See also: IRC log
<trackbot> Date: 04 December 2013
<Guus> chair: Guus
<scribe> scribe: sandro
guus: busy agenda, expecting 75 minutes
<pfps> minutes look OK
gregg: Have the tests been moved into their final location?
<ivan> I did
<AndyS> e.g. http://www.w3.org/2013/N-TriplesTests/
page now points: The test suite is at https://dvcs.w3.org/hg/rdf/file/d2aa3c0cdec7/rdf-turtle/tests-nt.
guus: PLEASE contact your AC rep to vote on JSON-LD
guus: rdf-mt is being challening
pfps: We've had two sort-of
responses, one from Olivier and one from Michael
Schneider
... Olivier's Corese -- I looked at it -- doesn't produce the
expected answer for 7 of the 42 tests.
<gkellogg> https://dvcs.w3.org/hg/rdf/raw-file/default/rdf-mt/reports/index.html
pfps: biggest reason is they
don't do anything follows from a contradiction
... I think we once decided not to test that.
... I suggest we change those tests to be directly a
contradiction, instead of anything-follows-from
... containerMembershipProperties ... lots of systems don't do
anything with them if they don't occur in the input
... A solution is to just put it in the input as a
crutch.
... Also, something with integer and string, I still need to
look into that.
guus: How many of the 7 would you consider real failures
pfps: None.
... Very few people actually implement "entailment" per se;
instead the implement a sort of "completion" thing. So they'll
detect contractions, but they wont implement a couple of the
structural rules, since those are done at the matching phase at
the end
... Oh, maybe one. Inclusion detection between integer and
string. That may be a failure.
... If you say integer is a subclass of string, that
... If you say integer is a subclass of string, that's false,
and Corese doesn't detect that.
... I need to ping Olivier back a little more.
... There was an error in the manifest that I corrected.
<markus> shouldn't the report at https://dvcs.w3.org/hg/rdf/raw-file/default/rdf-mt/reports/index.html show the content of the turtle files? I just get "....ttl not loaded" everywhere; gkellogg !?
<PatH> FWIW, XSD says integers aren't strings.
pfps: I sent personal messages to the people I thought were appropriate, but haven't gotten encouraging responses.
<PatH> OK, neveeerrrrr miiind.
sandro: maybe me, or the chairs, should try contacting them as well?
pfps: Maybe they're working on them. :-)
<PatH> tried jos de Roo?
sandro: What about Mike Dean & associated Silk folks?
<PatH> Parliament doing RDF reasoning sounds rather unlikely to me.
<ericP> PatH, jos would probably use euler, which is actually two cwm clones, one implemented in C# and the other in prolog
<PatH> i think it would be conforming to simply report contradictions as an error.
<ericP> PatH, does prolog implement "everything follows from a contradiction"?
pfps: A rule based system can do all of these except Everything Follows from a Contradiction
<PatH> EricP, if you mean, does it actually draw those ocnclusions, no.
sandro: yes, let's change the
everything-follows-from-a-contradiction
... WHat about Jan Weilemaker?
<PatH> YOu can map it all into a FOL style syntax.
pfps: Oxford is close, except for
predicate variables
... If you map to FOL it's not triples any more
... whatever representation, it needs to handle variables in
its predicate position.
I use one predicate: rdf(s,p,o)
pfps: The Oxford folks are saying that's not as efficient.
PROPOSED: Change the semantics tests to use an explicit contradiction marker, in the test manifest, instead of having everything following from a contradiction.
<PatH> I think they should be required to explode, but this MUST be interpreted as an error condition.
<Guus> +1
<pfps> +1
<ivan> +1
<gkellogg> +1
<PatH> +1
<zwu2> +1
+1
<ericP> +1
<AZ> +1
<pchampin> +1
<markus> +1
<davidwood> +1
RESOLUTION: Change the semantics tests to use an explicit contradiction marker, in the test manifest, instead of having everything following from a contradiction.
pfps: I'll make those changes and talk to Olivier about re-running his stuff.
sandro: I think we need to resolve PR a week from today.
guus: We made our lives difficult with semantics tests.
<PatH> Even with the 2004 semantics, RDFS allows contradictions. What was done with such tests then?
pfps: I'll send contact list to chairs
guus: We'll try to keep track of them over the coming weeks.
gregg: Got result from Dave
Backet, who is NOT doing TriG
... We have 2 impls for N-Quads and TriG (Jena and Ruby RDF)
and 3 for NTriples, with raptor.
guus: So all four languages are
fine, as far as exit criteria
... All impl reports up to date?
gregg: small update to N-Triples
guus: Excellent.
eric: Are Turtle parsers N-Triples parsers?
gregg: Not really, because of the Negative Syntax tests.
guus: So our only CR-exit concern is rdf-mt
guus: It looks like Turtle are all resolved, yes?
ericP: Yes.
guus: Proposal from Andy on closing 4 Turtle issues. Okay?
<Guus> http://lists.w3.org/Archives/Public/public-rdf-wg/2013Nov/0159.html
PROPOSED: close issue-172 issue-173 issue-174 issue-175 as per http://lists.w3.org/Archives/Public/public-rdf-wg/2013Nov/0159.html
<AndyS> +1
<ivan> +1
<gkellogg> +1
<pfps> +1
<zwu2> +1
andy: They've all been accepted by the commenters, details in issue writeups
+1
<TallTed> +1
<markus> +1
<pchampin> +1
RESOLUTION: close issue-172 issue-173 issue-174 issue-175 as per http://lists.w3.org/Archives/Public/public-rdf-wg/2013Nov/0159.html
<AZ> +1
<yvesr> +1
guus: Need handling of comments on N-triples and N-Quads. Maybe we need another editor, since Gavin hasn't been doing it.
andy: If necessary
... Okay, I'll do it.
guus: Thanks!
<davidwood> +1
<davidwood> Thanks, Andy
guus: Comments on escape characters? Real issue?
andy: I don't think so, but I
haven't looked closely yet.
... It's editorial, "Would you clarify..."
guus: I replied to David Booth
today
... Maybe we should send some reminders
... IRIs and globals, and M. Schn.
<scribe> ACTION: guus send reminders for comments 1 and 2 [recorded in http://www.w3.org/2013/12/04-rdf-wg-minutes.html#action01]
<trackbot> Created ACTION-331 - Send reminders for comments 1 and 2 [on Guus Schreiber - due 2013-12-11].
<PatH> I think we already sent a response to mIchael, no?
(the leftovers from Last Call)
<PatH> Ah, ok.
guus: Just reminders.
... Please start getting your documents in order for Proposed
Recommendation.
<PatH> What needs to be done?
<AndyS> CR #13 is a little more than editorial.
guus: Document is fine, aside from Boiler Plate?
pfps: Yes.
markus: Section on Literals, maybe.
davidwood: Section at the bottom on differences is Way Out Of Date. I propose cutting it and instead having a pointer to the What's New Note.
guus: Turtle document status?
ericP: I believe the document is in order.
<davidwood> Concepts Appendix A is out of date and should be removed in favor of the What's New note.
guus: Schema and XML later
... decision on 11th or 18th. Hm.
... Schema is coming along and pretty good.
<PatH> I trust you. Guus.
( Note Publication Moratorium: starting 18 December 2013 through 1 January 2014 )
<PatH> months????
<PatH> <strangled noises />
guus: I put my name in the Schema acknowledgement sections, but didnt list myself as an editor.
<PatH> OK
guus: I removed the schema
appendix.
... it was a courtesy, not it's a discourtesy,.
sandro: Title. Ugh. Oh well.
guus: Fabien isn't here. THere's a day of editorial work to day. I'll might do it if Fabien can't. We didn't promise it in any case.
<davidwood> NO
<davidwood> :)
<pchampin> I'm buying!
sandro: My sense is that the nsdoc is NOT the WG's responsibility, but I'll definitely ask for WG review.
guus: Sounds okay.
guus: Pat, what about your reviews?
pat: I think the document is basically okay, although it could use some editorial work and polish.
guus: Plan is to publish in December and then republish at Rec time.
an: I've read the comments; I'll reply soon. Some discussion to do.
AZ: Most of them should be fine.
guus: You CAN leave Open Issue in the document. It doesn't need to be consensus in this version.
sandro: That suggests we'll all be doing RDF WG stuff during January
pat: That's life.
guus: Please go ahead without my review.
guus: Very extensive review from Pat. Ongoing discussion. Figures are not quite right yet.
<yvesr> (figures should be there this evening)
guus: I've adopted MOST of Pat's comments. I hope you like the rephrasing about Statements, Pat.
<PatH> i need to go over the latest version.
guus: main thing left is how we talk about Multiple Graphs.
PatH: I'm not up on the latest email from Guus.
guus: On Multiple Graphs: <reading>
<tbaker> what is the time-frame for review and revision of the Primer?
PatH: A Literal is a Resource.
sandro: Literals and IRI are types of Terms; they denote Resources
Pat: Right.
guus: I'm thinking about publishing soon so we can do a revision for Rec.
Pat: I'm still not happy about the "Meta" stuff.
guus: We'll put in some warning notes about that.
ivan, I don't agree
<yvesr> so who should?
<yvesr> PatH, we had an explicit requirement to add a provenance example to the primer
pat: THe Primery must say: ONLY DO THIS IF YOU HAVE OUT-OF-BAND AGREEMENT, YOU CAN'T DO THIS ON THE OPEN WEB.
markus: Even with that warning, people will still do it.
pat: I'm hearning people who think we're giving them semantics.
<ivan> can we refer to Antoine's document in the primer?
<yvesr> sandro, and which are the two different ways?
<Arnaud> is the primer the place for airing this kind of dirty laundry??
<PatH> +1 Arnaud
It's the place for telling people things they need to know, yes?
and everyone wants to know.
Yves, the two ways are graph-name-is-URL-for-that-content and graph-name-denotes-that-content.
<yvesr> Guus, +999
guus: It'd be good to have
provenance example in primer.
... And weird to have a big feature of RDF that's not explained
in Primer.
<PatH> i think we should not have a provenance example in the primer. RDF as specified does not support provenance.
ivan: The Provenance example is usable only if you pick one of Antoine's semantics.
markus: What about Sandro's Direct Datasets. Can we use that?
ivan: That's what Antoine's document does.
pat: But Antoine's document doesn't ....
markus: Bring them inband.
sandro: ANyone can define the terms
pat: Antoine's spec clarifies several options.
<AZ> +1 to Pat here
pat: But not enough to really tell you how to use.
sandro: How about if the Primer says "We've provided syntax, but not enough semantics to make them useful, so stay tuned."
pat: Yes.
... So give an example that doesn't use metadata.
guus: But we have this
requirement of a Provenance Example.
... Someone mail me an example?
... We promised it to Prov WG
... They do the stuff in the nanopublications example.
... Hopefully we can decide on two notes next week; maybe PR
for syntax specs.
ADJOURNED
<Guus> trackbot, end meeting
This is scribe.perl Revision: 1.138 of Date: 2013-04-25 13:59:11 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/di/did/ Succeeded: s/deRoo/de Roo/ Succeeded: s/repled/replied/ Succeeded: s/an:/AZ:/ Succeeded: s/pfps/PatH/ Succeeded: s/PatH:/PatH,/ Found Scribe: sandro Inferring ScribeNick: sandro Default Present: AndyS, pfps, Guus_Schreiber, rdfwg, Sandro, AZ, David_Wood, Ivan, yvesr, TallTed, markus, gkellogg, pchampin, ericP, zwu2, [IPcaller], PatH, Arnaud Present: AndyS pfps Guus_Schreiber rdfwg Sandro AZ David_Wood Ivan yvesr TallTed markus gkellogg pchampin ericP zwu2 [IPcaller] PatH Arnaud Found Date: 04 Dec 2013 Guessing minutes URL: http://www.w3.org/2013/12/04-rdf-wg-minutes.html People with action items: guus WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]