**** BEGIN LOGGING AT Mon Sep 18 07:30:52 2000 UTC-0500 07:30:52 --> DanC (connolly@adsl-208-190-203-22.dsl.kscymo.swbell.net) has joined #topicmap-rdf 07:33:13 --- DanC has changed the topic to: Is this where we meet, EricM? 08:58:27 --- DanC has changed the topic to: topicmap/rdf chat http://dublincore.org/2000/09/18-rdftmagenda.html 09:58:29 --> EricFrees (EricFrees@179.minneapolis-03-04rs.mn.dial-access.att.net) has joined #topicmap-rdf 09:58:33 # Appears as XENO. 10:00:22 --> emiller (emiller@krash.dev.oclc.org) has joined #topicmap-rdf 10:02:07 --> RalphS (swick@24-6-128.wireless.lcs.mit.edu) has joined #topicmap-rdf 10:02:44 danc are you joining up on teleconference 10:03:04 oh... it's that time. I was working on http://www.w3.org/2000/09/ml54/Overview.html 10:04:38 -> http://dublincore.org/2000/09/18-rdftmagenda.html agenda & background 10:05:27 * DanC finds phone details... 10:06:07 attendance: regrets from brickley 10:06:18 attendance: swick, connolly, emiller (chair) present (others...) 10:06:50 attendance: Eric Freese present 10:07:27 EricM: goal: let's not let another get-together happen with no closure. 10:07:37 did I hear that Michel is here? 10:07:43 yes he is 10:08:10 attendance: Michel Biezunski present 10:08:26 how about Newcomb ? is that him now? 10:08:32 yep 10:08:38 EricM: how about an update of the [???] meeting 10:09:26 Newcomb: after some meeting-greeting stuff at the meeting, we decided to try to have something done for XML 2000 10:09:34 ... DTD, (something else? help?) 10:09:47 ... we formed several subcommittees 10:09:55 ... 13 Oct in England (something? help?) 10:09:59 several models 10:10:01 , etc. 10:10:12 UML model 10:10:33 --> tim-lurk (timbl@politburo.w3.org) has joined #topicmap-rdf 10:11:06 Micheal: version 1 of the spec by XMl 2000 10:11:15 Michel: more than that: it's v1.0 of the spec that is our goal for XML 2000 10:11:45 emiller, pls make the point that this record is for public consumption 10:12:07 attendance: Steven R. Newcomb present 10:12:16 crud... i'm not (yet) logging into the web 10:12:57 * DanC goes to set up logging into HTTP... brb... 10:13:43 Newcomb: we took the "draconian" step of closing the group to new members 10:13:54 EricM: membership of what exactly? I'm not quiet clear 10:14:59 Newcomb: we're trying to get an XML-appropriate, web-appropriate, (other?) result... 10:15:22 ... we're trying to make certain that those doing the work are completely in control; this means we had to reach deeply into our own pockets to get this done. 10:15:43 result is also planned to be compliant with the ISO standard 10:15:54 Newcomb: if you showed up for any of the founding meetings (there were 3 of them) you're a member; you can only miss one meeting (consecutively) or you lose your membership 10:16:06 Newcomb: proceedings are completely public 10:17:25 Newcomb: we're currently hosted by idealliance, but it's a matter of written records that we can fire that at any time 10:17:32 Swick: topicmaps.com? is that your site? 10:17:45 Newcomb: no, that's peppers 10:18:05 -> http://www.egroups.com/group/xtm-wg xtm-wg stuff referenced by Michel 10:18:27 Swick: connolly sent a summary of the lunch in Montreal which got me pretty enthusiastic about this effort 10:19:01 Swick: first thing that happened when I tried to follow the xtm-wg pointers was a request for personal information that I'm not willing to divulge. 10:19:43 Swick: what facilities of egroups are you using? just a maling lists? 10:20:17 Newcomb: just the mailing list and the file share 10:20:49 Newcomb: we could, of course, run these services under our own domain, but we haven't found time to do so. 10:21:48 Swick: there's a mention of an interesting-sounding paper by [??] on simultaneous RDF and topicmaps; pointer? 10:22:16 Connolly: er... think so... it's in the EXML proceedings, which I have on CD, and I think it's online 10:22:18 -> http://www.xml.com/pub/2000/09/13/extremes2.html "The Free-For-All in Montreal 10:22:20 [11] Simultaneous Topic Maps and RDF metadata structures in SVG 10:22:20 , David Dodds, Open Text 10:22:32 no url that I know of 10:23:32 I see http://infoloom.com/tmarticl.htm 10:23:39 swick, pointer pls? 10:23:45 -> http://www.infoloom.com/extreme2000/latebreaking.htm 10:23:46 [[[ 10:23:46 Topic Maps, Next Generation, a presentation made by Michel Biezunski at the Extreme Markup 2000 Conference, in Montreal, August 2000. 10:23:47 10:23:48 ]]] 10:23:52 infoloom.com site nice site for topicmap info... helpful for me in bridging the RDF topicmap gap 10:25:23 http://www.w3.org/RDF/Interest/#discussion 10:25:35 `rdf interest group listserv 10:25:47 possible place for topicmap / rdf discussions 10:27:02 Michel: looking for update on the current status of RDF 10:27:32 DanC: responds M&S W3C REC feb 1999, jan 2000 CR for RDF schema 10:28:40 DanC: software in rapid development (somewhat immature) ... applications more mature (dmoz.org) 10:29:08 DanC: Dublin core one of the mature applications 10:29:18 DanC: begining of API and test-suites, etc 10:29:50 DanC: RDF/Topicmap proposal - internchange syntax is replaceable with anythign that captures the RDF model in pointy brackets 10:30:14 DanC: contraints include 2 resources id'd by URI with a typed relation (also a URI) 10:30:21 DanC: only constraint on the RDF syntax is that there are two endpoints and an arc, any of which can be a URI 10:30:23 DanC: more contraints...? 10:30:26 thanks! 10:31:20 Steve: would the W3C reconize the new syntax? 10:31:24 DanC: sure 10:33:39 other mature/ing RDF application: dublin core 10:33:49 one I didn't mention (orally): Rich Site Summary (rss) 10:35:32 ... various bits not very well-recorded 10:36:06 Swick: some "graph in XML" stuff seems to lack the idea that parts of the graph might be in other documents 10:36:27 Swick: I've seen discussion of "merging topicmaps" which is a key property of the RDF model 10:37:15 Michel: I've been reading about the semantic web and the role RDF plays in it, and I'm struck by the similarity with what we're trying to do with topicmaps... merging, evolvability. 10:37:42 Michel: agenda request: models and concepts from semantic web 10:38:25 2nded by emiller, connolly 10:38:44 Newcomb: mergability was the requirement that led to "the whole thing" 10:39:17 ... we were the davenport group before [???] 10:39:47 aside: connolly and Newcomb met at a davenport meeting in '91 http://www.w3.org/People/Connolly/events/#y1991 10:40:25 Newcomb: topicmaps came from the davenport work on capturing the information typically lost in the process of building an index (OpenGroup was building a master index of their doc) 10:41:12 Michel: docbook went one way (a DTD for documentation) and topicmaps went another... things seem to be coming back together now. 10:41:37 timecheck? 40 minutes into a 120 minute meeting, right? 10:41:48 right 10:42:19 --- scribe-bot #topicmap-rdf 10:42:19 --> scribe-bot (scribe-bot@adsl-208-190-203-22.dsl.kscymo.swbell.net) has joined #topicmap-rdf 10:42:19 * scribe-bot starts log at http://shoal.ods.org/2000/09tr//2000-09-18T154219Z-%23topicmap-rdf.html (/home/connolly/public_html/2000/09tr//2000-09-18T154219Z-#topicmap-rdf.html) 10:42:34 oops... 10:42:35 <-- scribe-bot has quit (normal exit.) 10:43:02 what was that? 10:43:10 --- scribe-bot :No such nick/channel 10:43:21 a scribe robot I'm developing. 10:43:33 --- scribe-bot :No such nick/channel 10:44:08 Swick: are topicmaps designed to express "approved" properties? or ...? 10:44:27 Newcomb: the latter; new properties are just as good as old/approved ones 10:44:42 --- scribe-bot #topicmap-rdf 10:44:42 --> scribe-bot (scribe-bot@adsl-208-190-203-22.dsl.kscymo.swbell.net) has joined #topicmap-rdf 10:44:42 * scribe-bot starts log at http://shoal.ods.org/2000/09tr/2000-09-18T154442Z-%23topicmap-rdf.html (/home/connolly/public_html/2000/09tr/2000-09-18T154442Z-#topicmap-rdf.html) 10:46:22 testing log... 10:46:56 oops... the log is really at http://shoal.ods.org/~connolly/2000/09tr/2000-09-18T154442Z-%23topicmap-rdf.html 10:46:59 but it is working. 10:47:06 -> http://shoal.ods.org/~connolly/2000/09tr/2000-09-18T154442Z-%23topicmap-rdf.html log 10:47:36 EricM: I found that 15 line DTD that was supposed to be common to all topicmaps [did I get that right? close?] 10:49:30 [scribe missed some here] 10:49:39 close 10:49:44 this is what I was asking 10:50:05 Michel: you characterized an RDF/topicmap correspondence; I have a different model of it. 10:50:49 Newcomb: I observe, with happiness, that we're not talking about whether we want to help each other, but just how. 10:51:42 Newcomb: we may need a face-to-face to get into the deep philosophical discussions necessary to figure out how these things impinge on each other 10:54:34 Connolly: I'd like to take a specific application and explore the correpondence that way 10:56:14 ... teleconference logistics questions 10:57:06 http://www.infoloom.com/gcaconfs/WEB/paris2000/bookfront.HTM 10:57:07 -> http://www.infoloom.com/gcaconfs/WEB/index.htm GCA stuff on infoloom 10:57:27 DanC take a look at url... example of paris 2000 topicmap 10:57:37 -> http://www.infoloom.com/gcaconfs/WEB/paris2000/bookfront.HTM example topicmap from paris 2000 10:58:53 a Reading of the URI, which makes datespace-URIs look so much the better ;-) 10:58:58 Newcomb: got it. 10:59:26 Michel takes the floor and explains... 10:59:36 [[[ 10:59:37 Ralph also finds the path: 10:59:37 -> http://www.infoloom.com/faq.htm 10:59:37 --> http://www.infoloom.com/topmap.htm 10:59:37 ---> http://www.y12.doe.gov/sgml/sc34/document/0129.pdf ISO/IEC 13250 10:59:37 ]]] 11:00:09 "topic types" == RDF classes 11:02:21 austrailia, etc. instances of class Country (i.e. topicmap type country) 11:02:28 which were we to click on? 11:03:16 Michel's copy differs from what we see at... 11:03:20 -> http://www.infoloom.com/gcaconfs/WEB/ts3420/tp3420.HTM U.S.A. 11:03:34 so we go to... 11:03:40 -> http://www.infoloom.com/gcaconfs/WEB/ts464/tp464.HTM U K 11:04:40 Contains is an association role 11:05:32 "association role" in topicmaps is called "Property" in RDF 11:05:43 Connolly: does "Contains" have a URI? 11:05:54 Michel: not in this case, but it could have one 11:07:00 EricM: if I used the same id you did for "Contains", in a thing about states/countries, could we merge data [something like that] 11:07:10 Michel: we have this notion of topic identiy [...] 11:07:19 [push that very interesting question; I hope we get back] 11:07:28 where are we now? 11:08:05 ah... 11:08:11 -> http://www.infoloom.com/gcaconfs/WEB/ts1235/tp1235.HTM Berkshire 11:08:51 Contains and Belongs_to are inverses 11:09:38 scope =? reification 11:09:47 Michel: May 1998 is the scope of the association 11:10:02 Michel: scope is a set of teams [?] 11:11:14 Steve: "May 1998" is the name of a topic 11:12:42 yes, it looks like reification to me: [Berkshire -- Belongs_to --> England] --topicmap:scope--> May_1998 11:12:45 scope =? bag 11:14:33 right... but it looks like topicmap:scope is the *only* relation? 11:14:45 there can be several scopes on a topic 11:15:02 I suspect a closed-world assumption around "these are all the scopes" 11:15:12 reification is tough in TMs 11:15:33 Michel: "Mentioned" is [an instance of?] an "occurrence role" 11:16:59 occurrence role =? properties 11:18:54 connolly: are occrences intrinsic (part of the identity of a topic) or not? 11:18:55 answer: not 11:19:22 timecheck: 40 minutes 11:19:26 left 11:20:05 Newcomb: as I recall, thursday evening[?] in Montreal, we discussed... 11:20:50 ... an RDF processor either knows or doesn't know, i.e. has or hasn't seen a [...] 11:23:29 Connolly: [...] 11:23:49 Newcomb: so you do have the concept of bounded object set in Hytime? 11:24:06 Connolly: yeah... the SHOE folks do this... assertion by reference 11:24:28 ... and exhaustive enumerated lists 11:24:53 EricM: this notion isn't in the RDF standard vocabulary. 11:25:25 Newcomb: perhaps it ought to be; I was baffled that it wasn't in RDF (and that it was taken out of xlink, for that matter) 11:25:49 Swick: [... something about constraintproperties...] 11:27:20 Connolly: are you suggesting that we can use ConstraintProperty as a hook for assertion by reference? e.g. cardinality? 11:27:24 Swick: yeah... something like that 11:27:58 Michel: I have been trying to map RDF constraints into topicmap vocabulary... it seem to map to at least 3 things: scope, [...?] 11:30:53 Swick: to say "in order to process these data, you have to fetch X, Y, and Z" presumes what sort of processing the agent is doing 11:35:01 [I think we just reached an important agreement, but we didn't write it down...] 11:35:22 [... something like: we need an RDF property (identifier) for topicmap:necessaryForProcessing] 11:36:36 Connolly: what's it called? 11:36:44 Newcomb: bounded object set. 11:36:46 side note: Danbri reminds me to reference Ron Daniel's work on XLink-RDF mapping http://www.w3.org/RDF/Interest/#docs 11:39:32 how about this: http://www.w3.org/2000/09/tmr#boundedObjectSet 11:39:55 Michel: [aside] names for concepts are called "public topics" in XTM 11:40:34 Ralph: concepts can be modelled as public topics - they do not necessarily equate 11:40:50 thanks for the clarification, EricF 11:41:10 the TM public identifier maps to the URI 11:42:10 a public topic has an identifier (URI) by which it is known 11:42:31 this is how merging is done 11:42:41 push: difference between boundedObjectSet and shoe:useOntology 11:45:30 Connolly: next ftf opportunity looks like March 2001 in Austin -- gca [knowledge management?] conference 11:47:29 Michel: [v interesting questions...] 11:48:05 Newcomb: re those questions: I'm here because MSM folks said topicmap and RDF folks should be locked in a closet and not let out, and I'm heeding his recommendation. 11:53:49 proposed: next metting 11amEDT Monday, 25 Sep 2000 11:54:22 confirmed: Freese 11:54:23 Swick 11:54:26 Newcomb 11:54:28 Michel 11:54:29 Miller 11:54:30 Connolly 11:54:47 invited: danbri. 11:58:14 ACTION Connolly: distribute pointer to record to www-rdf-interest and xtm-wg 11:58:57 EricF: how about taking an example and working out the details in both models? 11:59:07 Several: yes, let's 11:59:14 ACTION Michel: send pointer to [???] 11:59:31 ADJOURN.