IRC log of rdf-wg on 2013-10-09

Timestamps are in UTC.

15:00:47 [RRSAgent]
RRSAgent has joined #rdf-wg
15:00:47 [RRSAgent]
logging to
15:00:49 [trackbot]
RRSAgent, make logs world
15:00:49 [Zakim]
Zakim has joined #rdf-wg
15:00:51 [trackbot]
Zakim, this will be 73394
15:00:51 [Zakim]
ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start now
15:00:52 [trackbot]
Meeting: RDF Working Group Teleconference
15:00:52 [trackbot]
Date: 09 October 2013
15:01:21 [Arnaud]
Arnaud has joined #rdf-wg
15:01:40 [TallTed]
Zakim, this is rdf
15:01:40 [Zakim]
ok, TallTed; that matches SW_RDFWG()11:00AM
15:01:43 [TallTed]
Zakim, who's here?
15:01:43 [Zakim]
On the phone I see [GVoice], OpenLink_Software
15:01:45 [Zakim]
On IRC I see Arnaud, Zakim, RRSAgent, pfps, ivan, gkellogg, AndyS, TallTed, davidwood, yvesr, manu, trackbot, sandro, ericP
15:01:52 [TallTed]
Zakim, OpenLink_Software is temporarily me
15:01:53 [Zakim]
+TallTed; got it
15:01:53 [TallTed]
Zakim, mute me
15:01:53 [Zakim]
TallTed should now be muted
15:02:02 [Zakim]
15:02:28 [pfps]
zakim, who is here?
15:02:28 [Zakim]
On the phone I see [GVoice], TallTed (muted), Danny
15:02:29 [Zakim]
On IRC I see Arnaud, Zakim, RRSAgent, pfps, ivan, gkellogg, AndyS, TallTed, davidwood, yvesr, manu, trackbot, sandro, ericP
15:02:35 [Zakim]
15:02:40 [pfps]
zakim, gvoice is me
15:02:40 [Zakim]
+pfps; got it
15:02:45 [Zakim]
15:02:53 [Zakim]
15:03:00 [ivan]
zakim, mute me
15:03:00 [Zakim]
Ivan should now be muted
15:03:15 [Zakim]
15:03:19 [AndyS]
zakim, IPcaller is me
15:03:19 [Zakim]
+AndyS; got it
15:03:28 [PatH]
PatH has joined #rdf-wg
15:03:45 [Zakim]
15:03:55 [pfps]
zakim, who is here?
15:03:55 [Zakim]
On the phone I see TallTed (muted), Danny, Arnaud, Ivan (muted), PatH, AndyS
15:03:56 [markus]
markus has joined #rdf-wg
15:03:58 [Zakim]
On IRC I see PatH, Arnaud, Zakim, RRSAgent, pfps, ivan, gkellogg, AndyS, TallTed, davidwood, yvesr, manu, trackbot, sandro, ericP
15:04:15 [Zakim]
+ +1.415.686.aaaa
15:04:19 [gkellogg]
zakim, I am aaaa
15:04:19 [Zakim]
+gkellogg; got it
15:04:30 [Zakim]
+ +1.540.898.aabb
15:04:31 [Zakim]
15:04:37 [davidwood]
Zakim, I am aabb
15:04:37 [Zakim]
+davidwood; got it
15:05:06 [gkellogg]
scribe: gkellogg
15:05:24 [gkellogg]
scribenick: gkellogg
15:06:00 [davidwood]
Zakim, who is here?
15:06:00 [Zakim]
On the phone I see TallTed (muted), Danny, Arnaud, Ivan (muted), PatH, AndyS, gkellogg, davidwood, EricM
15:06:02 [Zakim]
On IRC I see markus, PatH, Arnaud, Zakim, RRSAgent, pfps, ivan, gkellogg, AndyS, TallTed, davidwood, yvesr, manu, trackbot, sandro, ericP
15:06:10 [pfps]
zakim, who is talking?
15:06:17 [Zakim]
15:06:18 [markus]
zakim, ??P28 is me
15:06:18 [Zakim]
+markus; got it
15:06:19 [davidwood]
PROPOSED to accept the minutes of the 9 Oct telecon:
15:06:19 [davidwood]
15:06:19 [davidwood]
15:06:22 [Zakim]
pfps, listening for 10 seconds I heard sound from the following: davidwood (18%)
15:06:24 [sandro]
zakim, EricM is temporarily Sandro
15:06:24 [Zakim]
+Sandro; got it
15:07:33 [davidwood]
RESOLVED: Accept the minutes of the 9 Oct telecon:
15:07:53 [davidwood]
15:08:00 [gkellogg]
topic: open action items
15:08:31 [pfps]
I changed the status of a number of items that appear to be done
15:10:06 [gkellogg]
davidwood: guus has an agenda item about public responses. His preferences would be that editors send an ack of receipt and create an issue for it.
15:10:23 [gkellogg]
… Other members of the group should not comment on public-rdf-comments until we've had a chance to discuss it.
15:10:53 [gkellogg]
… (either email or on telecon). We should agree if anyone other than editors will make a response.
15:11:18 [sandro]
pfps, are you going to be able to get on the phone soon?
15:11:23 [davidwood]
pfps, do you want us to wait to discuss the test suite for Semantics? Will you be able to join by phone later?
15:11:32 [PatH]
peter, can we do the sematnics test suite or does it need phone input?
15:12:08 [pfps]
let's wait for a bit
15:12:12 [davidwood]
15:12:23 [gkellogg]
topic: Turtle
15:12:33 [sandro]
pfps, we're moving on for now, hoping you'll join us soon.
15:12:54 [ericP]
Zakim, please dial ericP-mobile
15:12:54 [Zakim]
ok, ericP; the call is being made
15:12:55 [Zakim]
15:13:03 [gkellogg]
davidwood: issue-119, spec should reference the test suite
15:13:09 [AZ]
AZ has joined #rdf-wg
15:14:14 [gkellogg]
action: ericp to update Turtle ED to reference the test suite.
15:14:14 [trackbot]
Created ACTION-306 - Update turtle ed to reference the test suite. [on Eric Prud'hommeaux - due 2013-10-16].
15:14:31 [gkellogg]
ericp: I had to do some stuff by hand to the publication, so that needs to be copied back.
15:14:32 [Zakim]
+ +81.16.5.aacc
15:14:43 [gkellogg]
15:14:43 [trackbot]
action-152 -- Gavin Carothers to Create new issue for :'s in the local part of prefix names -- due 2012-03-21 -- CLOSED
15:14:43 [trackbot]
15:14:44 [AZ]
Zakim, aacc is me
15:14:44 [Zakim]
+AZ; got it
15:15:00 [Zakim]
15:15:12 [gkellogg]
ericp: We're not inventing much, and there's no current candidate for how to do this reification.
15:15:31 [gkellogg]
… Did it disappear?
15:15:33 [Zakim]
15:15:36 [gkellogg]
sandro: I don't think so.
15:15:59 [Zakim]
15:16:18 [sandro] 2.17 Reifying Statements: rdf:ID
15:16:19 [gkellogg]
ericp: I prepose we say that there's no existing proposal for this; we're largely sticking with Turtle from before we got docs. There's no plan to do something about this, very sorry.
15:16:35 [gkellogg]
davidwood: fine with ericp giving a response.
15:16:48 [zwu2]
zwu2 has joined #rdf-wg
15:16:50 [gkellogg]
ericp: not that WG has provided a standard for datasets and "named graphs"
15:16:57 [gkellogg]
davidwood: point him to TriG
15:17:19 [gkellogg]
action: ericp to respond to Joe Rockmore relating to issue-152 and close if okay
15:17:19 [trackbot]
Created ACTION-307 - Respond to joe rockmore relating to issue-152 and close if okay [on Eric Prud'hommeaux - due 2013-10-16].
15:17:44 [gkellogg]
eric: another substantial issue which none addressed.
15:18:02 [gkellogg]
… There was a question from Jan ???; is there any versioning on Turtle.
15:18:06 [Zakim]
+ +1.619.663.aadd
15:18:16 [gkellogg]
… Can I say that there is no proposal for versioning turtle.
15:18:17 [zwu2]
zakim, +1.619.663.aadd is me
15:18:17 [Zakim]
+zwu2; got it
15:18:37 [gkellogg]
… It would help people if using accept headers to use new or old turtle.
15:18:56 [gkellogg]
… It's possible that adding a media-type parameter would keep us from having to do another LC
15:19:04 [gkellogg]
davidwood: it would be nice to do something.
15:19:16 [gkellogg]
ericp: it would help people trying to make the transformation.
15:19:28 [AndyS]
15:19:35 [Zakim]
15:19:36 [gkellogg]
action: ericp to write up proposal for adding a parameter to the media type registration, if it doesn't require a new LC
15:19:36 [trackbot]
Created ACTION-308 - Write up proposal for adding a parameter to the media type registration, if it doesn't require a new lc [on Eric Prud'hommeaux - due 2013-10-16].
15:19:49 [ericP]
15:19:52 [AndyS]
media type games do not work.
15:19:57 [Zakim]
15:20:09 [gkellogg]
ericp: I think it's comment 33 in the CR LC comments
15:20:12 [ericP]
15:20:40 [Zakim]
15:21:07 [gkellogg]
topic: JSON-LD
15:21:08 [AndyS]
This is a substantive change to TTL if the MIME type registration changes IMO
15:21:09 [ericP]
issue: media type parameter for Turtle
15:21:10 [trackbot]
Error creating an ISSUE: an internal error occurred. Please mail <> with details about what happened.
15:21:20 [ericP]
issue: media type parameter for Turtle
15:21:20 [trackbot]
Error creating an ISSUE: an internal error occurred. Please mail <> with details about what happened.
15:21:37 [gkellogg]
davidwood: what's the implication of semantics and comments not being in CR for JSON-LD PR transition?
15:22:01 [ericP]
issue: media type parameter for Turtle
15:22:02 [trackbot]
Error creating an ISSUE: an internal error occurred. Please mail <> with details about what happened.
15:22:03 [gkellogg]
… Right now, JSON-LD and JSON-LD-API point to Concepts LC with informative semantics, schema and turtle.
15:22:15 [gkellogg]
… JSON-LD-API has norm to semantics and inform to Turtle.
15:22:18 [AndyS]
15:22:29 [gkellogg]
… I thought we could leave a norm ref to an LC doc
15:22:41 [Souri]
Souri has joined #rdf-wg
15:22:42 [ericP]
action: beat the tracker into submission and create an issue for a version media type parameter in Turtle
15:22:42 [trackbot]
Error finding 'beat'. You can review and register nicknames at <>.
15:22:46 [gkellogg]
sandro: The normal rule is you can only go back one level, a PR can only point to a CR, not LC.
15:23:00 [gkellogg]
… Although, we've been moving more to "how stable is this material"
15:23:02 [ivan]
15:23:10 [gkellogg]
… We'd be asking for a waver on the rules to do this.
15:23:13 [ivan]
zakim, unmute me
15:23:13 [Zakim]
Ivan should no longer be muted
15:23:16 [Zakim]
15:23:23 [gkellogg]
… I think JSON-LD isn't ready due to another dependency problem.
15:23:40 [gkellogg]
… Even if we could go to PR, do we want the REC to point to something other than the other REC?
15:24:15 [gkellogg]
sandro: we can't make a change to the rec after publication.
15:24:18 [PatH]
15:24:29 [davidwood]
ack ivan
15:24:30 [gkellogg]
davidwood: we've done with LC for semantics and concepts, we just need to clean up
15:24:55 [gkellogg]
ivan: the issue is when we publish REC for JSON-LD, the clean way to do it is to publish them all together.
15:25:15 [gkellogg]
… Anything else will lead to long discussions. I think we're only talking about 1-2 weeks.
15:25:30 [gkellogg]
… An edited document needs to go through same AC voting.
15:25:49 [gkellogg]
sandro: we're talking about 2 months; they want to go no, and the rest are in December
15:27:16 [gkellogg]
ivan: I don't see this as being solvable. Going to CR now, providing we get the other docs in as well.. The publication of the REC would wait a few months, but in some sense, it looks more ridiculous if we publish a REC in Dec and an edited version a few months later.
15:27:30 [gkellogg]
sandro: not sure the world can tell the difference between PR/CR and REC
15:27:44 [gkellogg]
… I also don't think JSON-LD is ready because of Promises.
15:27:54 [PatH]
pfps, we are talking baout JSON publication being delayed by references to RDF.
15:28:03 [gkellogg]
ivan: I believe the dependencies go to a non-normative section.
15:28:26 [gkellogg]
… In this sense, we may have to go through a second LC until Promises is resolved.
15:28:36 [gkellogg]
davidwood: the shortest path is to move concepts and semantics to CR
15:28:56 [ivan]
zakim, mute me
15:28:56 [Zakim]
Ivan should now be muted
15:29:20 [gkellogg]
… If we can get concepts & semantics to CR we can push JSON-LD docs to PR
15:29:27 [davidwood]
Topic: LC Comments on Concepts, Semantics
15:29:27 [gkellogg]
topic: Concepts and Semantics
15:29:35 [PatH]
pfps, not yet.
15:29:49 [davidwood]
15:29:49 [trackbot]
ISSUE-142 -- LC comment: rdfs:Graph ? comment -- open
15:29:49 [trackbot]
15:29:51 [gkellogg]
davidwood: we're now looking at issue-142
15:30:42 [gkellogg]
pfps: I believe we discussed last week and declined it.
15:30:46 [ivan]
because nobody answered
15:30:49 [ivan]
15:30:53 [gkellogg]
sandro: someone was supposed to reply to him
15:30:55 [ivan]
zakim, unmute me
15:30:55 [Zakim]
Ivan should no longer be muted
15:31:05 [PatH]
I dont think a formal reply was sent (?)
15:31:44 [gkellogg]
ivan: I think pfps is right, we forced through a resolution to close it, but we stopped there and didn't give an action to send a reply back to Jeremy
15:31:59 [gkellogg]
path: I agreed with pfps
15:32:17 [gkellogg]
ivan: we agreed that the REC will not touch anything around datasets anymore
15:33:11 [AndyS]
15:33:11 [trackbot]
action-304 -- Peter Patel-Schneider to Send a proposed response to jeremy to the wg for comments -- due 2013-10-09 -- CLOSED
15:33:11 [trackbot]
15:33:18 [gkellogg]
sandro: pfps took the action on this, and action-304 on pfps to send to WG
15:33:36 [gkellogg]
pfps: that's been done (I think)
15:33:59 [AndyS]
15:34:09 [gkellogg]
path: the email thread ended with everyone agreeing, but no-one doing anything
15:35:26 [gkellogg]
sandro: meta-comment, please put "-" between "issue" and number so tracker can pick them up.
15:35:57 [gkellogg]
pfps: Jeremy's comment in issue-38, and who to track in issue-142
15:36:16 [gkellogg]
path: issue-142 quotes from Jeremy's comments, but doesn't mention OWL.
15:36:31 [gkellogg]
davidwood: what I see in Jeremy's comments was issue-35
15:36:45 [gkellogg]
sandro: I believe he said that one was supplanting the other.
15:36:56 [gkellogg]
pfps: then something is messed up with ???
15:36:59 [davidwood]
Looking at
15:36:59 [markus]
I think this response is for ISSUE-151:
15:37:31 [gkellogg]
davidwood: the LC Comments say that pfps proposed to issue-38, which relates to issue-151.
15:37:32 [AZ]
ISSUE-38 and ISSUE-35 are very similar
15:37:57 [gkellogg]
path: at some level, they're the same issue.
15:38:10 [gkellogg]
sandro: I think in our reply we can address several at once.
15:38:23 [AZ]
ISSUE-35 is about rdfs:Graph while ISSUE-38 is more generally about any vocabulary that would allow to talk about graphs
15:39:01 [gkellogg]
davidwood: how about combining issue-35, issue-38, issue-142, and issue-151
15:39:29 [gkellogg]
pfps: I would like to excise the owl:imports issue from the other bits. If we don't, it will continue to infect every other issue.
15:39:51 [gkellogg]
… I'd like to get a clean break on owl:imports and say that the WG isn't going to touch that.
15:40:20 [gkellogg]
sandro: I don't think that's the specific issue, just an example. You need to say how it works with RDF
15:40:45 [gkellogg]
davidwood: The proposed response is that it's out of scope for this WG
15:40:58 [gkellogg]
sandro: it's not, owl:imports is specifically, but the issue it raises isn't
15:41:16 [davidwood]
Peter's proposed response:
15:41:49 [gkellogg]
sandro: I agree with this response, but it disagrees with pfps
15:42:02 [Zakim]
15:42:07 [pfps]
That's my message!
15:42:15 [davidwood]
Further, if you look at the top two lines of I think you will agree that that the response column entries are swapped.
15:42:45 [sandro]
s/this response/what davidw is saying/
15:43:00 [gkellogg]
davidwood: pfps' proposed response to Jeremy says that the practice ...
15:43:03 [Zakim]
15:43:16 [gkellogg]
… The WG won't be addressing this, and it should be raised against OWL.
15:43:52 [gkellogg]
… sandro and I said the response is insufficient, as Jeremy is just using this as an example to illustrate his point on graph naming. We need to respond to that, and not on owl:imports specifically.
15:44:29 [gkellogg]
pfps: if we're doing owl:imports and graph names in issue-38, then I think it is about owl:imports.
15:44:47 [gkellogg]
… How can this not be about owl:imports
15:45:02 [gkellogg]
path: because the only relevant point is that owl:imports references a graph.
15:45:21 [gkellogg]
sandro: he's not saying that owl:imports doesn't work, but it doesn't work when you put it in a dataset
15:45:43 [gkellogg]
… Datasets are a new thing, and we should say how existing stuff, like owl:imports, works with the new stuff we're creating.
15:46:00 [gkellogg]
davidwood: Jeremy's saying that the new stuff breaks owl:imports
15:46:30 [AndyS]
In owl:imports it refers to a g-box.
15:46:38 [AZ]
owl:imports does not force the imported thing to denote a graph at all
15:46:38 [gkellogg]
path: it's not that it breaks owl:imports, it breaks the presumption that owl:imports <foo> that <foo> references a graph.
15:46:46 [ivan]
15:46:54 [gkellogg]
pfps: owl:imports doesn't work that way, it's very delicate.
15:46:59 [ivan]
zakim, unmute me
15:46:59 [Zakim]
Ivan was not muted, ivan
15:47:19 [AZ]
+1 pfps
15:47:25 [gkellogg]
… Whatever the common practice is seems to be working, but it has nothing to do with RDF per-say, but with the OWL layer on top of RDF.
15:47:43 [gkellogg]
… The RDF WG does nothing to prevent this from working, and shouldn't do anything else about it.
15:47:47 [davidwood]
ack ivan
15:48:22 [gkellogg]
ivan: we are arguing over what a reaction might be. Let's send the response back to Jeremy; strictly speaking pfps is right.
15:48:49 [gkellogg]
… I expect that Jeremy will respond by saying that the real issue is about datasets; Let's not put words into Jeremy's mouth.
15:49:13 [gkellogg]
sandro: can we handle the other issue first, so it's clear we're not going to do anything about datasets (issue-142)
15:49:29 [ivan]
zakim, mute ivan
15:49:29 [Zakim]
Ivan should now be muted
15:49:34 [pfps]
somehow the issues got conflated, which may have been my mistake somewhere
15:49:50 [gkellogg]
davidwood: we should respond to issue-35 and issue-142 first
15:50:08 [gkellogg]
… Do you have a proposed response for issue-35?
15:50:17 [gkellogg]
pfps: I may have mixed up the issues.
15:51:03 [pfps]
will do
15:51:13 [gkellogg]
davidwood: would you take a look at the issues and make sure they're referenced on the summary page?
15:51:33 [davidwood]
15:51:33 [trackbot]
ISSUE-147 -- LC comment: Definition of "Generalized RDF" -- open
15:51:33 [trackbot]
15:52:02 [pfps]
15:52:22 [gkellogg]
sandro: is the connects on issue-147 reflected in the ED? If so, we can send a response to say we accepted your wording.
15:52:41 [gkellogg]
davidwood: I think he wanted to move the definition to semantics.
15:53:02 [gkellogg]
path: we decided to not move it to semantics, but there was some wording from AZ that I think we should use on Concepts.
15:53:17 [gkellogg]
… This avoids the main objection of david booth's
15:53:31 [pfps]
so someone has to make the change to concepts and send out the message - my proposed response is now in concepts but Pat didn't like that
15:53:52 [gkellogg]
… The advantage of this wording is that it avoids david's objection. One should not say that a generalized triple/graph is a triple/graph
15:54:09 [gkellogg]
… In general, it's appropriate to be in concepts, and not semantics
15:54:24 [gkellogg]
davidwood: in section 7 of concepts, there's a note.
15:54:37 [pfps]
the latest version of concepts has my proposed wording
15:54:39 [gkellogg]
… This says that use is non-standard and may result in interoperability problems.
15:55:08 [gkellogg]
… That note was put in place based on david's phraseology. This issue is a separate request to move to semantics.
15:55:20 [Zakim]
15:55:23 [gkellogg]
path: I agree we shouldn't do this.
15:55:55 [gkellogg]
… He also doesn't like the wording, which is still in the concepts draft. I think we can better phrase this, as suggested by AZ and modified by me in email.
15:56:07 [Zakim]
15:56:25 [sandro]
15:57:14 [pfps]
That wording is fine by me.
15:57:18 [gkellogg]
issue: davidwood to make an editorial change to concepts in answer to issue-147
15:57:18 [trackbot]
Error creating an ISSUE: an internal error occurred. Please mail <> with details about what happened.
15:57:47 [gkellogg]
issue: davidwood to make an editorial change to concepts in answer to action-147
15:57:47 [trackbot]
Error creating an ISSUE: an internal error occurred. Please mail <> with details about what happened.
15:58:27 [gkellogg]
close issue-157
15:58:27 [trackbot]
Closed issue-157.
15:58:36 [PatH]
"the error is in the message" - Marshall McLuhan
15:58:57 [sandro]
RRSAgent, pointer?
15:58:57 [RRSAgent]
15:59:25 [gkellogg]
action: davidwood to make an editorial change to concepts in answer to issue-147
15:59:25 [trackbot]
Created ACTION-309 - Make an editorial change to concepts in answer to issue-147 [on David Wood - due 2013-10-16].
15:59:35 [gkellogg]
close issue-158
15:59:35 [trackbot]
Closed issue-158.
16:00:08 [gkellogg]
16:00:08 [trackbot]
issue-148 -- LC comment: IRIs do *not* always denote the same resource -- open
16:00:08 [trackbot]
16:00:25 [Zakim]
16:00:35 [gkellogg]
16:00:35 [trackbot]
issue-149 -- LC comment: Intuitive summary needs to be scoped to interpretations -- open
16:00:35 [trackbot]
16:01:01 [gkellogg]
path: my preferred response to issue-149 is to simply delete the section.
16:01:19 [Zakim]
16:01:25 [gkellogg]
… It's a short 2-3 paragraphs with an intuitive understanding that's causing more problems than it's solving.
16:01:51 [Zakim]
16:02:01 [pfps]
I think that David's response is acceptable.
16:02:08 [Zakim]
16:02:22 [gkellogg]
pfps: I would prefer putting in davids ???
16:02:50 [gkellogg]
davidwood: I think he'd' prefer putting in david woods suggestion rather than just removing.
16:02:56 [pfps]
I'm happy with David's response, but if Pat wants to rip out the section, then let's try that.
16:03:04 [gkellogg]
path: I would rather not do that. There's a long history to this.
16:03:34 [markus]
16:03:34 [gkellogg]
… I think putting this in could cause more problems than it solves
16:03:39 [sandro]
16:03:55 [gkellogg]
davidwood: it's for path and pfps to work out as editors
16:04:07 [gkellogg]
pfps: question is, who to respond?
16:04:19 [gkellogg]
path: my response would be that we just remove the section.
16:04:44 [gkellogg]
pfps: my proposed response could be changed to respond
16:05:13 [gkellogg]
action: pfps to respond to david black on issue-149
16:05:13 [trackbot]
Created ACTION-310 - Respond to david black on issue-149 [on Peter Patel-Schneider - due 2013-10-16].
16:06:31 [pfps]
I think that all we need is direction from the chairs as to who to ask
16:06:35 [pfps]
16:06:36 [gkellogg]
16:06:36 [trackbot]
issue-150 -- LC Comment: references and acknowledgements -- open
16:06:36 [trackbot]
16:07:57 [gkellogg]
davidwood: removal of editors was a ReSpec problem that would be addressed, but I haven't figured it out yet.
16:07:57 [ivan]
I believe this was a 'private' response, that was not added to tracker...
16:08:44 [markus]
respec supports this
16:08:49 [gkellogg]
sandro: if you think it's a problem, sent to spec-prod
16:09:00 [gkellogg]
davidwood: it used to flow through, and doesn't any more.
16:09:10 [gkellogg]
16:09:18 [gkellogg]
… Or look through archives.
16:09:50 [markus]
16:10:00 [gkellogg]
pfps: quick update on test suite
16:10:05 [gkellogg]
topic: semantics test-suite
16:10:14 [gkellogg]
pfps: I've got my part done, but waiting for landing page.
16:10:35 [gkellogg]
… AZ has a set of "torture tests", but it might be good to try to run them through an existing implementation first.
16:10:48 [markus]
that explains how to fix your respec issue
16:10:58 [gkellogg]
path: is it the intention that all RDF engines should pass this, or is it just a collection of "hard problems"
16:11:21 [gkellogg]
sandro: everything that claims to be an RDF simple-entailment reasoner needs to pass, but do we define this?
16:11:44 [gkellogg]
path: some of AZ's tests were rather obscure. An engine that couldn't handle them would still be useful.
16:12:12 [ericP]
Zakim, please dial ericP-mobile
16:12:12 [Zakim]
ok, ericP; the call is being made
16:12:14 [Zakim]
16:12:23 [gkellogg]
… They depend on very arcane properties of different datatypes. For example, two datatypes that only have one item in their intersection, which can force classes to be singletons.
16:12:38 [gkellogg]
… This is clever, but maybe not to useful for an actual implentation
16:13:00 [gkellogg]
AZ: We did this in 2004 someplace.
16:13:13 [gkellogg]
sandro: it was in RDF tests, which we're not bringing forward.
16:13:21 [gkellogg]
AZ: we need chair guidance for how to proceed
16:13:36 [AZ]
16:13:39 [gkellogg]
sandro: this comes back to cygri's point about semantics being optional, but we didn't get consensus.
16:13:40 [PatH]
16:14:30 [zwu2]
has to go to another meeting. Bye guys.
16:14:43 [Zakim]
16:14:46 [gkellogg]
sandro: most systems which handle RDF don't think of themselves as entailment- or consistency-checkers
16:15:05 [gkellogg]
… There could be some code which would be different if semantics was different
16:15:17 [gkellogg]
… Nothing about value-spaces of datatypes.
16:15:33 [gkellogg]
pfps: there should be code which at least looks up the class hierarchy.
16:15:44 [gkellogg]
path: I don't think you'd expect RDF reasoners to do that.
16:15:55 [Zakim]
16:16:05 [pfps]
the first RDF working group was before we were born
16:16:29 [gkellogg]
sandro: that was on RDFS, not RDF, right?
16:16:33 [gkellogg]
pfps: yes.
16:16:47 [gkellogg]
sandro: for non-RDFS, simple entailment ...
16:17:05 [PatH]
datatypes are now in basic RDF, not an extension to RDFS
16:17:11 [gkellogg]
pfps: the tricky ones from AZ were RDFS and positivie/negative integer, which intersect on a singleton
16:17:53 [gkellogg]
path: in addition, you take that singleton and say some very strange things, like an rdf:Property is a subclass of it
16:18:15 [gkellogg]
sandro: you're assuming there's a conformance clause?
16:18:35 [gkellogg]
path: I'm worried that people will take the test suite and fail it and be upset, which they shouldn't be
16:18:48 [gkellogg]
… We include datatypes in basic RDF now.
16:19:05 [gkellogg]
sandro: my point is that most implementations that claim to be RDF but not RDFS don't do such reasoning.
16:19:19 [gkellogg]
… I think RDF-Concepts defines such a system.
16:19:30 [gkellogg]
davidwood: yes it does, but not say anything about that.
16:19:56 [Zakim]
16:20:05 [AZ]
Zakim, ??P27
16:20:05 [Zakim]
I don't understand '??P27', AZ
16:20:07 [AZ]
Zakim, ??P27 is me
16:20:07 [Zakim]
+AZ; got it
16:20:11 [gkellogg]
… In relation to blank-nodes, we talk about systems for implementing skolemization, but don't define what that is.
16:20:32 [gkellogg]
sandro: I think Pat has something in mind, and it makes sense to define what such a system must do.
16:20:45 [gkellogg]
path: why are we publishing test cases? What's the point?
16:20:59 [gkellogg]
sandro: it only makes sense if we think people are implementing something here.
16:21:10 [pfps]
I updated them because someone thought that it was a good thing to do.
16:21:11 [gkellogg]
davidwood: we're running out of time today.
16:21:55 [Zakim]
16:21:56 [Zakim]
16:21:57 [Zakim]
16:22:00 [Zakim]
16:22:06 [Zakim]
16:23:04 [Zakim]
16:24:28 [sandro]
16:47:32 [Zakim]
16:47:34 [Zakim]
16:47:37 [Zakim]
16:47:38 [Zakim]
16:47:47 [Zakim]
16:52:47 [Zakim]
disconnecting the lone participant, Danny, in SW_RDFWG()11:00AM
16:52:49 [Zakim]
SW_RDFWG()11:00AM has ended
16:52:49 [Zakim]
Attendees were TallTed, Danny, Arnaud, pfps, Ivan, PatH, AndyS, +1.415.686.aaaa, gkellogg, +1.540.898.aabb, davidwood, markus, Sandro, EricP, +81.16.5.aacc, AZ, Souri, zwu2,
16:52:50 [Zakim]
... [GVoice]
18:37:34 [Zakim]
Zakim has left #rdf-wg