14:56:55 RRSAgent has joined #rdf-wg 14:56:55 logging to http://www.w3.org/2011/09/14-rdf-wg-irc 14:56:57 RRSAgent, make logs world 14:56:57 Zakim has joined #rdf-wg 14:56:59 Zakim, this will be 73394 14:56:59 ok, trackbot; I see SW_RDFWG()11:00AM scheduled to start in 4 minutes 14:57:00 Meeting: RDF Working Group Teleconference 14:57:00 Date: 14 September 2011 14:58:18 SW_RDFWG()11:00AM has now started 14:58:25 + +1.650.917.aaaa 14:58:36 zakim, aaaa is me 14:58:36 +gavinc; got it 14:58:54 +Guus_Schreiber 14:59:28 +Tony 14:59:40 Zakim, Tony is me 14:59:41 +Scott_Bauer; got it 15:00:16 zwu2 has joined #rdf-wg 15:00:23 zakim, code? 15:00:23 the conference code is 73394 (tel:+1.617.761.6200 sip:zakim@voip.w3.org), zwu2 15:00:24 AndyS has joined #rdf-wg 15:00:58 +??P27 15:01:01 +??P36 15:01:07 zakim, ??P36 is me 15:01:07 +AndyS; got it 15:01:10 Zakim, ??P27 is me 15:01:10 +NickH; got it 15:01:11 +zwu2 15:01:16 pchampin has joined #rdf-wg 15:01:18 +Peter_Patel-Schneider 15:01:26 ww has joined #rdf-wg 15:01:35 zakim, dial ivan-voip 15:01:35 ok, ivan; the call is being made 15:01:35 pfps has joined #rdf-wg 15:01:42 +Ivan 15:01:50 AlexHall has joined #rdf-wg 15:02:06 + +1.443.212.aabb 15:02:10 Zakim, aabb is me 15:02:11 +??P38 15:02:17 nono 15:02:22 +ww; got it 15:02:24 Zakim, ??P38 is me 15:02:29 zakim, mute me 15:02:36 +ww; got it 15:02:42 NickH should now be muted 15:02:48 Zakim, mute me 15:02:56 ww should now be muted 15:03:19 zakim, who is on the phone? 15:03:19 On the phone I see gavinc, Guus_Schreiber, Scott_Bauer, NickH (muted), AndyS, zwu2, Peter_Patel-Schneider, Ivan, ww (muted), ww.a 15:03:41 +OpenLink_Software 15:03:49 Zakim, OpenLink_Software is temporarily me 15:03:49 +MacTed; got it 15:03:51 Zakim, mute me 15:03:51 MacTed should now be muted 15:04:02 +Sandro 15:04:02 one of the ww's is me, how do i fix it? 15:04:37 I can help Guss 15:04:44 sorry Guus 15:04:56 Scribe: zwu2 15:05:23 AlexHall: sorry about that 15:05:33 Zakim, aabb is AlexHall 15:05:41 sorry, ww, I do not recognize a party named 'aabb' 15:05:49 zakim, ww.a is AlexHall 15:05:49 +AlexHall; got it 15:05:57 zakim, who is on the call? 15:05:57 On the phone I see gavinc, Guus_Schreiber, Scott_Bauer, NickH (muted), AndyS, zwu2, Peter_Patel-Schneider, Ivan, ww (muted), AlexHall, MacTed (muted), Sandro 15:06:30 topic: introductions 15:06:32 thanks, ww 15:06:40 PROPOSED to accept the minutes of the 7 Sep telecon: 15:07:01 Zakim, please dial ericP-office 15:07:01 ok, ericP; the call is being made 15:07:02 +EricP 15:07:11 AZ has joined #rdf-wg 15:07:14 RESOLVED: accept the minutes of the 7 Sep telecon 15:07:19 zakim, mute me 15:07:19 Ivan should now be muted 15:07:32 topic: action reviews 15:07:59 Guus: David put up F2F agenda 15:08:29 ... draft agenda http://www.w3.org/2011/rdf-wg/wiki/F2F2#Agenda 15:08:49 ... open actions 15:09:01 ACTION-39? 15:09:01 ACTION-39 -- Sandro Hawke to look at respec's handling of references -- due 2011-04-21 -- CLOSED 15:09:01 http://www.w3.org/2011/rdf-wg/track/actions/39 15:09:27 Action-67? 15:09:27 ACTION-67 -- Patrick Hayes to review Pierre-A's comments on SPARQL graph store update protocol -- due 2011-07-06 -- OPEN 15:09:27 http://www.w3.org/2011/rdf-wg/track/actions/67 15:09:45 action Guus: remind Pat on action-67 15:09:46 Created ACTION-87 - Remind Pat on action-67 [on Guus Schreiber - due 2011-09-21]. 15:09:52 action-73? 15:09:52 ACTION-73 -- Fabien Gandon to implement http://www.w3.org/2011/rdf-wg/meeting/2011-08-03#resolution_2 -- due 2012-01-05 -- OPEN 15:09:52 http://www.w3.org/2011/rdf-wg/track/actions/73 15:09:57 Guus: leave it open 15:10:02 action-77? 15:10:02 ACTION-77 -- Richard Cyganiak to add strawman rdf dataset proposal to concepts -- due 2011-08-24 -- OPEN 15:10:02 http://www.w3.org/2011/rdf-wg/track/actions/77 15:10:10 ... is Richard in? 15:10:22 I think he is out these days 15:10:30 action-79? 15:10:30 ACTION-79 -- Guus Schreiber to discuss with FabGandon about where to put examples and what to do with primary before F2F -- due 2011-09-29 -- OPEN 15:10:30 http://www.w3.org/2011/rdf-wg/track/actions/79 15:10:37 ... needs to be continued, for later 15:10:42 Action-82? 15:10:42 ACTION-82 -- Sandro Hawke to draft well-known URI template and propose WG resolution that it is "stable" enough for IETF. -- due 2011-09-14 -- OPEN 15:10:42 http://www.w3.org/2011/rdf-wg/track/actions/82 15:10:48 ... Sandro, did you make progress 15:10:52 sandro: no 15:11:01 action-86? 15:11:01 ACTION-86 -- Sandro Hawke to provide example for how code is simpler with language-tag options 3 and 4 vs 1 and 2. -- due 2011-09-14 -- OPEN 15:11:01 http://www.w3.org/2011/rdf-wg/track/actions/86 15:11:25 +??P58 15:11:59 proposed to close action-86 15:12:11 mischat has joined #rdf-wg 15:12:37 Topic: F2F planning 15:12:51 Guus: wiki page: http://www.w3.org/2011/rdf-wg/wiki/F2F2 15:12:59 ... 11 people at BBC, 15:13:13 ... peter, will you attend? 15:13:21 swh has joined #rdf-wg 15:13:25 peter: may not know until Tues morning 15:13:42 + +33.4.77.42.aacc 15:14:00 zakim, +33.4.77.42.aacc is me 15:14:10 +AZ; got it 15:14:38 Guus: draft agenda for two days http://www.w3.org/2011/rdf-wg/wiki/F2F2#Agenda 15:14:59 ... we have more time now due to timezone difference 15:15:14 ... any comments? 15:15:31 q+ 15:15:48 zakim, who is speaking 15:15:48 I don't understand 'who is speaking', zwu2 15:16:00 thanks 15:16:25 Guus. by the time of Boston f2f, 15:16:38 ... we should have some new material 15:16:59 ... for RDF Primer 15:17:07 ... unlikely to be complete 15:17:18 zakim, unmute me 15:17:18 Ivan should no longer be muted 15:17:52 q+ 15:17:56 ivan: one of the two hotels does not have any more room 15:18:01 I will pass on the message to yvesr 15:18:10 q- 15:18:54 ... can you send out that hotel info? 15:19:13 comments on london hotels very true! 15:19:32 the pound is weak at the moment, people should be "ok" 15:19:41 action Guus: provide information on london hotel 15:19:41 Created ACTION-88 - Provide information on london hotel [on Guus Schreiber - due 2011-09-21]. 15:20:18 Topic: liaison with Provenance WG 15:20:34 Guus: we will use RDF WG channel 15:20:52 ... zakim code is RDFG 15:21:13 ... attendance is in http://www.w3.org/2011/rdf-wg/wiki/Meetings:Telecon2011.09.15 15:21:22 pchampin has joined #rdf-wg 15:21:29 ... not sure who will attend from Provenance WG 15:21:40 ... two main objectives 15:21:47 ... 1. requirement 15:22:25 ... work out use cases 15:22:56 ... have joint actions to work on the examples (concrete) 15:23:40 +1 to concrete examples 15:23:45 i thought that the provenance group where going to give us concrete examples? 15:23:52 zakim, mute me 15:23:52 Ivan should now be muted 15:24:06 topic: RDF Dataset proposal 15:24:24 Guus: richard is essential to this discussion, He is not here 15:24:27 +??P59 15:24:35 zakim, ??P59 is me 15:24:35 +pchampin; got it 15:25:02 ... not sure if we can discuss this item 15:25:25 zakim, unmute me 15:25:25 Ivan should no longer be muted 15:25:41 ivan: we should not go into discussion 15:26:19 sandro: the doc does not look much different to me 15:26:42 Souri has joined #RDF-WG 15:27:11 sandro: it is not a great proposal 15:27:21 ... his design is minimal 15:27:25 +Souri 15:27:49 s/great/sufficient/ 15:28:34 sandro: applications need their vocab to do something useful 15:28:59 ... we need to proceed with more use cases 15:29:14 q? 15:29:18 q- 15:29:21 -Peter_Patel-Schneider 15:29:43 Topic: language tagged literals 15:30:13 Guus: only two proposals look promising 15:30:33 ... 1 or 2d 15:31:05 ... either we do nothing, or we go with 2d 15:31:22 ... we don't have to make a final decision today 15:31:32 ... for resolutoin, we could go for 2d 15:31:52 ... if we can sufficient push back, then we choose 1 and do nothing 15:32:00 sandro: I agree that approach makes sense 15:32:21 ... doing 2d at provisonal at risk 15:32:26 +1 to what to sandro says 15:32:35 ... let community decide if it is desired 15:32:58 Guus: not sure much is going to change before F2F 15:33:24 ... Pat said he prefers 4 15:34:06 -> http://www.w3.org/2002/09/wbs/46168/tagged_literals/results results of the poll 15:34:08 ... I will write an email to the group 15:34:30 For the records, 2d means: "The abstract syntax has a lexical form and language tag (like in RDF 2004). The value is assigned directly (like in RDF 2004), bypassing the datatype. The datatype has an empty lexical space and empty L2V mapping." 15:35:14 Or, more simply -- it's the same as we have now, except instead of an EMPTY datatype for tagged strings, there's a particular value for those strings. 15:35:24 action Guus: email to group stating a possible resolution for tagged literals 15:35:25 Created ACTION-89 - Email to group stating a possible resolution for tagged literals [on Guus Schreiber - due 2011-09-21]. 15:36:18 Guus: aob? 15:37:32 telecon adjourned 15:37:37 sandro: two axes: 15:37:44 ... current: (lex, (datatype | langtag) 15:37:49 ) 15:38:04 Choice One: or just 15:38:32 isn't the current: | ? 15:39:13 Choice Two: smoosh the lang tag into Lexrep (option 4) or into datatype URI (option 3a). 15:39:31 -zwu2 15:39:49 -Souri 15:40:09 current: | | 15:40:14 -NickH 15:40:38 current, post xsd:string normalization: | 15:42:22 agreed, option 4 is v. bad 15:42:25 +1 to -1 of smushing into the string 15:42:32 Option 4 is bad 15:42:48 +1 as long as we're talking about the RDF abstract syntax 15:43:15 -AZ 15:43:19 opt 4 is very very bad AKA it will be ignored - APIs will hide it completely so does not exist 15:43:22 language tags REQUIRE micro parsing already btw 15:43:23 several of us were originally interested but DT-lang but it had problems. 15:43:37 ivan: my problem with 3 is the vast number of language tag URIs that we'd have. that might cause problems. 15:44:03 gavinc - not really - can treat (unchecked) as strings. 15:44:27 that's true, okay... langMatches() requires micro parsing 15:44:28 true if checking -- but can "assume" valid 15:44:40 ivan: owl2 semantics, generatng things like differentFrom for each datatype 15:44:48 Yes, langMatches is ... interesting. 15:45:59 http://tools.ietf.org/html/rfc4647 15:46:36 gavin; I don't object to 3a due to microparsing -- since you need to "microparse" language tags anyone. 15:47:08 +1 to Ivan: 3a makes no statements about relations between the language types 15:47:39 AndyS: the mental framework from datatypes does not apply to the theory of lanuage tags. 15:47:55 q+ to ask a question about 3a and datatypes 15:48:07 ack me 15:48:09 pchampin, you wanted to ask a question about 3a and datatypes 15:48:09 Also JJC email that relationship of DT != lang relationships 15:48:13 (ah, that makes sense, Andy.) 15:48:24 we need to A/B test by teaching folks the different models and seeing which they "get" in the sense of having intelligent intuitions 15:48:43 I prefer 2d because it apprears to be the smallest change that does the trick, but maybe someone can convince me this is wrong 15:49:11 -Scott_Bauer 15:49:24 zakim, mute me 15:49:24 pchampin should now be muted 15:49:28 Turtle funkyness, "Bob"@en == "Bob"^^lang:EN == "Bob"^^lang:en 15:49:52 my question is: are those datatypes (in option 3a) different in any way form other datatypes? 15:50:04 I think not, but some ppl seem to think so 15:50:16 datatype("Bob"@en) doesn't throw an error 15:50:39 http://lists.w3.org/Archives/Public/public-rdf-wg/2011May/0425.html 15:50:55 gus, we lost yhou 15:51:05 JJC's message from our archives 15:51:06 s/gus/guus/ 15:51:07 ok thx 15:51:14 I'll look at that in more detail 15:51:20 -.- 15:51:38 -AlexHall 15:51:48 ^^lang:en-fr 15:52:00 sorry, seem to have problems 15:52:09 "Bob"^^lang:en "Bob"@en 15:52:16 @gavin: we already have "bob"@en == "bob"@EN 15:52:28 and we don't have to allow lang:EN... could normalize to lang:en 15:52:44 Yes! but .. "1"^^xsd:int != "1"^^xsd:INT 15:52:47 + a prefix declaration. 15:52:53 ok, you each get 3billion people... 15:53:00 :-= 15:53:04 :-) 15:53:51 -MacTed 15:53:54 I wasn't aware any proposal was deprecating @ 15:54:05 what happens in (shudder) RDF/XML? 15:54:14 @AndyS no, this was not the intent of 3a akaic 15:54:58 trix? 15:55:13 [my phone apprears to be broken, sorry] 15:55:25 (I have never seen TriX in use) 15:55:37 zakim, unmute me 15:55:37 pchampin should no longer be muted 15:55:47 okay, you're close to convincing me to give up Option 3..... 15:56:11 we can normalize @EN to lang:en 15:56:23 the parser could do the smackdown on lang:eN-Fr 15:56:30 zakim, mute me 15:56:30 pchampin should now be muted 15:56:43 "Bob"@en == "Bob"^^lang:en == "Bob"^^lang:en != "Bob"^^lang:EN 15:56:49 yes 15:56:54 right. 15:56:55 "Bob"@en == "Bob"^^lang:en == "Bob"@EN != "Bob"^^lang:EN 15:59:40 so maybe we have make the lang-tag -> URI mapping fold to lower case. 15:59:56 ivan: i think technically that would be okay. 15:59:59 IRI equivalence: http://tools.ietf.org/html/rfc3987#section-5.1 16:00:07 AndyS: Didn't we ask the OWL people? 16:00:15 "Applications using IRIs as identity tokens with no relationship to a protocol MUST use the Simple String Comparison" 16:00:18 ivan: Ian Horrocks said it was okay, but maybe that was just in theory. 16:01:56 gavin: xpath defines language tags as case insentivie, too. 16:02:22 eric: So everyone says they are case insensitive, so normalizing would be okay. 16:03:06 Plain literals have a lexical form and optionally a language tag as defined by [RFC-3066], normalized to lowercase. 16:03:12 gavin: rdf spec says "normalized to lowercase" 16:03:59 lang:en owl:sameAs eric:HeLLo 16:04:43 the proposal here for handling case is: @en == @EN == lang:en but NOT lang:EN 16:04:58 zh-Hant 16:05:07 (not that anyone is arguing in favor of this.) 16:06:26 I think the @ syntax has to case fold, otherwise it will be too strange 16:07:15 for sparql term matching, swh? 16:07:27 sandro, yes 16:07:50 I would ban "foo"^^lang:bar in turtle, though that's a little strange 16:07:54 makes the abstract simpler simpler 16:08:05 and seems to make everything else more complex :) 16:08:22 http://www.ietf.org/rfc/rfc5646.txt sec 4.5 16:08:29 datatype("bob"@en) does not do something errorish 16:08:48 what does currently datatype("chat"@en) currently return? 16:09:00 nothing, it's an error 16:09:20 ok, so anyway, all options that aim to give a datatype to *every* literal will change SPARQL, right? 16:09:34 Yep! 16:09:52 is there an option on the table that will change SPARQL less than the others? 16:10:02 Yep! Do nothing. 16:10:06 :) 16:10:41 -AndyS 16:11:50 [i'm back] 16:12:01 q+ 16:12:02 +??P30 16:12:10 zakim, ??P30 is me 16:12:10 +AndyS; got it 16:12:11 eric: Gavin says the reason for 2d over 1 is so you can do datatype("chat"@en). 16:12:19 zakim, please do not cut me off 16:12:19 I don't understand 'please do not cut me off', AndyS 16:12:41 gavin: Folks in existing queries should be using NOT LANG(...) 16:13:05 pchampin - tech answer is SPARQL does not require lang tags ; practically, impls care about STR() most. 16:13:10 eric: there are current queries that will break with this change. 16:13:22 FILTER (LANG(?a) != "en-fr") 16:13:35 gavinc: but those are funky queries, so it's okay. 16:13:50 then LANG() then LANGMATCHES() 16:13:58 eric: yeah, hard to find an example 16:14:19 andy: FILTER with DATATYPE then LANG_MATCHES 16:14:59 no, that was already in SPARQL 16:15:02 gavinc: We already broke that by removing Plain Literals. 16:15:27 ok 16:15:36 good point 16:15:36 which was exactly our intent :-) 16:15:42 okay. I'm convinced to support 2d. 16:15:49 sandro, you had a question for me which i did not hear and therefore did not answer 16:15:51 I move too 16:16:26 (in my heart I still love 3a, but ... it does seem like there are a lot of complications....) 16:16:33 bye everybody 16:16:38 -Ivan 16:17:04 for the record, I'm still convinced that 2d can be interpreted as 3b 16:17:07 FILTER( (LANG(?a) != 'cd' | true) (?a -> '1'@ab) (?a -> '2'@cd) (?a -> '3'^^xsd:int)) 16:17:10 => (?a -> '1'@ab) 16:17:11 if that's any confort to Sandro ;) 16:17:24 topic: turtle in html 16:17:29 -ww 16:17:33 AlexHall has left #rdf-wg 16:18:10 -Guus_Schreiber 16:19:16 nasty!!! 16:20:04 16:21:36 Sandro, what convinced you to support 2d? 16:23:16 davidwood, Gavin and AndyS pointing out tricky corner cases that I didnt know how to solve well. 16:23:36