2 of JSON-LD-API failed to be published yesterday, so we have a chance to put in two AT RISK notes.
16:00:12 [sandro]
... one on Native Types for Number, and other other for how to reference Futures
16:00:26 [AndyS]
Currently, rdf:langString is defined in MT (I'm suggesting it go in concepts with the other datatypes mentioned)
16:00:28 [gkellogg]
PROPOSAL: Publish the latest (
16:00:29 [gkellogg]
16:00:30 [gkellogg]
) JSON-LD Algorithms and API 1.0 specification as a 2nd Last Call
16:00:31 [gkellogg]
Working Draft on Thursday, May 16th 2013 with a Last Call period of 3 weeks.
16:00:35 [sandro]
.. we don't want to resolve these today, just put in the issue markers.
16:01:00 [sandro]
PROPOSAL: Publish the latest ( JSON-LD Algorithms and API 1.0 specification as a 2nd Last Call
16:01:01 [sandro]
16:01:06 [gkellogg]
16:01:11 [TallTed]
16:01:12 [AndyS]
16:01:13 [PatH]
16:01:16 [yvesr]
16:01:19 [pfps]
+0 (lacking knowledge)
16:01:34 [davidwood]
16:01:49 [davidwood]
FEATURE AT RISK 1 is handled by
16:01:50 [sandro]
RESOLVED: Publish the latest ( JSON-LD Algorithms and API 1.0 specification as a 2nd Last Call
16:02:07 [sandro]
davidwood: Any other business?
16:02:07 [davidwood]
Topic: AOB
16:02:15 [davidwood]
16:02:21 [pfps]
no comments on semantics
16:02:29 [Zakim]
16:02:41 [sandro]
16:02:48 [Zakim]
16:02:49 [Zakim]
16:02:49 [Zakim]
16:02:50 [Zakim]
16:02:50 [Zakim]
16:02:52 [Zakim]
16:02:53 [PatH]
16:03:11 [Zakim]
16:03:13 [Zakim]
16:03:33 [Zakim]
16:05:02 [Zakim]
16:19:27 [davidwood]
davidwood has joined #rdf-wg
16:19:32 [gavinc]
gavinc has joined #rdf-wg
16:31:52 [AndyS]
There are no action items for working through the consequences of the resolution on blank nodes for named graphs.
16:37:59 [Zakim]
16:42:59 [Zakim]
disconnecting the lone participant, yvesr, in SW_RDFWG()11:00AM
16:43:00 [Zakim]
SW_RDFWG()11:00AM has ended
16:43:00 [Zakim]
Attendees were +1.408.992.aaaa, Guus_Schreiber, yvesr, AndyS, pfps, Sandro, TallTed, +1.415.686.aabb, gkellogg, +1.540.538.aacc, Souri, AZ, davidwood, ericP, pchampin, PatH
17:17:28 [davidwood]
davidwood has joined #rdf-wg
17:36:37 [davidwood]
davidwood has joined #rdf-wg
17:40:49 [gavinc]
While we're at it why not just replace all the nodes with blank nodes?
17:41:18 [gavinc]
Everything is a variable! It's very unconstrained
17:41:20 [gavinc]
Everything is a variable! It's very unconstrained
17:55:01 [AndyS]
Hmm - this will be the only place where [] is allowed and is a unique parse term - everywhere else it's a possible blankNodePropertyList
17:55:47 [AndyS]
(well, I find it interesting)
18:03:33 [gavinc]
_:b {_:b rdf:value _:b } is very strange.
18:08:37 [AndyS]
Currently, I'm merely updating RIOT parser without putting extra code on the critical path for NT and NQ.
18:16:12 [davidwood]
davidwood has joined #rdf-wg
18:29:17 [Zakim]
Zakim has left #rdf-wg
19:17:19 [gkellogg]
gkellogg has joined #rdf-wg
20:21:30 [davidwood]
davidwood has joined #rdf-wg