14:51:08 RRSAgent has joined #rdf-wg 14:51:08 logging to http://www.w3.org/2011/03/30-rdf-wg-irc 14:51:23 rrsagent, make records public 14:51:44 Scribe: gavinc 14:53:21 Zakim has joined #rdf-wg 14:53:44 hsbauer has joined #rdf-wg 14:53:54 zakim, this is rdfwg 14:53:54 ok, gavinc; that matches SW_RDFWG()11:00AM 14:54:46 73394 14:55:27 SteveH__ has joined #rdf-wg 14:55:56 +??P19 14:56:02 mbrunati has joined #rdf-wg 14:56:04 +Tony 14:56:07 Zakim, +??P19 is yvesr 14:56:07 sorry, yvesr, I do not recognize a party named '+??P19' 14:56:13 Zakim, P19 is yvesr 14:56:13 sorry, yvesr, I do not recognize a party named 'P19' 14:56:32 +??P21 14:56:36 Zakim, ??P19 is yvesr 14:56:36 +yvesr; got it 14:56:39 Zakim, ??P21 is me 14:56:39 +SteveH__; got it 14:57:01 zakim, who is here 14:57:01 gavinc, you need to end that query with '?' 14:57:03 zakim, who is here? 14:57:03 On the phone I see +1.707.861.aaaa, yvesr, Tony, SteveH__ 14:57:14 zakim, +1.707 is me 14:57:14 +gavinc; got it 14:57:17 +David 14:57:19 Zakim, mute me 14:57:19 yvesr should now be muted 14:57:46 +[IPcaller] 14:58:17 zakim, [IPCaller] is me 14:58:17 +mbrunati; got it 14:58:21 AxelPolleres has joined #rdf-wg 14:59:10 zwu2 has joined #rdf-wg 14:59:17 + +20598aabb 14:59:18 +AxelPolleres 14:59:34 +OlivierCorby 14:59:38 +Peter_Patel-Schneider 14:59:40 AndyS has joined #rdf-wg 14:59:50 ericP has joined #rdf-wg 14:59:54 pfps has joined #rdf-wg 14:59:58 +zwu2 15:00:03 +EricP 15:00:11 zakim, mute me 15:00:11 zwu2 should now be muted 15:00:13 zakim, what is the conference code? 15:00:13 the conference code is 73394 (tel:+1.617.761.6200 tel:+33.4.26.46.79.03 tel:+44.203.318.0479), AndyS 15:00:26 zakim, thank you 15:00:26 you are very welcome, AndyS 15:00:41 AlexHall has joined #rdf-wg 15:00:47 Chair: Guus Schreiber 15:00:51 AZ has joined #rdf-wg 15:00:54 +Sandro 15:01:05 +[IPcaller] 15:01:05 cmatheus has joined #rdf-wg 15:01:14 zakim, IPCaller is me 15:01:14 +AndyS; got it 15:01:20 + +1.443.212.aacc 15:01:24 zakim, who is here? 15:01:26 On the phone I see gavinc, yvesr (muted), Tony, SteveH__, David, mbrunati, +20598aabb, AxelPolleres, OlivierCorby, Peter_Patel-Schneider, zwu2 (muted), EricP, Sandro, AndyS, 15:01:31 ... +1.443.212.aacc 15:01:45 zakim, mute me 15:01:45 gavinc should now be muted 15:01:49 zakim, +1.443.212.aacc is me 15:01:49 +AlexHall; got it 15:02:06 zakim, unmute me 15:02:06 gavinc should no longer be muted 15:02:20 zakim, who is on the call? 15:02:20 On the phone I see gavinc, yvesr (muted), Tony, SteveH__, David, mbrunati, +20598aabb, AxelPolleres, OlivierCorby, Peter_Patel-Schneider, zwu2 (muted), EricP, Sandro, AndyS, 15:02:24 ... AlexHall 15:02:40 Topic: Admin 15:03:11 +Ronald 15:03:17 zakim, who is making noise? 15:03:28 sandro, listening for 10 seconds I heard sound from the following: +20598aabb (22%) 15:03:31 Zakim, I think Ronald is me 15:03:31 I don't understand 'I think Ronald is me', AZ 15:03:42 Zakim, Ronald is me 15:03:42 +AZ; got it 15:03:47 zakim, who is making noise? 15:03:58 sandro, listening for 10 seconds I could not identify any sounds 15:04:02 zakim, who is making noise? 15:04:07 subway? 15:04:12 sandro, listening for 10 seconds I heard sound from the following: +20598aabb (28%) 15:04:14 +??P5 15:04:17 +??P8 15:04:18 zakim, mute aabb 15:04:18 +20598aabb should now be muted 15:04:25 :) 15:04:27 zakim, mute me 15:04:27 AZ should now be muted 15:04:27 zakim, unmute aabb 15:04:29 +20598aabb should no longer be muted 15:04:40 AZ, you were making the noise. 15:04:46 Zakim, i am ??P8 15:04:46 +webr3; got it 15:04:53 +Souri_ 15:05:11 Souri has joined #rdf-wg 15:05:28 Guus: Proposing accepting minutes, no objections, accepted. 15:05:35 Sorry looks like i was not marked present. 15:05:41 Minutes OK, but we should figure out who Tony is. 15:06:16 My name didn't make the list 15:06:28 zakim, who is here 15:06:28 gavinc, you need to end that query with '?' 15:06:30 zakim, who is here? 15:06:30 On the phone I see gavinc, yvesr (muted), Tony, SteveH__, David, mbrunati, +20598aabb, AxelPolleres, OlivierCorby, Peter_Patel-Schneider, zwu2 (muted), EricP, Sandro, AndyS, 15:06:36 ... AlexHall, AZ (muted), ??P5, webr3, Souri_ 15:07:02 Subtopic: Actions 15:07:15 thomas sent regrets too 15:07:28 Guus: Most people with actions are not here. 15:07:57 Guus: Richard did Nathans action, group JOSN use cases on wiki 15:08:00 close action-22 15:08:01 ACTION-22 Group json use cases on wiki closed 15:08:13 + +1.850.377.aadd 15:08:25 Subtopic: F2F1 15:08:40 regrets for F2F 15:08:44 Guus: Has everyone confirmed attendance? 15:09:05 Zakim, i am ??P5 15:09:05 +cmatheus; got it 15:09:12 Guus: Some people still haven't RSVP'd 15:09:21 pathayes has joined #rdf-wg 15:09:26 Guus: Remote will be hard with timezones. 15:09:34 well, I'll remotely attend the audio portion, if we have a tel# 15:09:44 Topic: Turtle Task Force 15:10:00 q+ 15:10:01 apologies for 10 mins late. 15:10:20 http://www.w3.org/2011/rdf-wg/track/products/3 15:10:26 Guus: Only two open issues, likely need more 15:10:44 Guus: ??? had a thread on Turtle issues. 15:10:57 s/???/Peter and Eric/ 15:11:25 Eric: We need a mapping from Turtle to Triples 15:11:57 Sandro: Are we going to produce such a mapping? 15:12:02 s/Eric/Peter/ 15:12:02 -> http://www.w3.org/2010/01/Turtle/#sec-parsing Editor's Draft to which pfps referred 15:12:17 -> http://lists.w3.org/Archives/Public/semantic-web/2008Jan/0128 15:12:35 s/Are we/Maybe the ISSUE is: Are we/ 15:13:17 Eric: Recursive or linear definition? How does concatenation work? 15:13:24 q- 15:14:49 sandro has changed the topic to: Agenda: http://lists.w3.org/Archives/Public/public-rdf-wg/2011Mar/0693.html 15:15:05 sandro has changed the topic to: Agenda: http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2011.03.30 15:15:09 Guus: Can see a perpose for both of them 15:15:11 Guus: Can see a purpose for both of them. Might mean adding an appendix or section to current document. 15:15:12 Eric: Was going to code up both, but sprained wrist 15:15:14 Guus: Is that do able? 15:15:15 Peter: I am only interested in having a good spec. I have my own ideas on what is best, but think having more then one way is worse. 15:15:17 I don't want to have two ways to turn Turtle to triples. 15:15:35 Eric: It is somewhat distracting to have two 15:15:40 Peter: Agree. 15:16:09 Guus: Is there a way to work this out? 15:16:15 q? 15:16:28 Eric: I don't think either of us think the other is wrong. 15:16:33 Peter: Just misguided 15:17:02 Pat: Are we sure they are different? 15:17:39 ???: Sounds like we have agreement, accept for agreement 15:17:45 s/???/Pat/ 15:18:00 general note: pfps' approach is perhaps more similar to other well used specs (html and webapps ones) and thus may be more familiar to readers 15:19:49 Peter: A good spec doesn't have different interpretations. A recipe often does. 15:20:25 Eric: My main objection is to the recursive method vs the linner method 15:20:42 andy: A good set of test cases will help 15:20:49 +1 15:20:59 s/linner/linear 15:21:22 AndyS: What happens around BNodes, what happens around escaping 15:21:24 andy: layering -- escaping, identifying terminals, etc 15:22:22 q+ 15:22:34 q- 15:22:40 AndyS: Deifine how each part is converted into each bit of a triple so that it can be reused (??) 15:23:44 EricP: Using this grammer in TriG. Notion of reusing the grammer was part of the reason for giving it a linear parsing model. So that it could be reused in cases where there is a document after. Stole what I could from SPARQL 15:24:04 EricP: Reuse of shared understanding from one spec or another 15:24:24 +q 15:24:40 andy: Yes, sharing between SPARQL and Turtle spec is good. 15:24:45 -> http://lists.w3.org/Archives/Public/semantic-web/2008Jan/0128 this one 15:25:01 +1 15:25:08 Sandro: Reusing and sharing grammer? 15:25:12 -> http://lists.w3.org/Archives/Public/semantic-web/2008Jan/0128 pfps's mapping from Turtle grammar to an RDF graph 15:25:15 Peter: There is no grammer in mine. 15:25:29 Peter: It takes the AST, and turns the AST into triples 15:25:50 Peter: Takes AST, silent to sequences, silent to Turtle doc 15:26:30 Sandro: Add Peters spec to Turtle document? 15:26:37 s/Sandro/Guus 15:26:51 Peter: Silly to have a recipe and then a spec 15:26:59 Eric: Both are recipes 15:27:12 Peter: Silly have a ??? followed by a ??? 15:27:46 s/???/procedural spec/ s/???/non-procedural spec/ 15:27:48 eric: Who has looked at both of these? 15:28:06 gavin: I did. 15:28:12 andy: As did I 15:28:24 gavin: I got stuck on how whitespace is to be handled in Turtle. 15:28:39 gavin: Got it mostly working, but I have no idea if this is correct. 15:28:52 10. 15:28:57 sandro, thank you 15:29:02 q? 15:29:06 -q 15:29:27 My parser does tokenization then parsing -- and it's streaming so working on a unit of same-subject makes sense. 15:29:27 ISSUE: How do we parse "18." in Turtle? 15:29:28 Created ISSUE-18 - How do we parse "18." in Turtle? ; please complete additional details at http://www.w3.org/2011/rdf-wg/track/issues/18/edit . 15:29:50 ah - i thougt i raised that one but must have forgot 15:30:16 http://lists.w3.org/Archives/Public/public-rdf-wg/2011Feb/0131.html 15:30:38 Guus: Peter do you mind being on Turtle TF? 15:30:47 Peter: I like Graphs more then Turtle ;) 15:31:28 i can edit the issue 15:31:40 Guus: Two other issues with Turtle 15:31:47 (ISSUE-18) 15:32:20 Eric: I think all the terminals have an s at the end of them are supposed to be defined in the sparql spec. 15:32:35 Eric: Lexical tokens are all the same as SPARQL 15:32:41 eric: re ISSUE-1 I think all the terminals have a link back to the SPARQL spec, except BASE and PREFIX. A few of the lower level productions are different, but the lack of variables makes it difference. 15:32:42 close but not the same e.g. [ :p 123 ] . 15:33:07 close but not the same e.g. [ ns:property 123 ] . (due to Ivan) 15:33:08 INSER DATA { } doesn't allow variables does it? 15:33:10 ISSUE: Should TURTLE allow triples like "[ :p 123 ]." as SPARQL does ? 15:33:10 Created ISSUE-19 - Should TURTLE allow triples like "[ :p 123 ]." as SPARQL does ? ; please complete additional details at http://www.w3.org/2011/rdf-wg/track/issues/19/edit . 15:33:12 *INSERT 15:33:44 Same with "(1 2 3) ." -- uncommon 15:33:49 -> http://www.w3.org/2010/01/Turtle/#sec-grammar-grammar 15:34:16 the diagram from n3 is quite good [related] http://www.w3.org/DesignIssues/diagrams/n3/venn 15:34:39 s/INSER DATA/INSERT DATA/ 15:34:57 -> http://www.w3.org/2010/01/Turtle/#sec-diff-sparql 15:35:01 Sandro: How are the languages related? How do I get from one to the other? 15:35:16 sandro: I want a complete and precise listing in this spec of how it differs from SPARQL. (base, prefix, variables, maybe terms - as - triples) 15:35:39 SPARQL grammar cheats - just says "don't use variables" -- not part of formal grammar (grammar explosion) 15:36:28 sandro: just the bit of SPARQL that INSERT { ... } 15:37:06 sandro: it's smaller than BGP, it's just the INSERT part, not the WHERE part. 15:37:32 INSERT DATA can include GRAPH :-) 15:39:11 AndyS, true 15:39:28 CONSTRUCT template then? 15:39:29 sandro: why don't 6-16 have links to SPARQL? 15:39:40 eric: they are very similar, but different eg in allowing variables. 15:39:57 Eric: http://www.w3.org/2010/01/Turtle/#sec-grammer 4.4, numbers with s at the end are the same as SPARQL. 6-16 don't match as Turtle doesn't have variables 15:41:12 peter: The SPARQL spec does the standard cheat of saying "you need whitespace where you need whitespace". ... a b 18. a b ... 15:41:33 eric: greedy lexer consumes the dot, simpler solution. 15:41:34 agree 15:41:38 sandro: the turtle spec does the same... 15:41:50 s/:/,/ 15:41:50 "White space (production ws) is used to separate two tokens which would otherwise be (mis-)recognized as one token" 15:42:05 yes 15:42:09 I hate that. 15:42:10 q+ 15:42:35 gavin: i have an issue on that. I can parse everything, unless it's a double. I can't figure out how to parse what you just said. 15:43:08 gavin: parser had to comment out the bit with numbers with dots in them. 15:43:15 double has an 'e' 15:43:22 peter: typically, you do a greedy lexer, doing doubles and decimals. 15:43:40 peter: just about every programming language has this. 15:44:19 https://github.com/norcalrdf/pymantic/blob/master/pymantic/parsers.py 15:44:20 AndyS: re LALR(1) parser. I think we want to go for LL(1), actually, so it can be done via Recursive Descent. Still need the right tokenizer. 15:44:47 ericP: I haven't noticed the difference manifesting for this. 15:45:24 q+ 15:45:28 q- 15:46:00 q+ 15:46:10 Sandro: Speaking as user. Not a grammar geek. Don't want to have to put a space before the period. I don't care if makes the parser harder. 15:46:53 peter: you only need to space after a number. 15:47:01 sandro: so code puts a space there 15:47:16 AndyS: SPARQL allows for 18. to represent a number, the charter would not allow us to change that 15:47:24 andy: SPARQL 1.0 for better or worse requires the space after a number, and SPARQL 1.1 by its charter can't change that. 15:47:36 andy: I'm not saying it's a good decision, but that's where we are. 15:47:56 andy: Prefix names do not end in a . 15:47:59 AndyS: Prefix names in SPARQL do not end with a . 15:48:01 ack AndyS 15:48:11 ack sandro 15:48:21 Subtopic: Turtle and Triples 15:48:57 q+ 15:48:59 Guus: Deprecate N-Triples and just use Turtle? N-Triples developed for test cases. 15:49:21 +1 to deprecating ntriples 15:49:31 peter: Yes, get rid of N-Triples! 15:49:51 q+ 15:49:58 zakim, unmute me 15:49:58 zwu2 should no longer be muted 15:50:18 AndyS: Easier to put N-Triples into Turtle, used as a dump format and import format 15:50:22 +1 15:50:25 +1 15:50:55 15:51:10 eric: repeating AndyS: I'd like a standard around N-Triples, it is out there, it is used, lets put it on a standards basis. 15:51:11 turtle: a . 15:51:12 ntriple: a . 15:51:28 +1 to and a mime type 15:51:28 q+ to discuss differences between turtle and ntriples 15:51:35 what is required to turn an n-triples doc into a Turtle doc? 15:51:38 sandro: Can it be done as a small section of the Turtle grammar, specifuing the restrictions? 15:51:44 AndyS: Yes, and the mime type. 15:52:05 pfps, nothing iirc 15:52:26 sandro: would N-Triples be a sublanguage of Turtle? 15:52:29 AndyS: Yes. 15:52:59 AndyS: Could call it N-Turtle, if we want to distinguish it from N-Triples. 15:53:04 +q ... and N-Quads? ;) 15:53:18 -q ... 15:53:25 -q N-Quads? 15:53:28 -q ;) 15:53:35 +q to talk about N-Quads 15:53:38 ericP: Difference is in IRIs. In N-triples, we we have to do percent-decoding. 15:53:52 [sorry, wasn't looking at the q] 15:54:04 ericP, we need it to be UTF-8 and a subset of turtle 15:54:13 AndyS: I would make N-Turtle, and have it be utf-8. 15:54:44 AndyS: I do come across utf-8 N-triples, and also ascii, with percent-encoding. 15:54:44 +1 likewise, many didn't know it couldn't be ascii in a recent thread on semantic-web 15:54:56 Eric: Some of it is broken, and which way do we want to unbreak it. 15:55:03 +1 use IRIs 15:55:04 N-Turtle is preferrable UTF-8 15:55:04 q- 15:55:04 s/ascii/utf8 15:55:05 +1 and likewise for nquads 15:55:10 Guus: +1 use IRIs 15:55:26 m029206__ has joined #rdf-wg 15:55:27 steve:+1 Andy 15:55:37 SteveH: I agree with Andy. We should make N-Triples a subset of Turtle, give it a mimetype and make it UTF-8 15:55:39 -yvesr 15:56:03 Need a different name (RDF-data?) and explain that it's N-Triples made into a standard. 15:56:10 Yvesr: Oracle position, it's really easy to parse 15:56:17 zwu2: Oracle uses N-Triples heavily. We don't want it deprecated. Yes, it's long but it's REALLY EASY TO PARSE. Folks zip it up. So backward compatibility is important to us. 15:56:23 Ah, wrong name 15:56:50 zwu2: Existing N-Triples should be valid as well. 15:56:54 guus: Oracle okay with using IRIs? 15:57:05 zwu2: Yes, I think 15:57:08 +??P11 15:57:13 Zakim, ??P11 is yvesr 15:57:13 +yvesr; got it 15:57:26 ericP: This change would affect the meaning of any NTriples documents with "%" in them. 15:57:42 q+ to disagree with ericP 15:57:44 +1 to raising an issue 15:57:48 +q to ask rhetorcally if N-Turtle would be harder to parse than N-Triples, pace backwards compatibility 15:57:50 ack zwu2 15:57:54 -q 15:57:57 q- 15:58:28 q? 15:58:28 ACTION: AndyS to make concrete "N-Turtle" proposal 15:58:28 Created ACTION-23 - Make concrete "N-Turtle" proposal [on Andy Seaborne - due 2011-04-06]. 15:58:42 text/ntriples+turtle would be consistent with RFC3023 §7 ¶1 15:58:43 ericP, I'm not sure http://www.w3.org/TR/rdf-testcases/#ntriples says you should decode %s 15:58:44 q- 15:59:09 +1 to Steve. 15:59:11 SteveH: I'm not sure N-Triples doc says you need to unpack %-encoding 15:59:16 SteveH: N-Triples doesn't say you have to un % encode things 15:59:18 Eric: Oh, that would be excellent 15:59:24 ack gavinc 15:59:24 gavinc, you wanted to talk about N-Quads 15:59:35 Q-Turtle 15:59:43 NT has it's own esacaping via \u 15:59:49 yah 15:59:49 q? 16:00:50 sandro: basic non-answer is that N-Quads is part of the [GRAPHS] task force, and it hasn't made enough progress yet to know. 16:01:12 Guus: I think that's all the issues. Unless we missed something? 16:01:26 raise issue against product turtle so we come back to it when [graph] done with it 16:01:42 Sandro: Issue-19, diffrence between SPARQL and Turtle. In N3 and SPARQL you can do a BNode production... you can make a statement by just having it [] 16:01:53 e.g. [ ns:foo 132 ; ns:bar ] . 16:02:37 AndyS: Yes, it should be part of Turtle. 16:02:45 Pat: me too. 16:02:50 Pat: Yes, it should be in Turtle. 16:02:51 sounds good to me 16:02:55 +1 16:02:57 +1 16:03:09 raptor parses it too, fwiw 16:03:15 Be careful that "[] ." does no parse 16:03:20 cygri has joined #rdf-wg 16:03:32 Be careful that "[] ." does not parse 16:03:40 I'll speak up here. 16:03:49 [ 132 ; 23 ] . = two triples in raptor 16:04:11 Gavin: I think most parsers work this way already 16:04:17 Sandro: Maybe gather more data? 16:04:20 -EricP 16:04:25 Jena RIOT accepts it (as 2 triples) 16:04:46 excellent, if Jena and Raptor and rdflib are happy with it, I'm happy with it. :-) 16:04:48 my parsers do too, two triples. 16:04:56 Gavin please add my name to attendees 16:04:58 Guus: I hope we can make some progress on this at the F2F 16:05:02 the parsers speak... 16:05:22 Topic: General Issues 16:05:36 Sandro: Should we talk more about the F2F? 16:05:57 Guus: We won't get to these general issues today. 16:06:32 Guus: Are there any women in this WG? 16:06:34 Gavin: No. 16:06:42 -cmatheus 16:07:11 ACTION: Andy to write collect issues and write a proposal standardizing N-Triples. 16:07:11 Created ACTION-24 - Write collect issues and write a proposal standardizing N-Triples. [on Andy Seaborne - due 2011-04-06]. 16:07:25 AxelPolleres has left #rdf-wg 16:07:53 me on JSON TF, I'll try to do my best .) for the F2F 16:08:14 -webr3 16:08:26 AxelPolleres has joined #rdf-wg 16:09:06 Sandro: Job of the presentor is to make sure the people at the F2F have a good idea of what's going on 16:09:29 you dont have to use powerpoint.... 16:09:57 SteveH: Yeah, I'll do the Turtle TF presentation at the F2F 16:10:42 Guus: Richard in the best position to give this overview 16:11:04 s/this/Graph 16:11:52 Guus: JSON, hard to summarize 16:11:58 AxelPolleres has left #rdf-wg 16:12:05 -yvesr 16:12:08 AxelPolleres has joined #rdf-wg 16:13:02 Sandro: I'll ask Thomas to present JSON TF at the F2F 16:13:14 Guus: Good to do a summary of the Other Things. 16:13:46 Sandro: Willing to present deprecation issues 16:13:59 Sandro: Make Ivan do the other ones. 16:14:32 q? 16:14:45 Guus: Thank you all. See you next week. 16:14:51 - +1.850.377.aadd 16:14:58 bye 16:15:02 -AxelPolleres 16:15:03 -OlivierCorby 16:15:06