From RDF Working Group Wiki
Revision as of 16:11, 15 February 2012 by Eric
Please justify/explain non-obvious edits to this page, in your "edit summary" text.
15:56:36 <RRSAgent> RRSAgent has joined #rdf-wg 15:56:36 <RRSAgent> logging to http://www.w3.org/2012/02/08-rdf-wg-irc 15:56:38 <trackbot> RRSAgent, make logs world 15:56:40 <trackbot> Zakim, this will be 73394 15:56:40 <Zakim> ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 4 minutes 15:56:41 <trackbot> Meeting: RDF Working Group Teleconference 15:56:41 <trackbot> Date: 08 February 2012 15:57:08 <AndyS> zakim, this is 73394 15:57:08 <Zakim> ok, AndyS; that matches SW_RDFWG()11:00AM 15:58:55 <ivan> zakim, dial ivan-voip 15:58:55 <Zakim> ok, ivan; the call is being made 15:58:56 <Zakim> +Ivan 15:58:57 <swh> swh has joined #rdf-wg 15:59:36 <Zakim> +Guus 15:59:48 <gavinc> gavinc has joined #rdf-wg 15:59:49 <yvesr> Zakim, who is on the phone? 15:59:49 <Zakim> On the phone I see ??P5, Ivan, Guus 15:59:54 <yvesr> Zakim, ??P5 is me 15:59:54 <Zakim> +yvesr; got it 15:59:56 <Zakim> +Peter_Patel-Schneider 16:00:10 <Arnaud> Arnaud has joined #rdf-wg 16:00:12 <pfps> pfps has joined #rdf-wg 16:00:34 <Zakim> +OpenLink_Software 16:00:38 <Zakim> + +33.9.54.07.aaaa 16:00:40 <MacTed> Zakim, OpenLink_Software is temporarily me 16:00:40 <Zakim> +MacTed; got it 16:00:41 <MacTed> Zakim, mute me 16:00:41 <Zakim> MacTed should now be muted 16:00:43 <Zakim> +??P13 16:00:53 <AZ> zakim, +33.9.54.07.aaaa is me 16:00:53 <Zakim> +AZ; got it 16:00:54 <NickH> Zakim, ??P13 is me 16:00:54 <Zakim> +NickH; got it 16:00:59 <Zakim> + +1.707.861.aabb 16:00:59 <AZ> zakim, mute me 16:01:01 <Zakim> AZ should now be muted 16:01:01 <Zakim> + +1.408.996.aacc 16:01:02 <NickH> Zakim, mute me 16:01:03 <Zakim> NickH should now be muted 16:01:10 <gavinc> Zakim, aabb is me 16:01:10 <Zakim> +gavinc; got it 16:01:13 <AlexHall> AlexHall has joined #rdf-wg 16:01:20 <Zakim> + +1.443.212.aadd 16:01:31 <AlexHall> zakim, aadd is me 16:01:31 <Zakim> +AlexHall; got it 16:01:33 <Zakim> +Sandro 16:01:45 <sandro> sandro has changed the topic to: 8 Feb -- http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2012.02.08 16:01:47 <Guus> zakim, who is here? 16:01:48 <Zakim> On the phone I see yvesr, Ivan, Guus, Peter_Patel-Schneider, MacTed (muted), AZ (muted), NickH (muted), gavinc, +1.408.996.aacc, AlexHall, Sandro 16:01:55 <Zakim> On IRC I see AlexHall, pfps, Arnaud, gavinc, swh, RRSAgent, AZ, ScottB, Zakim, LeeF, AndyS, Guus, MacTed, mischat, ivan, danbri, SteveH, yvesr, davidwood, mdmdm, manu, trackbot, 16:02:02 <Zakim> ... manu1, NickH, sandro, ericP 16:02:04 <Zakim> +??P17 16:02:09 <swh> Zakim, ??P17 is me 16:02:10 <pchampin> pchampin has joined #rdf-wg 16:02:20 <Zakim> +swh; got it 16:02:24 <Zakim> +Tony 16:02:42 <ScottB> Zakim, Tony is me 16:02:42 <Zakim> +ScottB; got it 16:03:46 <Zakim> +??P24 16:03:51 <AndyS> zakim, ??P24 is me 16:03:51 <Zakim> +AndyS; got it 16:04:02 <Zakim> +davidwood 16:04:15 <Zakim> + +44.117.230.aaee 16:04:23 <danbri> zakim, +44.117.230.aaee is danbri 16:04:25 <Zakim> +danbri; got it 16:05:27 <danbri> (we've got a breather...) 16:05:43 <swh> scribe: swh 16:05:48 <swh> scribenick: swh 16:06:29 <pfps> minutes look fine 16:06:35 <swh> PROPOSED: accept minuites of last week 16:06:53 <zwu2> zwu2 has joined #rdf-wg 16:06:55 <swh> RESOLVED 16:06:57 <zwu2> zakim, code? 16:06:57 <Zakim> the conference code is 73394 (tel:+1.617.761.6200 sip:firstname.lastname@example.org), zwu2 16:07:07 <swh> Pending review items 16:07:14 <swh> close ACTION-136 16:07:14 <trackbot> Sorry... adding notes to ACTION-166 failed, please let sysreq know about it 16:07:29 <Zakim> + +1.650.265.aaff 16:07:35 <FabGandon> FabGandon has joined #rdf-wg 16:07:38 <zwu2> zakim, +1.650.265.aaff is me 16:07:38 <Zakim> +zwu2; got it 16:09:10 <Zakim> +[Sophia] 16:10:53 <Zakim> +EricP 16:11:47 <Zakim> -AndyS 16:12:15 <Zakim> +[IPcaller] 16:12:16 <danbri> (was I audible?) 16:12:21 <AndyS> zakim, IPCaller is me 16:12:22 <Zakim> +AndyS; got it 16:13:15 <Zakim> + +1.707.318.aagg 16:13:25 <cgreer> cgreer has joined #rdf-wg 16:13:55 <swh> One more review since last week 16:14:49 <swh> AlexHall: there's a lot of cleanup in XSD around defn's of lexical and value spaces, and mapping. RDF doesn't say anything about them. Just refers to them. No action needed. 16:14:59 <Zakim> +??P37 16:15:10 <swh> … one change that needs discussion is distinction between identity and equality 16:15:11 <sandro> zakim, who is talking? 16:15:14 <Guus> zakim, who is talking? 16:15:22 <Zakim> sandro, listening for 10 seconds I heard sound from the following: Guus (14%), AlexHall (19%), AndyS (13%) 16:15:33 <Zakim> Guus, listening for 10 seconds I heard sound from the following: Guus (9%), swh (5%), AlexHall (94%) 16:16:00 <swh> … has implications around entailment - for eg. +0 and -0 are distinct under XSD 1.1, but were equiv under 1.0 16:16:20 <swh> … NaN has implications for SPARQL, but not RDF 16:16:30 <Zakim> -AndyS 16:16:55 <swh> … we might need to write some text in the semantics document to make this clear 16:17:00 <Zakim> +??P24 16:17:04 <AndyS> zakim, P24 is me 16:17:04 <Zakim> sorry, AndyS, I do not recognize a party named 'P24' 16:17:11 <AndyS> zakim, ??P24 is me 16:17:11 <Zakim> +AndyS; got it 16:18:04 <swh> … should probably include duration as well as the datetime etc. datatypes in the types that are good for use with RDF 16:18:17 <swh> Guus: we should raise that as an issue 16:18:29 <swh> AlexHall: there's already an older issue 16:18:44 <AlexHall> http://www.w3.org/2011/rdf-wg/track/issues/66 16:19:46 <swh> Guus: should add that review to ISSUE-66 16:20:36 <swh> Turtle 16:20:56 <swh> Guus: things came up this week, from Ivan w.r.t. multiline comments 16:21:20 <swh> gavinc: read through it, not had time to write an email 16:21:35 <swh> … comment that we should do it like python is a problem because python doesn't 16:21:45 <swh> … have multiline comments 16:21:59 <Souri> Souri has joined #rdf-wg 16:22:00 <sandro> ? CSS has multiline comments. 16:22:05 <swh> … CSS only has single line comments, I don't see it as major problem 16:22:07 <sandro> ??? 16:22:16 <Zakim> +Souri 16:22:17 <Zakim> -zwu2 16:22:31 <swh> Danny Ayres had a comment 16:22:47 <danbri> re CSS .... /* I thought it \n did */ 16:22:47 <swh> gavinc: yes, CSS does have multiline comments 16:23:13 <swh> … we could do what python does, but seems like a large change 16:23:30 <swh> Guus: ask gavinc or ericP to respond to Danny 16:23:36 <swh> ericP: do we need a descision? 16:23:45 <sandro> http://www.w3.org/TR/CSS1/#comments "similar to those in the C programming language" 16:23:52 <swh> … we want to keep alignment with SPARQL etc. 16:23:54 <ivan> q+ 16:24:22 <gavinc> sandro, Yeah, I have have no idea where my brain was 16:24:32 <swh> ericP: I feel like I have lack of authority 16:24:48 <swh> Guus: might be best to raise an issue 16:24:50 <sandro> Yeah, I think this needs a WG resolution. 16:25:27 <AndyS> ack ivan 16:25:45 <swh> ivan: was a discussion danny raised on swig, not formally raised on this group, Ivan just drew groups attention to it 16:25:58 <Zakim> + +1.603.438.aahh 16:26:21 <zwu2> zakim, +1.603.438.aahh is me 16:26:21 <Zakim> +zwu2; got it 16:26:22 <swh> Guus: I think it would be in the spirit to regard this as a comment 16:26:42 <swh> … ericP and gavinc, please take an action 16:27:01 <swh> ACTION: ericP to repsond to multiline comments comment of Danny Ayres 16:27:01 <trackbot> Created ACTION-142 - Repsond to multiline comments comment of Danny Ayres [on Eric Prud'hommeaux - due 2012-02-15]. 16:27:22 <swh> Guus: 2nd issue, raised by Alex, on local name escapes 16:27:48 <swh> AlexHall: one issue is a typo… [noise] 16:28:18 <swh> … double \ was shown as introducing a char escape squence 16:28:30 <swh> … the other I was confused by appearance of % escape encoding in local part 16:28:44 <swh> … want clarification that they're not treated as escapes 16:29:02 <swh> … Andy confirmed for SPARQL, but not Turtle editors 16:29:37 <AlexHall> rdf:foo%20bar 16:29:38 <swh> ericP: you concern is whether I can say %68 and have it be equivalent to the unencoded version 16:29:51 <gavinc> on comments, Python still doesn't have multi line comments, nor Perl, and Ruby has really really funky multi line comment 16:29:59 <AndyS> f-o-o-%-2-0-b-a-r 16:30:16 <swh> AlexHall: if I include a % in the localname, is it equiv. to the unescaped version 16:30:23 <swh> ericP: [writing example] 16:30:31 <ericP> my:foob%61ar == my:foobar ? 16:31:16 <swh> ericP: I believe that the % has to stay in there - can't take %s out and have equivalence 16:31:19 <sandro> I think you mean: my:foob%61r -- my:foobar 16:31:52 <swh> AlexHall: I'm worried that people might thing it gets deescaped 16:32:23 <swh> ericP: ok, we need a bit of text saying you're not intended to unescape during processing 16:32:46 <swh> gavinc: possibly just reference the RFC doc 16:33:08 <Guus> q? 16:33:10 <swh> §5.3.1- simple string comparison 16:35:09 <ericP> ACTION: ericP to propose text to say that %nn is *NOT* unescaped while parsing Turtle 16:35:09 <trackbot> Created ACTION-143 - Propose text to say that %nn is *NOT* unescaped while parsing Turtle [on Eric Prud'hommeaux - due 2012-02-15]. 16:35:58 <swh> gavinc: do want to look at publishing a new draft of Turtle soon 16:36:03 <Zakim> - +1.707.318.aagg 16:36:09 <ivan> +1 to gavinc 16:36:12 <swh> … we have Turtle in HTML, grammar has changed, 16:36:49 <swh> ericP: were there any changes we made without consensus 16:37:01 <swh> Guus: may need formal review if WG wants 16:37:48 <swh> Topic: named graphs 16:37:59 <swh> Guus: want to talk a bit about exchanging data 16:38:14 <swh> … message from AndyS 16:38:32 <AndyS> I don't think we have consensus 16:38:37 <swh> Guus: AndyS, do you think we reached consensus 16:39:09 <swh> AndyS: I brought it up because it was a priority before 16:39:29 <swh> … I think Pat and I are agreed about whether it would be best if we published the smenatics of the 4th column 16:39:42 <swh> … I think we'd like to see it published, but need back compat 16:40:26 <swh> AndyS: if you look at dbpedia they have a 4th col, but I don't know what it means, they're not complying to a published method, but it is a usecase 16:40:53 <swh> Guus: so, you can't use the 4th column as an IRI to fetch the triples 16:40:58 <swh> AndyS: no 16:41:04 <swh> … it's got a hash on the end 16:41:11 <swh> … it's quite profile, and it exists 16:41:11 <MacTed> Zakim, unmute me 16:41:11 <Zakim> MacTed should no longer be muted 16:41:21 <Guus> q? 16:42:11 <swh> Guus: what is the relationship between the 4th col and the triple? 16:42:19 <swh> MacTed: I don't know, trying to get answer 16:43:05 <swh> ACTION: MacTed to investigate what the relationship is, and document it 16:43:05 <trackbot> Created ACTION-144 - Investigate what the relationship is, and document it [on Ted Thibodeau - due 2012-02-15]. 16:43:17 <swh> AndyS: the consequence is that quads are not just an internal issues 16:43:24 <swh> ivan: I don't understant 16:43:34 <swh> AndyS: I phrased the usecase as being about TriG 16:43:53 <swh> … it's from the extraction project, not the running service 16:46:08 <swh> ACTION-144: relationship between 4th col and triple 16:46:08 <trackbot> ACTION-144 Investigate what the relationship is, and document it notes added 16:46:26 <swh> Guus: usecase discussion between sandro and AndyS 16:46:43 <davidwood> q+ to ask about the details of the use case 16:46:53 <swh> … to illustrate how you go about identifying time-varying gboxes 16:47:01 <davidwood> q- 16:47:07 <swh> … AndyS, sandro, do you think your desgins are the same 16:47:23 <swh> sandro: we're talking about the same pattern 16:47:38 <swh> Guus: that makes it worth exploring in more detail 16:47:50 <swh> … in general I'l like to explore more solution designs and apply to usecase 16:48:07 <Guus> http://www.w3.org/2011/rdf-wg/wiki/TriG-REST 16:48:40 <swh> Guus: we started exploring sulution designs 16:48:50 <swh> … try to give a natural language explanation of what it means 16:48:58 <swh> … so non RDF geeks can understand 16:49:04 <swh> … started writing down examples 16:49:17 <swh> … my question to WG is, is this kind of approach useful? 16:49:17 <Zakim> -AZ 16:49:33 <AndyS> Useful to explore this pattern 16:49:58 <swh> … in the 3rd usecase sandro has a statement about static graph container, those appear in AndyS's solution too 16:50:07 <swh> … do we want to define that, and if so, in what namespace 16:50:12 <ivan> q+ 16:50:17 <swh> no moee namespaces please 16:50:23 <sandro> +1 16:50:26 <swh> Guus: is it useful to look at these designs? 16:50:36 <AndyS> q? 16:50:37 <sandro> +1 on looking at solutions in more detail 16:50:41 <swh> +1 16:50:59 <pchampin> +1 16:51:01 <Zakim> +AZ 16:51:02 <swh> ivan: we certainly have to move on and look at possibilities 16:51:04 <AZ> zakim, mute me 16:51:04 <Zakim> AZ should now be muted 16:51:31 <swh> … what I don't fully understand is that the usecase means the semantics of GET, but does it mean that you would have other semantics? 16:51:49 <swh> Guus: how do we do the page structuring? I thought one page per solution 16:52:07 <swh> … tried to come up with nat lang description of TriG RESR 16:52:14 <swh> s/RESR/REST/ 16:52:30 <MacTed> AndyS - do you have a sample fourth-column value from those DBpedia dumps? 16:52:33 <swh> sandro: at a tutorial level the text description is ok 16:52:45 <swh> ivan: I would prefer to see all on one page 16:52:50 <swh> Guus: no problem 16:52:54 <MacTed> (even better, a full row or two from any of those N-quad files) 16:52:55 <sandro> http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs-Designs 16:53:02 <AndyS> MacTed - the load page has samples IIRC 16:53:12 <swh> sandro: I was doing the same thing 16:53:28 <AndyS> ... not sure if they are representative enough. 16:54:00 <swh> sandro: my impression right now is that most of the group is not following this closely enough to make descisions about it 16:54:26 <swh> Guus: I prefer to see people owning solutions 16:54:47 <swh> sandro: don't want to get too emotionally involved 16:55:07 <ericP> scribenick: ericP 16:55:08 <Zakim> -swh 16:55:28 <ericP> Guus: happy to write more examples for this solution 16:56:13 <ericP> … i'll try to stay impartial in my defense of this solution 16:56:36 <sandro> http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs-Designs 16:56:38 <ericP> … and there's the equality use case? 16:57:15 <ericP> sandro: in TF-Graphs-Designs, the first is "trig state" (earlier called "trig rest") (subject of guus's page) 16:57:26 <ericP> … second is "trig equality" 16:57:47 <ericP> Guus: where the lable is a placeholder for the set of triples, instead of pointing to it 16:59:05 <ericP> sandro: 3rd is the n3 style of explicitly naming the relation (scroll down to "graph object") 17:00:48 <ericP> ... "Graph Objects" is triples, where nodes can be graphs 17:01:09 <zwu2> q+ 17:01:19 <ivan> q- 17:01:22 <ericP> Guus: could be seen at quints, e.g.: 17:01:23 <danbri> (regarding 'bigger than RDF', I always think of RDF as http://en.wikipedia.org/wiki/Flatland ... and quads as adding an extra dimension...) 17:01:44 <ericP> … eg:s1 eg:p1 eg:o1. eg:g1 rdf:graphState 17:01:52 <yvesr> it is possible to store such things in a quad store (that's how my SWI-Prolog N3 implementation works) 17:02:09 <ericP> sandro: steve argued that there's a problematic computational overhead to this approach 17:03:06 <ericP> … distinction between Graph Objects and Graph Datatypes is that former has graphs in the object position and latter has turtle literals 17:04:41 <sandro> the type of the thing identified by the label. 17:04:41 <pchampin> re 5., I would be more comfortable with something more keywordish, like "@graphLabelRelation rdf:graphState" 17:04:47 <ericP> AndyS: [re: Relation Flag"] i was imagining the the modifier applied to each triple, e.g. eg:s eg:p eg:o. eg:g 17:06:21 <ericP> Guus: in AndyS's Oct mail, he used types like StaticGraphContainer 17:06:38 <ericP> … these are close to Graph Objects 17:07:11 <ericP> sandro: with trig state, there are precise semantics which i think are distinguishable [from andy's mail] 17:07:50 <AndyS> q? 17:07:53 <Guus> q? 17:07:55 <ericP> … calling Andy's proposal "typed @@1" 17:08:28 <ericP> zwu: with graphs as objects, how could you enforce equality? 17:08:57 <ericP> ericP: i.e. have the same extension? 17:09:12 <ericP> zwu: right, owl:sameAs has powerful semantics 17:09:40 <ericP> sandro: hasn't come up in the use cases, but you could say 17:10:09 <AndyS> """The built-in OWL property owl:sameAs links an individual to an individual.""" 17:10:12 <ericP> … ^h^h^h utter some inconsistency which was computationally hard to catch 17:10:28 <Souri> +1 to calling it something else 17:10:53 <sandro> ivan: later, we can say sameas is sameas sameas 17:10:57 <pchampin> sameGraph ? 17:10:58 <ericP> … so if we use owl:sameAs, are we bringing in baggage? happy to use something else 17:11:06 <ericP> … sameGraphAs 17:11:23 <MacTed> -1000 sameGraphAs 17:11:30 <ericP> zwu: triple order and bnodes make it hard 17:11:32 <MacTed> sameGbox maybe 17:11:32 <AndyS> It is expensive to test for equality. See JJC paper. 17:12:00 <MacTed> (implying also sameGsnap, sameGtext) 17:12:06 <pchampin> @MacTed no, as I get it, the idea is to identify the g-snap here, to the g-box 17:12:44 <gavinc> Please see http://en.wikipedia.org/wiki/Graph_isomorphism 17:12:49 <sandro> MacTed, how about rdf:isGSnap ? 17:13:01 <ericP> sandro: i think we can factor out that computation because it's not necessary for the use cases 17:14:29 <ericP> [general exceptance of the term "GSnap"] 17:14:55 <ericP> Guus: would like to execute the provenance scenario with these different designs 17:15:12 <ericP> Arnaud: have we defined GSnap et al? 17:15:28 <sandro> http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs-Designs#Graph_Objects edited to use rdf:isGSnap 17:15:32 <ericP> sandro: plan is that these terms won't make it into the final specs 17:15:59 <ericP> ACTION: Guus to merge his page with Sandro's 17:15:59 <trackbot> Created ACTION-145 - Merge his page with Sandro's [on Guus Schreiber - due 2012-02-15]. 17:17:03 <ericP> ACTION: Guus to write down the provenance scenario example, as well as those in AndyS's and Steve's email, and the one from DBPedia 17:17:03 <trackbot> Created ACTION-146 - Write down the provenance scenario example, as well as those in AndyS's and Steve's email, and the one from DBPedia [on Guus Schreiber - due 2012-02-15]. 17:17:19 <Zakim> -Peter_Patel-Schneider 17:17:55 <ericP> Guus: let's start with the three in sandro's message, plus the one in Andy's message and the keeping-inference-separate examples 17:18:19 <AZ> bye 17:18:22 <Zakim> -AZ 17:18:23 <zwu2> bye 17:18:27 <danbri> bye! 17:18:27 <Zakim> -zwu2 17:18:29 <Zakim> -Ivan 17:18:37 <MacTed> s/general exceptance/general acceptance/ 17:18:38 <Zakim> -Souri 17:18:41 <Zakim> -yvesr 17:18:43 <Zakim> -NickH 17:18:47 <Zakim> -AlexHall 17:18:48 <Zakim> -gavinc 17:18:52 <Zakim> -ScottB 17:19:03 <danbri> http://www.w3.org/2009/CommonScribe/ ? # SPECIAL MARKER FOR CHATSYNC. DO NOT EDIT THIS LINE OR BELOW. SRCLINESUSED=00000318