Chatlog 2011-05-11

From RDF Working Group Wiki
Revision as of 16:24, 11 May 2011 by Lfeigenb (Talk | contribs)

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.

15:04:10 <LeeF> topic: Admin
15:04:12 <Souri> Souri has joined #rdf-wg
15:04:29 <LeeF> Guus: PROPOSED to accept the minutes of the 04 May telecon: http://www.w3.org/2011/rdf-wg/meeting/2011-05-04
15:04:38 <LeeF> RESOLVED: Accept the minutes of the 04 May telecon http://www.w3.org/2011/rdf-wg/meeting/2011-05-04
15:04:46 <LeeF> subtopic: action items
15:04:56 <LeeF> Guus: F2F poll has been setup
15:05:04 <PatHayes> I still have an action item, I think, but I can't find the details of what exactly it is.
15:05:16 <ericP> ericP has changed the topic to: RDF-WG site meeting - Code: 26631, Agenda: http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2011.05.11
15:05:17 <LeeF> close ACTION-41
15:05:18 <trackbot> ACTION-41 Set up poll about which site you'd use if we have a video link, pref murray hill vs cambridge/mit, and oct 4-5 vs oct-12-13. closed
15:05:20 <cygri_> cygri_ has joined #rdf-wg
15:05:25 <ericP> ericP has changed the topic to: RDF-WG weekly meeting - Code: 26631, Agenda: http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2011.05.11
15:05:43 <LeeF> Guus: Open action on cygri re: ISSUE-15 options
15:05:50 <LeeF> ... stays pending
15:06:22 <SteveH> Zakim, aaaa is [Garlik]
15:06:22 <Zakim> +[Garlik]; got it
15:06:32 <SteveH> Zakim, [Garlik] has SteveH and mischat 
15:06:32 <Zakim> +SteveH, mischat; got it
15:06:40 <LeeF> ACTION-21?
15:06:40 <trackbot> ACTION-21 -- Manu Sporny to create a doodle poll to find a time to have a call about RDF in JSON -- due 2011-03-23 -- CLOSED
15:06:40 <trackbot> http://www.w3.org/2011/rdf-wg/track/actions/21
15:06:45 <LeeF> ACTION-26?
15:06:46 <trackbot> ACTION-26 -- Patrick Hayes to write an description of action-21 -- due 2011-04-13 -- OPEN
15:06:46 <trackbot> http://www.w3.org/2011/rdf-wg/track/actions/26
15:06:47 <Zakim> +mhausenblas
15:06:58 <LeeF>  ACTION-26: actually about ISSUE-21, not ACTION-21
15:06:59 <cygri_> zakim, mhausenblas is temporarily me
15:06:59 <Zakim> +cygri_; got it
15:07:00 <trackbot> ACTION-26 Write an description of action-21 notes added
15:07:31 <LeeF> Guus: action on danbri continues until August
15:07:48 <LeeF> Guus: 3 actions regarding tools for spec authoring
15:08:04 <LeeF> gavinc: looked at it but haven't yet written it up
15:08:35 <PatHayes> Pat is puzzled. There does not appear to be an issue-21 listed.
15:08:57 <sandro> issue-21?
15:08:57 <trackbot> ISSUE-21 -- Can Node-IDs be shared between parts of a quad/multigraph format? -- open
15:08:57 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/21
15:09:02 <LeeF> subtopic: October F2F2
15:09:10 <LeeF> Guus: there's a new poll at http://www.w3.org/2002/09/wbs/46168/F2F2-EAST/
15:09:19 <LeeF> ... right now we've gotten 14 answers and there's a preference for the MIT location 
15:09:36 <LeeF> ... and a slight preference for 12-13 Oct
15:10:11 <mischat> I can make UK remote thing ... is what I tried to convey in the poll
15:10:34 <LeeF> Guus: we'll handle venues for F2F3 after deciding on F2F2
15:11:06 <LeeF> Guus: hope to make a decision on which tool to use for spec authoring next week
15:11:34 <LeeF> ACTION: Guus to look at spec authoring tools 
15:11:34 <trackbot> Created ACTION-46 - Look at spec authoring tools  [on Guus Schreiber - due 2011-05-18].
15:11:59 <LeeF> topic: ISSUE-12: Reconcile various forms of string literals
15:12:21 <AZ> zakim, unmute me
15:12:21 <Zakim> AZ should no longer be muted
15:12:22 <LeeF> See http://www.w3.org/2011/rdf-wg/track/issues/12 and http://lists.w3.org/Archives/Public/public-rdf-wg/2011May/0050.html
15:12:38 <LeeF> Guus: status?
15:12:47 <mischat> zakim, who is making noise ?
15:12:57 <Zakim> mischat, listening for 10 seconds I heard sound from the following: AZ (70%), Guus_Schreiber (35%)
15:13:00 <danbri> danbri has joined #rdf-wg
15:13:36 <LeeF> AZ: Not very concerned about the decision,  but reacted based on original decision that made the xsd:string URI archaic
15:14:22 <LeeF> AZ: Don't mind any kind of change to that proposal as long as it doesn't change the semantics of literals and not making xsd:string archaic because we still want to use xsd:string's in range restrictions of properties (e.g.)
15:14:31 <LeeF> q+ to point at Alex's comment
15:15:05 <Zakim> +OpenLink_Software
15:15:08 <gavinc> LeeF: Want to point at specific email from Allen. 
15:15:21 <MacTed> Zakim, OpenLink_Software is temporarily me
15:15:21 <Zakim> +MacTed; got it
15:15:23 <MacTed> Zakim, mute me
15:15:23 <Zakim> MacTed should now be muted
15:15:29 <AndyS> http://lists.w3.org/Archives/Public/public-rdf-wg/2011May/0090.html
15:15:34 <Zakim> + +31.20.598.aabb
15:15:41 <AlexHall> s/Allen/Alex
15:15:41 <gavinc> s/Allen/Alex
15:15:44 <LeeF> http://lists.w3.org/Archives/Public/public-rdf-wg/2011May/0090.html
15:16:35 <LeeF> AlexHall: basic thoughts are that as long as plain literal strings and xsd:string's are syntactically distinct, the software stack needs to be able to treat them as such
15:16:43 <LeeF> ... not a good idea to tell systems to silently convert from one to the other
15:17:05 <LeeF> ... we recognize they're semantically equivalent... a lot of discussion around SPARQL which is a syntactic query
15:17:20 <LeeF> ... discussion around whether it's the job of the RDF WG to address this issue, or for SPARQL WG
15:17:25 <LeeF> q-
15:17:33 <gavinc> +q RDF Interfaces
15:17:33 <ericP> q+ to say there's a value to having the semantic equivalence apparent in the graph
15:17:37 <LeeF> AlexHall: is the issue specific to SPARQL or wider than that?
15:17:40 <SteveH> 1 and "1"^^xsd:integer are syntactically different, but one gets transformed to the other, I don't see the difference�, except that historically "foo" and "foo"^^xsd:string were different, for crazy historical reasons
15:17:43 <gavinc> -q RDF, Interfaces
15:17:48 <PatHayes> q+
15:17:52 <gavinc> +q to talk about RDF Interfaces
15:17:57 <LeeF> LeeF: SteveH++
15:18:12 <ivan> ack ericP 
15:18:12 <Zakim> ericP, you wanted to say there's a value to having the semantic equivalence apparent in the graph
15:18:15 <LeeF> ericP: there's a fair value to having semantic equivalence apparent in the graph, and not just because of SPARQL
15:18:31 <Zakim> +NickH
15:18:41 <AlexHall> note, when i say "syntactic" i'm referring to the abstract syntax in RDF Concepts
15:18:54 <Guus> zakim, who is here?
15:18:54 <Zakim> On the phone I see FabGandon, AZ, Guus_Schreiber, Ivan, AndyS, gavinc, OlivierCorby, LeeF, mbrunati, Peter_Patel-Schneider, AlexHall, EricP, [Garlik], Souri, PatH, AxelPolleres,
15:18:57 <Zakim> ... cygri_, MacTed (muted), danbri, NickH
15:18:58 <LeeF> ericP: are there use cases that make us want to have both xsd:string and the plain literal in the same graph
15:18:59 <Zakim> [Garlik] has SteveH, mischat
15:19:01 <ivan> language tag
15:19:05 <LeeF> s/graph/graph?
15:19:21 <LeeF> ericP: if not, then the advice not to use one or the other is advice to parser specification authors
15:19:46 <LeeF> q?
15:20:01 <LeeF> PatHayes: Agree with ?Lee? that silent rewriting is a bad idea
15:20:15 <LeeF> ... I'd like to hear what Peter has to say about this issue
15:20:26 <LeeF> ... have a recollection that there was a strong case made to remove untyped literals of any kind
15:20:43 <LeeF> ... but pressure came from higher layers of the layer cake which motivated rdf:PlainLiteral
15:20:46 <ericP> why is silent rewriting bad? (or worse than having divergence of these representations?)
15:20:53 <LeeF> ... i'm puzzled as to why we're going in the other direction
15:21:08 <LeeF> ... this WG should seriously consider the arguments made previously that untyped literals should be deprecated
15:21:28 <LeeF> ... we do need to pay attention to the rdf:PlainLiteral typing idea
15:21:28 <ivan> +1 to Pat on rdf:PlainLiteral
15:21:30 <ericP> i thing that saying "use plain literals, but consider its type to be xsd:string" makes most folks happy
15:21:46 <AndyS> +1 to ericP - I'd like to understand what problems it causes
15:21:54 <LeeF> pfps: the problem with plain literals is they don't have a datatype and so it's hard to say that a property is restricted to plain literals
15:21:59 <ivan> andy, eric: language tag!
15:22:02 <AndyS> (the silent rewriting)
15:22:11 <ivan> q+
15:22:17 <ivan> ack PatHayes 
15:22:19 <gavinc> I thought they DO have a datatype?
15:22:21 <gavinc>  rdf:PlainLiteral?
15:22:28 <gavinc> http://www.w3.org/TR/rdf-plain-literal/#Definition_of_the_rdf:PlainLiteral_Datatype
15:22:29 <LeeF> pfps: recommendation in new OWL documents is that OWL processors should silently do the conversion
15:22:51 <PatHayes> Silent rewriting is bad because I certainly dont want ANYONE rewriting my RDF. My reasons for making it the way I make it might be private, but I dont want some other system second-guessing me.
15:22:54 <LeeF> ... of course in OWL everything is semantic, so it's not like it's changing anything as far as OWL is concerned
15:23:04 <ericP> ivan, can you describe a use case which reveals the language tag problem inherent in andy and my proposal?
15:23:19 <LeeF> pfps: recommendation was to use "foo" by itself over the wire, but internally consider it to be typed with rdf:PlainLiteral
15:23:19 <AndyS> I prefer lang tag and lang tagless behave similarly - more than xsd:string and untyped lang literals
15:23:47 <ivan> Eric, I want my name to properly written and flagged as Hungarian in a foaf file, and I cannot do that in xsd:string
15:24:23 <LeeF> pfps: for OWL, "foo"^^rdf:PlainLiteral is the same as "foo"^^xsd:string 
15:24:26 <LeeF> ivan: what about language tags?
15:24:30 <LeeF> pfps: no language tags here
15:24:32 <ericP> +1 to pfps's proposal
15:24:43 <PatHayes> Note, owl:sameAs, not 'same as'
15:24:54 <LeeF> gavinc: this comes up not just in parsing syntax, but in the recently published RDF interface working draft
15:25:18 <LeeF> ... when trying to implement it, you run into this problem, in that you have an expectation from programmers that native language strings get converted into _something_
15:25:30 <LeeF> ... it's very strange to try to figure out whether that should be xsd:string or rdf:PlainLiteral
15:25:40 <LeeF> ... no consensus in the APIs as to which one it actually does
15:25:44 <ivan> q+
15:25:48 <LeeF> ack gavinc
15:25:48 <Zakim> gavinc, you wanted to talk about RDF Interfaces
15:25:51 <ivan> ack gavinc 
15:26:00 <LeeF> gavinc: the current RDF interfaces WD points out that the RDF WG is working on this
15:26:05 <LeeF> ... so we need some sort of conclusion
15:26:21 <LeeF> ack ivan
15:26:46 <LeeF> ivan: the discussion around the interface is on the fact that at the moment, the RDF Concepts defines equality of 2 literals purely on lexical level
15:26:56 <AZ> AZ has joined #rdf-wg
15:27:06 <LeeF> ... so strings of unicode characters must be equal, plus datatypes equal, plus languages (if present) equal
15:27:10 <LeeF> ... no notion of conversion to any kind of value
15:27:15 <PatHayes> q+
15:27:15 <LeeF> ... so not sure if relevant at this point
15:27:46 <LeeF> ... long discussion on interface document, because for programmers it's not intuitive that when you have two strings that both stand for a number which has equal (mathetmatical) value, the two literals are still different 
15:27:56 <gavinc> Yes, "example" != "example"^^xsd:string 
15:28:38 <AZ> RDF semantics say they are equivalent with XSD entailmùent
15:28:40 <PatHayes> That is *syntactic* equality. But they denote the same value. No contradiction.
15:28:43 <ericP> not a prob if "example"^^xsd:string is silently converted to "example"
15:28:45 <LeeF> Guus: pfps said at face to face that they are the same
15:28:52 <gavinc> But Literal("example").valueOf == Literal("example"^^xsd:string).valueOf 
15:28:54 <LeeF> pfps: the issue is which level of entailment you want to live at
15:29:10 <LeeF> ivan: at core level there is no entailment, like in SPARQL
15:29:18 <LeeF> ericP: which is why SPARQL is the avatar for these problems
15:29:47 <LeeF> gavinc: when you use this in the interface, as soon as you use .valueOf() or use the API, they are _sometimes_ equal... which is strange!
15:30:01 <AndyS> SPARQL does not require all entailment : def for minimum.
15:30:03 <LeeF> ivan: true, but i'm sticking to the concepts there
15:30:33 <LeeF> Guus: seems clear we need to give more guidance, and we need a resolution that gives more guidance
15:30:41 <LeeF> ... marking as archaic doesn't seem to have concensus
15:30:46 <LeeF> PatHayes: i don't think there's a bug to repair here
15:30:54 <LeeF> ... we're just using the phrase "same as" in two different sense. 
15:30:59 <LeeF> ... in concepts we're talking about syntax
15:31:22 <ivan> q+
15:31:24 <LeeF> q+ to ask Steve to verbally mention his 1 vs. "1"^^xsd:integer analogy
15:31:26 <ivan> ack PatHayes 
15:31:34 <LeeF> Guus: people using this have trouble with the subtle difference
15:31:37 <LeeF> PatHayes: not very subtle
15:31:39 <LeeF> ack ivan
15:31:45 <gavinc> +q to sort of agree with PatHayes
15:31:47 <ericP> q+ to argue owl:sameAs will lead to cardinality challenges in SPARQL or rules with fresh variables in the head
15:32:08 <LeeF> ivan: the various tools around don't touch any sort of entailment by default, so that's what happens in SPARQL and RDF Interface
15:32:12 <LeeF> ... the bible stops at Concepts
15:32:18 <AlexHall> I think most people consider XSD-entailment too high a bar just to get string equivalence
15:32:24 <LeeF> ... anything about data type entailment is hidden in the cloud of the semantics document; it's rarely implemented
15:32:27 <cygri> +1 AlexHall
15:32:38 <tomayac> tomayac has joined #rdf-wg
15:32:41 <LeeF> ... is this a problem in the document?
15:32:44 <SteveH> +1 to AlexHall 
15:32:49 <Guus> zakim, who is here?
15:32:49 <Zakim> On the phone I see FabGandon, AZ, Guus_Schreiber, Ivan, AndyS, gavinc, OlivierCorby, LeeF, mbrunati, Peter_Patel-Schneider, AlexHall, EricP, [Garlik], Souri, PatH, AxelPolleres,
15:32:52 <Zakim> ... cygri_, MacTed (muted), danbri, NickH
15:32:53 <Zakim> [Garlik] has SteveH, mischat
15:33:02 <AndyS> D-entailment pulls in RDF and RDFS entailment - can we extract just "same value"?
15:33:03 <LeeF> PatHayes: is the problem that RDF is being used in a way that completely ignores its semantics?
15:33:09 <Souri> In practice, it is much easier to assume the kind of entailment that says "0010"^^xsd:integer = "10"^^xsd:integer ! Same could be true for "abc" and "abc"^^xsd:string!
15:33:24 <Zakim> +tomayac
15:33:33 <LeeF> PatHayes: is this just a problem of exposition?
15:33:37 <LeeF> Guus: essentially yes
15:33:53 <LeeF> ... we need to make very clear to the outside community that they should use the syntax in such a way that it doesn't give rise to the confusions that definitely exist
15:34:11 <Guus> q?
15:34:11 <LeeF> PatHayes: what confusions exist? we have 3 syntactic forms that are semantically equivalent. why don't we just say that that's what it is?
15:34:16 <LeeF> PatHayes: what is wanted?
15:34:17 <ericP> i think they want something *less*
15:34:20 <cygri> q+
15:35:30 <ivan> ack LeeF 
15:35:30 <Zakim> LeeF, you wanted to ask Steve to verbally mention his 1 vs. "1"^^xsd:integer analogy
15:36:02 <AndyS> q+ to talk about language tags
15:36:10 <SteveH> +1 to LeeF 
15:36:12 <ericP> +1
15:36:23 <ericP> (i think andyS is on board as well)
15:36:37 <LeeF> LeeF: we already have a precedent of different surface syntax mapping to the same abstract syntax with 1 and "1"^^xsd:integer
15:36:52 <ivan> q+
15:36:58 <LeeF> LeeF: why not do that with literals as well, so that both "foo" and "foo"^^xsd:string parse to the same abstract syntax term (such as "foo"^^xsd:string)
15:36:58 <AndyS> (nearly - seems to get strange for langs)
15:37:11 <LeeF> AndyS, I agree that langs makes it a little weird
15:37:27 <LeeF> gavinc: there is an issue where rdf term equality is defined in the Concepts document, that never gives a hint about semantic equivalence
15:37:28 <SteveH> I disagree, lang tags just means you have to do ""^^xsd:string -> ""
15:37:30 <AZ> +1, I think most people want XSD entailment
15:37:31 <LeeF> ack gavinc
15:37:31 <Zakim> gavinc, you wanted to sort of agree with PatHayes
15:37:43 <PatHayes> OK, how about this resolution. We take the rdf:PlainLiteral idea seriously, and (since we can change RDF) we say that engines SHOULD silently convert plain literals to typed literals with an explicit, syntactic type of rdf:plainLiteral.
15:38:07 <AZ> people want that "1.0"^^xsd:decimal = "1"^^xsd:decimal, among other things
15:38:19 <gavinc> +1 AZ
15:38:43 <PatHayes> AZ, that is true, they are equal. So they have what they want, now. 
15:38:53 <LeeF> ericP: when we say "same as" saying that xsd:string is same as untyped literal, that seems fine. if we do "owl:sameAs" then any system that preserves cardinality or does inference with fresh variables in the head will end up with different answers if we have "asdf" and "asdf"^^xsd:string being separate abstract syntax things but having rules that say you need to do owl:sameAs things
15:39:30 <LeeF> ericP: Lee's proposal seems to have some support - make one of them disappear from the abstract syntax
15:39:32 <AZ> PatHayes, they are equal only under D-entailment, where D includes {xsd:decimal}
15:39:46 <LeeF> ... the type of the "winner" is xsd:string
15:39:50 <LeeF> ack cygri
15:39:55 <LeeF> q- ericP
15:39:59 <LeeF> cygri: i see two options
15:40:01 <Guus> ack ericP
15:40:13 <LeeF> ... 1) what eric just said - make it so that there' sonly one option in the abstract syntax
15:40:19 <PatHayes> AZ, true. So, use that kind of entailment. Isnt this waht datatypes were invented for???
15:40:30 <LeeF> ... 2) leave the abstract syntax, and treat this as a usability problem of the specifications
15:40:50 <LeeF> ... as Pat noted, part of the problem is that to actually get value equality, you have to dive pretty deep
15:41:00 <PatHayes> The only-one-option option breaks because of language tagging, which cannot be got into xsd:string.
15:41:09 <LeeF> ... which gives you a lot of things orthogonal to the question of whether 1 == 1.0 and whether "a"^^xsd:string is the same as "a"
15:41:16 <AZ> PatHayes, I agree, we should tell people to use XSD entailment and stop caring about syntactic differences
15:41:30 <LeeF> ... maybe editorial work that can get literal equality stuff more in the foreground and decouple that from other datatype entailment stuff
15:42:23 <LeeF> AndyS: i like the idea of going to one datatype overall, but not necessarily comfortable with that being xsd:string
15:42:29 <ivan> +1 Andy
15:42:31 <SteveH> +1
15:42:37 <PatHayes> +1 AndyS
15:42:38 <LeeF> ... i think users expect that no lang tag and lang tag are closer together then no lang tag and xsd:string
15:42:48 <LeeF> ... you even see people that expect that without a lang tag match with a lang tag
15:42:59 <LeeF> ... e.g. see lots of questions online about querying dbpedia with a language tag
15:43:01 <Guus> +1 for pref of plain literal
15:43:13 <LeeF> ... so for user consistency would like to convert xsd:string to plain literals, no lang tag
15:43:16 <Souri> +1 to AndyS
15:43:16 <AndyS> ack me
15:43:16 <Zakim> AndyS, you wanted to talk about language tags
15:43:20 <LeeF> ... or create datatypes that reflect language tags
15:43:42 <LeeF> ericP: and datatype("asdf") is xsd:string?
15:43:44 <LeeF> AndyS: yes, in sparql
15:43:45 <PatHayes> We have a datatype which respects alnguage tags, it is rdf:PalinLIteral. ALl we need to do is to make this 'visible' in future RDF.
15:43:52 <PatHayes> PlainLIteral
15:44:04 <PatHayes> NOt PalinLIteral, aaaaargh
15:45:02 <cygri> q+
15:45:30 <LeeF> ivan: agree that xsd:string can't be the winner because of language tags
15:46:04 <LeeF> ivan: situation more complex because at the moment in the semantic documents everything for D-entailment is explicitly defined as an extension of RDFS entailment
15:46:09 <LeeF> ... so if we want to separate it, it's more of a change
15:46:30 <PatHayes> Ivan, very good point. I agree.
15:46:37 <gavinc> +1 Ivan, value equality of literals should not depend on RDFS 
15:46:43 <LeeF> q?
15:46:45 <LeeF> ack ivan
15:46:47 <LeeF> ack cygri
15:46:50 <PatHayes> We need to revise this 'layered' aspect of the sematnics in an y case.
15:46:54 <Guus> ack cygri
15:47:09 <LeeF> cygri: i don't know how much this layering of entailments matters
15:47:11 <AZ> Where is it said that D-entailment must be an extension of RDFS-entailment?
15:47:34 <PatHayes> The 'layering' was really just exposition, it is not a deep matter to do it more separated.
15:47:50 <LeeF> ... the mathematics of the situation shouldn't stop us from pointing out useful entailments 
15:47:53 <PatHayes> The entailments will change slightly, of course.
15:48:11 <ivan> AZ: "If D is a datatype map, a D-interpretation of a vocabulary V is any rdfs-interpretation I of V union {aaa: < aaa, x > in D for some x } ..."
15:48:15 <AndyS> AZ - the examples of D-ent use RDFS and the only test cases use RDFS at least
15:48:47 <LeeF> ivan: this is the only definition we have today
15:48:50 <PatHayes> NOt PURELY editorial, but it is do-able and I thinkwe should do it anyway.
15:48:51 <LeeF> ... and we need to live with / deal with that
15:48:55 <PatHayes> Ivan is right.
15:49:26 <PatHayes> IT is a technical change but its easy and I promise I will be able to do it. 
15:50:30 <PatHayes> LOL
15:51:04 <LeeF> Scribe: (general disagreement between ivan and cygri about how closely bound d-entailment is with rdfs entailment)
15:51:32 <ericP> ivan, cygri, i think cygri is saying we could factor existing entailment with text which invites e.g. SPARQL to say it works on a new "entailment1" which is graph entailment plus string entailment
15:51:55 <ericP> i prefer a stronger statement like "there never *was* any "abc"^^xsd:string"
15:52:04 <ivan> eric, I agree, that might be useful, but that is not an editorial change on the semantics document. That *all* I was saying...
15:52:28 <LeeF> Guus: would like to action the editors to consider this discussion and propose changes to concepts document
15:52:38 <SteveH> ericP, I'm not convinced that d(xsd:string) entailment doesn't make things worse
15:52:56 <ericP> ditto - i propose: 1. The form "abc"^^xsd:string is a deprecated form of "abc", which systems should silently convert. 2. The datatype of "abc" is xsd:string. 3. The datatype of "abc"@hu is rdf:PlainLiteral .
15:53:02 <PatHayes> +1 AZ
15:53:05 <AndyS> eric's q of a while ago - where do both xsd:string and simple literal occur together (on the web)?
15:53:07 <SteveH> ericP, yeah
15:53:42 <SteveH> ericP, except 3. is a bit odd, but maybe we have no choice there
15:53:48 <ivan> eric, and what is wrong to say that the datatype of "abc" is also rdf:PlainLiteral?
15:53:52 <ivan> why having the two?
15:54:12 <LeeF> Guus: strawpoll -- are changes to RDF Concepts abstract syntax needed?
15:54:21 <ericP> +1 to attacking this on the abstract syntax level
15:54:26 <LeeF> +1 
15:54:28 <SteveH> perhaps [question is too low levle]
15:54:30 <AlexHall> +1
15:54:33 <AndyS> I think lang+datatype will break code out there.  I'd be surprised if there wasn't assumption of one OR the other
15:54:34 <ivan> +1 if it works:-)
15:54:39 <AZ> -0.5
15:54:43 <PatHayes> Need to clarify if these are expositonal.ecitorial changes or changes to content. I cna t vote yet.
15:54:45 <SteveH> AndyS, I can promise you there is:)
15:54:45 <cygri> +-0
15:54:51 <LeeF> cygri: +-0
15:54:54 <pfps> +1
15:54:58 <Souri> +1
15:54:58 <cygri> PatHayes, changes to content
15:55:03 <mbrunati> +1
15:55:05 <PatHayes> Then +1
15:55:05 <gavinc> +0 
15:55:22 <LeeF> Guus: majority in favor, without details, i think this is something we should try to reach consensus around
15:55:36 <danbri> ∓0
15:55:36 <SteveH> q+
15:55:41 <LeeF> cygri: does anyone strongly believe that the abstract syntax should not be changed?
15:55:42 <AndyS> SteveH, yep - but I can change mime. It's people outside the WG ... who do data work 
15:56:30 <PatHayes> It is not enough. If there are changes to the anstract syntax, this will send ripples through everything. The semantics will need to be revised to fit.
15:56:58 <AZ> +1 ivan, make it an issue and we'll discuss it on the ML
15:57:14 <PatHayes> Im not afraid, but I want the earth to stop moving.
15:58:22 <LeeF> PatHayes: are we talking about a change to the content or the exposition?
15:58:24 <LeeF> ericP: content
15:58:24 <Souri> To understand it better: If someone submits two triples: <a> <name> "Dan" . <a> <name> "Dan"^^xsd:string . Should these be combined into just *one* triple?: <a> <name> "Dan" .
15:58:31 <LeeF> PatHayes: we ought to understand soon what that change _is_ then
15:58:49 <PatHayes> lol
15:59:11 <ericP> Souri, i say that's one triple
15:59:20 <SteveH> I'd /like/ it to be one triple
15:59:28 <ericP> i suspect that everyone that +1'd had that idea
15:59:28 <PatHayes> What does 'submit' mean???
15:59:31 <SteveH> but I have no clear idea following the discussion
15:59:47 <AndyS> This was my problem reading F2F minutes - need an answer (one triple preferably)
15:59:49 <AlexHall> Yes, one triple, with the abstract syntax of said triple still under discussion
16:00:06 <LeeF> +1 to One triple
16:00:06 <danbri> Can we couch this in terms of rdfcore style test cases? ie. what does some test data entail?
16:00:32 <AndyS> +1 to danbri
16:01:34 <ivan> <a> <b> "01234"^^xsd:integer . <a> <b> "1234"^^xsd:integer is another example for the same question, these are not only string issues
16:01:40 <LeeF> PatHayes: i understand if this is about surface form
16:01:53 <LeeF> ... i can't make sense if we're talking about making them the same "in the merge"
16:01:57 <LeeF> ericP: i think people are voting for the former
16:02:16 <Souri> +1 to Ivan's question
16:02:17 <LeeF> q+ to disagree with ivan's other example being the same
16:03:06 <cygri> These two are the same single triple: <a> <b> 1 . <a> <b> "1"^^xsd:decimal .
16:03:30 <AlexHall> string equality is so close to syntactic equality that it makes sense to approach that in the abstract syntax
16:03:32 <Zakim> +[Garlik.a]
16:03:32 <AZ> cygri, this is true in turtle because of syntactic sugar
16:03:34 <Zakim> -AZ
16:03:59 <Souri> In practice, value-based equality is what people expect for literal equality
16:04:01 <LeeF> q-
16:04:02 <Zakim> +AZ
16:04:12 <SteveH> q-
16:04:28 <LeeF> cygri: I can make a proposal
16:05:04 <ivan> ISSUE-40?
16:05:04 <trackbot> ISSUE-40 -- Skolemization advice in the RDF dcocument -- raised
16:05:04 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/40
16:05:07 <AndyS> XSD sort of separates xsd:integer and xsd:decimal and sort of doesn't  integer + integer => integer but same values and same type hierarchy
16:05:22 <LeeF> topic: ISSUE-40: Skelemization advice
16:05:31 <LeeF> Guus: it would be good to have discussion around ISSUE-40
16:05:36 <PatHayes> General question, how seriously attached to XSD should RDF be?
16:05:45 <LeeF> --> http://www.w3.org/2011/rdf-wg/track/issues/40 and http://www.w3.org/2011/rdf-wg/wiki/Skolemization
16:06:04 <LeeF> topic: Revisit RDF Postponed Issue
16:06:18 <LeeF> Guus: start with issue 51
16:06:20 <LeeF> ISSUE-51?
16:06:20 <trackbot> ISSUE-51 -- Revisit "Suggestion that the concept of context is missing from RDF" -- raised
16:06:20 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/51
16:06:24 <AndyS> PatHayes - I think it's good to pick a number system
16:06:50 <LeeF> PROPOSED: Close ISSUE-51 noting that it's a duplicate of work the WG is already doing with multiple graphs
16:06:58 <PatHayes> Andy, agree, but was just asking baout the general mood right now. 
16:07:15 <LeeF> +1
16:07:20 <Souri> +1
16:07:21 <AZ> +1
16:07:23 <pfps> +1
16:07:23 <cygri> +1
16:07:23 <mbrunati> +1
16:07:31 <LeeF> RESOLVED: Close ISSUE-51 noting that it's a duplicate of work the WG is already doing with multiple graphs
16:07:38 <LeeF>  ISSUE-51: RESOLVED: Close ISSUE-51 noting that it's a duplicate of work the WG is already doing with multiple graphs
16:07:39 <trackbot> ISSUE-51 Revisit "Suggestion that the concept of context is missing from RDF" notes added
16:07:41 <LeeF> close ISSUE-51
16:07:41 <trackbot> ISSUE-51 Revisit "Suggestion that the concept of context is missing from RDF" closed
16:07:43 <LeeF> ISSUE-52?
16:07:43 <trackbot> ISSUE-52 -- Revisit "How to indicate whether RDF embedded in another document is asserted" -- raised
16:07:43 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/52
16:08:01 <LeeF> PROPOSED: to resolve CLOSE - this is the responsibility of the enclosing document
16:08:08 <cygri> +1
16:08:10 <AndyS> +1
16:08:10 <danbri> (so this was a timbl thing, he wanted to say that RDF wasn't just a data model ...)
16:08:11 <SteveH> +1
16:08:12 <AZ> +1
16:08:14 <danbri> +1
16:08:44 <pfps> q+
16:09:07 <LeeF> ericP: trying to understand what the resolution means ... sometihng like N3, or some way to speak of "everything asserted in another document"?
16:09:30 <PatHayes> (real) named graphs provide a meachanism for this, but we dont have named graphs yet. Hey ho.
16:09:38 <pfps> no, this is like the RDF WG saying that RDF embedded in HTML is always/sometimes/never true
16:09:47 <AndyS> example - a graph diff has add triples and delete triples - delete triples not asserted
16:09:50 <LeeF> ack pfps
16:10:01 <LeeF> pfps: this is asking us whether RDF in an Adobe document (e.g.) should be asserted or not
16:10:06 <danbri> q+
16:10:07 <LeeF> ... and we have nothing to say about that
16:10:07 <AndyS> +1 to pfps
16:10:15 <Zakim> -AxelPolleres
16:10:16 <Zakim> -[Garlik.a]
16:10:40 <Zakim> -AndyS
16:10:45 <pfps> +1 to close
16:10:50 <mbrunati> +1 
16:10:54 <SteveH> +1
16:10:57 <LeeF> RESOLVED: to close ISSUE-52 - this is the responsibility of the enclosing document
16:11:02 <LeeF>  ISSUE-52: RESOLVED: to close ISSUE-52 - this is the responsibility of the enclosing document
16:11:03 <trackbot> ISSUE-52 Revisit "How to indicate whether RDF embedded in another document is asserted" notes added
16:11:04 <danbri> all we need to say is that RDF is descriptive, that it's the kind of stuff that can be interpreted as making claims about world. Other specs tell you when you've got some direct claims, vs quotes etc.
16:11:06 <LeeF> close ISSUE-52
16:11:06 <trackbot> ISSUE-52 Revisit "How to indicate whether RDF embedded in another document is asserted" closed
16:11:08 <LeeF> ISSUE-53?
16:11:08 <trackbot> ISSUE-53 -- Revisit "RDF is not just a data model; an RDF statement is an assertion" -- raised
16:11:08 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/53
16:11:11 <AxelPolleres> AxelPolleres has left #rdf-wg
16:11:42 <pfps> +1 to close 53 as indicated
16:11:53 <ericP> Guus: "RDF assertions can be used to make claims about the world" addresses this?
16:12:11 <ericP> danbri: timbl brought this up years ago
16:12:25 <ericP> ... he wanted to say that this could be used to talk about the world
16:12:35 <danbri> 'assertionable'
16:12:48 <LeeF> PROPOSED: Close ISSUE-53 with no change, and with the explanation that RDF semantics establishes that RDF statements can be used to make claims about the world. Figuring out who exactly is making those assertions is beyond the scope of the core technology. Some of these concerns may be addressed by the 'named graph' activity; others by W3C's new Provenance WG - http://www.w3.org/2011/prov/wik
16:12:48 <LeeF> i/Main_Page - and by other application vocabularies. See also http://www.w3.org/2005/Incubator/webid/charter for work linking cryptographically-assured notions of identity with RDF.
16:12:52 <AZ> it was originally raised by Dan Connolly apparently 
16:12:56 <Souri> +1
16:12:58 <ivan> +1
16:12:59 <SteveH> +1
16:13:02 <danbri> +1 (but i'm seconding my own text)
16:13:03 <AZ> +1
16:13:04 <mbrunati> +1
16:13:06 <cygri> tl;dr but +1
16:13:29 <LeeF> RESOLVED: Close ISSUE-53 with no change, and with the explanation that RDF semantics establishes that RDF statements can be used to make claims about the world. Figuring out who exactly is making those assertions is beyond the scope of the core technology. Some of these concerns may be addressed by the 'named graph' activity; others by W3C's new Provenance WG - http://www.w3.org/2011/prov/wik
16:13:29 <LeeF> i/Main_Page - and by other application vocabularies. See also http://www.w3.org/2005/Incubator/webid/charter for work linking cryptographically-assured notions of identity with RDF.
16:13:34 <LeeF>  ISSUE-53: RESOLVED: Close ISSUE-53 with no change, and with the explanation that RDF semantics establishes that RDF statements can be used to make claims about the world. Figuring out who exactly is making those assertions is beyond the scope of the core technology. Some of these concerns may be addressed by the 'named graph' activity; others by W3C's new Provenance WG - http://www.w3.org/201
16:13:34 <trackbot> ISSUE-53 Revisit "RDF is not just a data model; an RDF statement is an assertion" notes added
16:13:34 <LeeF> 1/prov/wiki/Main_Page - and by other application vocabularies. See also http://www.w3.org/2005/Incubator/webid/charter for work linking cryptographically-assured notions of identity with RDF.
16:13:37 <LeeF> close ISSUE-53
16:13:37 <trackbot> ISSUE-53 Revisit "RDF is not just a data model; an RDF statement is an assertion" closed
16:14:02 <pfps> +1 to not changing RDF/XML
16:14:02 <LeeF> PROPOSED: Close ISSUE-54 with no change, noting that this WG has no plans for substantially changing RDF/XML or the collection mechanism at this time.
16:14:06 <LeeF> +1
16:14:11 <ivan> ISSUE-54?
16:14:11 <trackbot> ISSUE-54 -- Revisit "RDF collection syntax should allow literals" -- raised
16:14:11 <trackbot> http://www.w3.org/2011/rdf-wg/track/issues/54
16:14:53 <cygri> +1 to proposal
16:14:56 <pfps> i changed the title
16:14:58 <LeeF> ivan: note that the title is misleading
16:15:10 <LeeF> RESOLVED: Close ISSUE-54 with no change, noting that this WG has no plans for substantially changing RDF/XML or the collection mechanism at this time.
16:15:15 <LeeF>  ISSUE-54: RESOLVED: Close ISSUE-54 with no change, noting that this WG has no plans for substantially changing RDF/XML or the collection mechanism at this time.
16:15:15 <trackbot> ISSUE-54 Revisit "RDF/XML collection syntax should allow literals" notes added
16:15:20 <LeeF> close ISSUE-54
16:15:20 <trackbot> ISSUE-54 Revisit "RDF/XML collection syntax should allow literals" closed
16:15:28 <LeeF> Guus: that's all, please fill in the F2F poll
16:15:29 <LeeF> ADJOURNED.
16:15:36 <mbrunati> bye
16:15:39 <AndyS> AndyS has left #rdf-wg
16:15:41 <Zakim> -mbrunati
16:15:42 <AZ> bye
16:15:43 <Zakim> -Souri
16:15:44 <Zakim> -FabGandon
16:15:45 <FabGandon> FabGandon has left #rdf-wg
16:15:45 <mbrunati> mbrunati has left #rdf-wg
16:15:46 <Zakim> -cygri_
16:15:46 <Zakim> -danbri
16:15:47 <Zakim> -PatH
16:15:47 <Zakim> -EricP
16:15:49 <NickH> bye!
16:15:50 <Zakim> -AlexHall
16:15:52 <Zakim> -OlivierCorby
16:15:54 <Zakim> -gavinc
16:15:54 <Zakim> -AZ
16:15:56 <Zakim> -Peter_Patel-Schneider
16:15:59 <Zakim> -NickH
16:16:03 <Zakim> -tomayac
16:16:04 <Zakim> -MacTed
16:16:45 <AlexHall> AlexHall has left #rdf-wg
16:16:56 <Zakim> -LeeF
16:16:58 <Zakim> -Guus_Schreiber
16:17:01 <LeeF> RRSAgent, make logs world
# SPECIAL MARKER FOR CHATSYNC.  DO NOT EDIT THIS LINE OR BELOW.  SRCLINESUSED=00000543