13:00:03 logger_1 (rdfcore-lo@tatooine.ilrt.bris.ac.uk) has joined #rdfcore 13:00:03 Users on #rdfcore: @logger_1 13:00:03 ChanServ has changed the topic to: 13:00:03 This channel has been registered with ChanServ. 13:18:56 AaronSw (Snak@c930384-a.hlndpk1.il.home.com) has joined #rdfcore 13:40:00 dajobe (jan.grajnt@phobos.hpl.hp.com) has joined #rdfcore 13:40:30 Hello 13:40:42 I'm confused - is this dajobe or jan.grant? 13:40:47 hi - dave 13:40:49 brb 13:40:51 dajobe has left channel 13:41:18 dajobe (dave.becke@phobos.hpl.hp.com) has joined #rdfcore 13:57:40 spetschu (spetschu@cr660071-a.rct1.bc.wave.home.com) has joined #rdfcore 13:57:49 bwm (bwm@phobos.hpl.hp.com) has joined #rdfcore 13:58:30 * AaronSw dials 13:58:41 * dajobe dialing 13:58:51 +Aaron 13:59:06 +danbri 13:59:19 +Frank 13:59:25 danbri? 13:59:28 delete that +danbri 13:59:44 gk (GK@194.205.99.125) has joined #rdfcore 13:59:59 +Steve 14:00:08 mdean (mdean@hh1114013.direcpc.com) has joined #rdfcore 14:00:32 mdean is now known as mdean_ 14:00:57 Note to Scribe: Attach IRC Log to bottom 14:01:02 of minutes 14:01:40 +GK 14:01:43 +Ora 14:02:18 +Bill, Mike, Jos 14:02:30 WELCOME, ROLL CALL, SCRIBE 14:03:03 regrets: danbri 14:03:05 bwm present 14:03:11 regrets from eric 14:03:17 regrets from artb 14:03:20 daveb present 14:03:24 frankb missing 14:03:29 danc regrets 14:03:32 ron daniel missing 14:03:38 bil dehora present 14:03:41 jos de roo, present 14:03:44 rael missing 14:03:48 jang regrets 14:03:58 martyn horner regrets 14:04:08 yoshiyuki, regrets 14:04:10 gk present 14:04:17 kopchenov missing 14:04:24 ora, manola present 14:04:28 nakimura regrets 14:04:32 steve present 14:04:36 pierre missing 14:04:39 aaron here 14:04:44 miked present 14:04:54 guha, phayes, sergey missing 14:05:07 REVIEW AGENDA 14:05:15 http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0114.html 14:05:27 - Agenda 14:05:32 REVIEW MINUTES OF MEETING on 6th July 2001 14:05:33 http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0052.html 14:05:40 no objections 14:05:46 approved 14:05:51 +Pat Hayes 14:06:06 REVIEW AGENDA 14:06:17 Frank has item about DAML joint comitte 14:06:24 coming early in the meeting 14:06:58 2001-06-29#5, 2001-06-08#2 - done 14:07:10 JUN-01-01-#9 - CONTINUED, next week 14:07:19 2001-07-06#1 - DONE 14:07:27 2001-06-22#5 - CONTINUED 14:07:39 2001-07-06#2 - CONTINUED 14:07:50 2001-07-06#3 - continued, sergey missing 14:08:14 DAML JOINT COMMITTEE 14:09:04 http://www.daml.org/2001/07/RDFS-DAML+OIL-coordination.html 14:10:50 person to read thru, compare with our issues list 14:10:57 get agreement with the working group 14:10:59 and send response to DAML 14:11:24 frankm: many technical issues already have had technical discussion 14:11:33 considerable technical discussion 14:11:44 bwm: two things - an acknowledgement and an assesment 14:11:45 sergey (melnik@Barnacle.Stanford.EDU) has joined #rdfcore 14:12:15 ACTION bwm: acknowledge receipt 14:12:27 ACTION frankm: Draft an assesment 14:12:43 bwm: should I indicate a timeframe for assessment? 14:13:16 gk: should be done fairly quickly -- don't need solutions, just issue pointers 14:13:22 bwm: by the end of the month, i'll say 14:14:35 - Formalizing N-Triples (5 mins) 14:14:39 http://purl.org/net/dajobe/2001/06/ntriples/ 14:14:59 bwm proposes: 14:15:00 o this becomes the 'official' definition of n-triples 14:16:00 few comments 14:16:18 bill dehora: in spec as an appendix 14:16:21 bwm: key part 14:16:28 ??: key phrase is working document 14:16:33 sergey: modified as we go along? 14:16:40 yes, that's the intention 14:16:50 any objections? 14:16:53 none. 14:16:59 o call for a maintainer 14:17:10 suggestion: dave and art get together 14:17:24 ACTION dave: resolve who will maintain document 14:17:37 aaron: can this be put in w3c space? 14:17:42 bwm, dave: yes 14:17:46 Aaron: if this is to be an official doc, it shoiuld be in W3C space 14:18:01 bwm: yes, it would move w/ a copyright notice, etc. 14:18:34 ACTION bwm: mail out message announcing registration 14:18:49 re: ACTION 2001-06-08#2: (5 mins) 14:18:56 Propose to update the issues list to point to this text. 14:19:06 no feedback from the list 14:19:12 any objections? 14:19:20 done 14:20:10 gk: not in a position to agree as a resolution, not studied it 14:20:15 bwm: people need more time to study? 14:20:20 talking about http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0039.html 14:20:24 gk: yes 14:20:33 ACTION all: review the document 14:20:39 it will return on next week's document 14:20:41 err agenda 14:21:02 Preparing for the f2f: (10 minutes) 14:21:10 registration above refers to f2f 14:21:38 sergey sent out email, ordering people agreed upon mostly 14:21:53 issue of scope: what needs to be done, what is expected result 14:22:09 people weren't sure aboout where to place it in priorities 14:22:15 all items are expensive topics 14:22:24 sergey's email - http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0033.html 14:22:32 the consensus seems that it's a first-cut summary of the three priorities 14:23:11 open mic = open microphone 14:23:27 brian: proposed the summarization, and divided the open issues 14:23:29 brian's list: http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0072.html 14:23:39 top group was priority because it affected abstract syntax 14:23:47 bottom group somewhat deemphasized 14:24:08 any other issues that affect abstract syntax? 14:24:11 please speak up 14:24:20 sergey: I'm raising namespaces 14:24:27 ACTION bwm: Open the namespaces issue 14:24:31 sergey is owner for issue 14:25:27 bwm: if we had a model theory, it'd be easy to answer questions 14:25:39 ... seems we have some logicians, so lets invite them to give a strawman 14:25:47 ... gives us a focus for discussion 14:25:55 gk: i agree 14:26:06 frank: second, jos: third 14:26:19 sergey: several times, discussion of common terms 14:26:23 seems very serious 14:26:38 before we have detailed discussions on model, we need to know about terms 14:26:54 some sort of itemized list of discussion 14:27:19 gk: put out a message reminding of terminology doc 14:27:25 started while back in rdfinterest 14:27:44 graham http://lists.w3.org/Archives/Public/www-rdf-interest/2001Jan/att-0064/01-RDFTerminologyConcepts.html 14:27:53 sorry -- i thought i had a call waiting beep 14:28:10 phayes: i've volunteered to put together straw model theory 14:28:15 ... probably incomplete 14:28:27 bwm: incomplete is fine, something to get started 14:28:44 hayes: i have to agree with term problem, but doing formalizing is way of getting definitions clear 14:28:49 * dajobe agrees 14:28:52 ... do them in parallel, not one before another. 14:29:10 frankm, gk, sergey: agree 14:29:37 aaron: also make sure that we answer the question 14:29:49 hayes: model theory is a sandbox, focuses the questions 14:30:01 i agree - let's not replace answering the questions 14:30:09 Pat: absolutrely -- think of the model theory as a mathematical sandbox to try out the answers 14:30:10 manola: an excercise in answering the question 14:30:26 pat: it gives us more detailed answers 14:30:44 action pat: produce the strawman 14:30:56 Review status of #rdfms-identity-anon-resources (15 mins) 14:31:06 brian: reviewing issues with high priority 14:31:24 about 50 mail messages on this topic 14:31:28 i hope there is some progress 14:31:46 gk: i think there is progress, recent exchange with frank, want to hear danc's comments 14:32:27 w/ frank we've narrowed issue down to one of query on one side and and issue of generation of a variable to capture semantics of there exists 14:32:40 other comments fit within that spectrum 14:32:44 (gk, am i getting this right?) 14:33:16 hayes: your dismissal of query, while correct, will not be accepted so well -- for many people its curcial 14:33:27 gk: ok, i suppose I have to wait for an argument 14:33:33 hayes: just a heads-up 14:34:09 manola: if this is an issue, and one of the clarifications to go in the m&s 14:34:43 query draws on semantics not inherent in rdf core 14:34:46 - gk 14:34:53 gk: can we broaden the discussion? 14:35:04 hayes: it would be useful if the documentation had explicit warnings 14:35:10 saying things are out of scope 14:35:25 in DAML we said in the spec things that people misunderstand 14:35:34 we're going to flag warning signs 14:35:41 frank: i agree, two basic approaches 14:36:11 1) we recognize sending rdf expression to agent as query 14:36:25 people are afraid to assign query semantics, those are not mandated 14:36:29 query folks are on their own 14:36:29 2 14:36:43 2) we could do the extension ourselves - not advocating, but a possibility 14:37:01 gk: basically agree -- all for warnings 14:37:06 be careful - don't stray too far 14:37:23 don't get stuck on difficult discussion 14:37:52 bwm as techie: gk, you think we have a resolution? what is it? 14:38:33 gk: current spec suggests key idea is that anon resources say "there exists something with these properties" 14:38:39 jos: 1 or only one, or more than one? 14:38:48 gk: not limited to being a single thing 14:39:00 i refer folks to frank's recent message 14:39:06 there is more precise wording 14:39:17 seems to come down to 2 ways of expressing idea 14:39:27 1) skolemization - creating identifier 14:39:45 2) introduction of some scoped variable, something which is more explicit 14:39:53 - an existentially quantified variable 14:40:20 bwm: which way 14:40:24 gk: problems each way 14:40:34 1: how do we generate it 14:40:38 2: how do we scope it? 14:40:50 not clear to me how we scope a variable in the abstract syntax 14:40:58 without collection of statements 14:41:05 bwm: link to the graph issue 14:41:12 gk: right, important point to capture 14:41:29 bwm-techie: on the genid approach, do you thinnk you'd recognize names as being "different" 14:41:35 or would it be indistinguishable 14:41:53 gk: my thinking is indistinguishable, but concepts having uid that's distinguishable.... 14:41:59 ... somewhat sitting uncomfortably 14:42:13 don't have a particularly strong view, slight preference towards scoped variable 14:42:23 jos: scope could be in the document itself 14:42:32 ntriples makes it quite clear 14:42:39 _:things are variable names 14:42:51 gk: i have a problem with scope=document -- what happens when you combine documents? 14:43:02 you need way of capturing boundaries 14:43:08 within the abstract syntax 14:43:41 frankm: issues to resolve include whether it's a syntactic exercise or something more substantial 14:44:08 if i merge two logic statemnts, variables don't refer to the same thing 14:44:43 hayes: have to rename variables for no clashes 14:44:49 i thought that's what URIs were supposed to do 14:45:02 why not just say they're URIs? 14:45:24 frank: let's be clear 14:45:38 bwm: need to move on w/ process and actions 14:45:48 sounds like good deal of progress 14:45:55 gk, what do you need? 14:46:21 gk: need to clarify choices 14:46:29 what pat says was what i saw as skolemization 14:47:40 ACTION frankm: clarify choices for anonymous resources with some ramifications 14:47:58 action on pat? 14:48:24 hayes: question about creating the 'same' genids? 14:48:51 ACTION graham: collect proposals from pat, frank, and any others 14:49:35 ACTION pat: provide his point of view 14:49:47 gk: i think after this we'll be ready to move on 14:50:07 bwm: some colleagues here are using anon resources and they believe it matters 14:50:17 ACTION bwm: write up summary of what they're doing 14:50:36 gk: whn i sent summary, there were several parts 14:50:42 one section was "goals for anon resources" 14:50:46 are there other goals? 14:51:10 ACTION all: submit anonymous resource use cases 14:51:22 ready to move on? 14:51:26 Review status of #rdfms-xmllang (10 mins) 14:51:39 sent summary late last night 14:51:48 three use cases 14:52:00 brian's summary http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2001Jul/0122.html 14:52:36 bwm: M&S is quite clear on this 14:52:45 it describes it as being part of a literal twice 14:52:56 there may be contradictions elsewhere 14:53:11 jang: his implementation used it the way M&S described, preferred that 14:53:33 martyn: sales and services says we don't model it that way -- we represent the concept with anonymosu resource 14:53:46 just wouldn't use M&S's way 14:54:22 ericm: provided seemingly similar use case -- couldn't determine whether m&s solution met requirements 14:54:29 think it probably would, but better followup 14:54:49 ACTION bwm: followup with eric on whether M&S meets his reqs 14:54:57 * gk losrt my telephione link for a minute or so 14:55:11 some questions: 14:55:15 o Do we understand what M&S says about xml:lang? 14:55:18 Yes 14:55:27 o Is what M&S describes broken - is there some fatal flaw that must be 14:55:27 fixed? 14:55:54 hayes: clearly not, since some have implemented it 14:55:59 aaron: timbl and danc have problems with it 14:56:10 hayes: timbl is quite clear, xmllang is a mistakr 14:56:37 hayes: uses term brain-damaged 14:56:51 s/hayes/gk/ 2 above 14:57:07 bwm: do we fix it or should we leave it? 14:57:24 gk: i see 3 possible ways -- 14:57:33 1) idea of literals as complex entities 14:57:36 2) model it in the graph 14:57:45 3) adopt both of those 14:57:52 bwm: let the market decide? 14:58:01 gk: maybe, let a thousand flowers bloom 14:58:16 hayes: something a little weird to represent it two different ways 14:58:33 might get contradictions 14:58:37 frank: sounds likely to happen 14:58:50 bwm: two minutes left 14:59:02 like graham's suggestion about writing up alternatives with pros and cons 14:59:18 i'd be happy to give the action to someone else 14:59:41 ACTION bwm: do a writeup 14:59:44 running out of time 14:59:50 meeting same time same place next week 14:59:56 hayes: same time as daml pi meeting 14:59:58 i'll send email 15:00:07 bwm: i'll bear that in mind for agenda 15:00:09 any other comments? 15:00:17 CLOSE OF MEETING 15:00:38 thanks aaron 15:00:43 no problem 15:00:51 btw, does anyone know where eric is? 15:00:53 back later 15:00:55 dajobe has quit (Leaving) 15:00:58 sergey has left channel 15:01:14 gk has left channel 15:02:31 spetschu has left channel