Session Start: Thu Dec 07 08:29:43 2000 [08:29] *** Now talking in #daml-dc [08:30] *** mdean has joined #daml-dc [08:30] *** DanC_DC sets mode: +o mdean [08:30] *** ora-in-dc has joined #daml-dc [08:31] present: Deb M, Dan C, Frank vH, Ian H, Ora L, Jeff H, Mike D (chair) [08:37] * DanC_DC logs into daml.org to make the agenda accessible ;-) [08:40] agenda chat (@@link) [08:40] Hefflin's chat on SHOE lessons learned... [08:41] note to self: they use n-ary relations; what was guha's rdfdb scalability point about binary-only-ness? [08:44] DanC: Q: why the variable limitations (cf http://www.cs.umd.edu/projects/plus/SHOE/spec.html#DeclaringInferences) [08:47] example on the whiteboard [@@anybody got a digital camera?] [08:51] Ian H: does SHOE figure out subsumption from rules? (e.g. about Professors and such...) [08:52] ======== [08:53] suppose there's a rule: Professor(x) -> Hated-Person(x) [08:54] and a query: find me ?x such that Professor(x) and Hated-Person(?x) [08:54] you get all the professors. [08:55] but does the system tell you, "did you realize that all professors are hated, and so you just asked for all professors?" [08:57] Lesson: very useful to get intensional answers ala "there cannot be any" rather than "I didn't find any." [08:58] you can use it in several ways: ** suppliment search/query answers, ** query optimization, ** finding related queries. [08:58] ======= [08:59] Q: what's the motivation for DEF-TYPE in shoe? [08:59] hendler: we haven't actually used it [09:00] s\hendler\heflin [09:01] oops ;-) [09:06] ===== [09:07] SHOE folks were reluctant to use URIs as ontology names at first; in hindsight, Hefflin doesn't defend that choice. [09:12] DanC: shoe:extends looks like daml:imports (hefflin: yup); what does revises mean? is it monotonic? [09:21] [...] [09:22] Deb: we can publish [several good ontologies] and have folks choose. [09:23] agenda request: let's talk about the communities around ontologies. [09:26] Deb: it takes a *long* time to realize how valuable, for example, the cyc upper ontology [09:26] general agreement that [hmm... not sure... come back to it later in this meeting] [09:26] ========= [09:27] general agreement that we should (a) focus on infrastructure stuff but (b) actively reach out to the various communities. [09:27] ======== [09:30] Q: you can't have a rule for all transitive properties, can you? [09:30] A: no... but I wonder about a macro [10:11] ===== [10:11] agenda review [10:11] Agenda adds [10:11] -- Concrete domains [10:11] -- Community ontologies, upper/base ontology [10:11] -- unique name assumption [10:11] -- equivalentTo: just for classes? [10:11] -- belief systems/claims [10:11] ============ [10:12] === Agenda adds [10:12] Agenda adds [10:12] -- Concrete domains [10:12] -- Community ontologies, upper/base ontology [10:12] -- unique name assumption [10:12] -- equivalentTo: just for classes? [10:12] -- duplication of presentation and semantic content [10:12] -- belief systems/claims [10:12] ... break. [10:17] *** ora-in-dc has left #daml-dc (ora-in-dc) [10:23] ===== Ora's presentation... [10:24] DC: are these slides for public consumption? [10:24] OL: yeah, sure. [10:27] *** IanH has joined #daml-dc [10:27] * DanC_DC waves to Ian. [10:27] Hi [11:39] standardization is not a good context in which to do design; it's a good context in which to do review [11:40] *** ora has joined #daml-dc [11:57] *** ora has quit IRC [12:18] http://www.daml.org/committee/charter.html [12:33] *** ora has joined #daml-dc [13:06] *** tim-w3m has joined #daml-dc [13:06] Ahhh .... I hoped this channel would be here. [13:06] .nick tim-bos [13:06] *** tim-w3m is now known as tim-bos [13:08] "New members of the committee may be selected by a vote of the current membership" <== noodling on membership rules. [13:08] * DanC_DC waves to tim [13:08] * tim-bos waves to folks in DC [13:09] tim, would you like to join by phone? [13:09] ... for 15 minutes or so? [13:09] By vote of current membership ... cliqueville [13:09] ... to chat about this charter stuff? [13:09] Well, I don't know whther that would be appropriate - on what topics? [13:10] ... ah [13:10] cliqueville. I made that point, but we're not sure what to do about it. [13:10] I could -- but I don't know a good solution - apart from using w3c rules [13:10] ok. If you're inspired, let us know. [13:11] I'd be happy to listen in anyway if there is a bridge [13:11] * tim-bos feels one should listen before talking [13:11] ok... we'll use IRC until it looks like a phone call looks necessary. [13:13] we're considering chartering ourselves as a W3C WG/IG. http://www.w3.org/Consortium/Process/Process-19991111/activities.html#GeneralCharters [13:14] Sounds good. I didn't think votes of commitee were better than program director's say for new emmbers - worse maybe. [13:15] By a copy of te charter or as a real IG? [13:15] huh? I think I mean as a real IG. [13:16] ok [13:16] we'd scrap the vote of the committe and just do a regular W3C CFP, grant the chair the power to invite experts, etc. as usual. [13:16] cool [13:16] there are issues with the size, in that case. [13:17] In fact the work is more like a WG [13:18] but the deliverable is more like a Note, no? [13:18] DAML-O, renamed RDF-O could be Rec trcak ... [13:18] ... in response t some comments at the AC that it was not open to W3C enough [13:18] yeah, but for a few months, we could just work on a Note-track thing, maybe? [13:18] WGs can prodce noets, no? [13:18] s/noets/notes [13:19] this is Jim using Ian's connection for a few minutes... [13:19] Hi Jim (Ian Horrocks?) [13:19] I'm a bit worried that if we throw this open as a w3c effort [13:19] that we will end up swamped by people who will never let us actually [13:19] get to SWELL... is there a way to keeps things focused on [13:20] the logic/ontology for a little while? Note may be solution [13:20] we never really found a good way of limiting the RDFS WG membership [13:20] This is the tension of openness. But foxuss you can completely control in the charter. [13:20] which was an unfrtunate thing... [13:20] hmmm [13:20] s\unfrtunate\unfortunate [13:21] Ora: Was itthe size too big orthe disparity fo view too large, or the turnover too great? [13:22] there were people whose skills/experience was not sufficient for the technical complaxity of the task(s) [13:22] we (read: me and a few others) spent large amounts of time educating others [13:22] The chair could have politely asked them not to take phone time [13:22] and got thus less work done [13:23] When XML started they just said you had to basically know SGML to join -- it was not a good idea. [13:23] But you can require expertise. [13:23] David Singer called one member, got the WG person changed... [13:24] There is a threshold of %effort -- expert effort. [13:24] Also you can set the tone with a few general messages - I am sorry, red this and that rather thanasking that on the group. [13:25] But beware of trying to eliminate smart people with different ideas. [13:25] we tried, it didn't help [13:25] btw... folks agree to invite Peter@lucent to attend. [13:25] tried to set the tone and make people read stuff, that is [13:25] ... good ... though he might refuse if he has to be invited [13:25] ACTION Deb: call Peter tonight. [13:26] Is the feeling that DAML-ont is more or less a done deal and we are talking about DAML-L now? [13:26] i.e. is this the daml-ont or daml-l WG we are consituting? [13:26] more or less done: I think you'll have to join us by phone to get a reasonable answer to that question. [13:26] What do I call? [13:26] tel:+1-703-284-4662 [13:27] brb [13:27] we haven't gotten to the technical stuff on the agenda :-{ [13:27] we had a cool chat around Hefflin's SHOE history slides. [13:28] Then we got into W3C process and charter stuff, and here we are. [13:30] back [13:30] [08:31] present: Deb M, Dan C, Frank vH, Ian H, Ora L, Jeff H, Mike D (chair) [13:30] +Hendler. [13:31] ta [13:32] TimBL: if it's going to have a deliverable, that's pretty much a WG, and you gotta describe what that deliverable is; you'd probably have to rename any existing proposal. [13:32] JAH: it seems things are proceeding faster that the timeline we earlier discussed. [13:33] TimBL: so maybe the DAML-ONT stuff is ready for standardization; but is the logic stuff [...oops.. missed the rest of it] [13:34] DebM: having my participation here funded is important. [13:34] JAH: let's talk about that offline. [13:38] JAH: if you're happy to have this ontology stuff become a W3C working group, perhaps everybody's happy... DARPA declares victory on the ontology layer of stuff, and starts noodling on the rules/logic stuff. [13:39] Maybe we need a ONT WG and an L IG [13:39] rules/logic would be an interest group in the new SW activity. [13:39] [gee... we could even call it www-rdf-logic] [13:41] JAH: once W3C has an official Semantic Web Activity launched, that eliminates the "no official W3C" thorn from earlier. [13:41] JAH: having done this once, I think I can do a better job of avoiding some problems next time. [13:42] TimBL: yeah, and once the community has seen that DARPA can jump-start something and hand it over, they'll trust more the next time around. [13:46] TimBL: yeah, it does seem that the Ontology stuff came together quickly, as there wasn't much novel design/research to do, but I don't see that happening for the logic layer, do you? [13:46] Putthe other way, we can't make the case that we could write a charter for a L WG. [13:49] a possibility: roll the Ontology stuff into the charter of the RDF schema WG. [not issue-free, of course[ [13:49] ] [13:49] ] [13:50] JAH Proposed: it looks like a WG that starts around March 2001 is what we all have in mind; meanwhile, we're doing exploring and doing preparation. [did I get that right?] [13:51] TimBL: there are issues with that too... preparation for a WG can be seen as closing the discussion before it starts. [13:51] * ora nodding to Deb's comments [13:52] Deb: yeah, I'd like to prepare more use cases and such to manage the risk of dropping features like cardinality. [13:56] * tim-bos realzies DanC unhappy but not sure what with -- the consitution ofthe joint committee? [13:57] (Deb had said that she was worried about stad sprocess weakening the spec by sieving out members) [13:57] I'm not unhappy; I just don't think we'll have public text governing the tuesday afternoon telcons soon. [13:57] s/members/features [13:58] * tim-bos lost the phone line [13:58] oops! I hung up; you didn't mean for me to do that? [13:59] no [13:59] * tim-bos mute [14:02] * tim-bos notes choice of chair is crucial in keeping new chair. [14:02] .... in public? [14:03] s/new chair/direction/ [14:11] * tim-bos has half an ear out - and reception comes and goes. [14:11] * tim-bos not to worry but don't assume I have heard everything! [14:12] right. [14:29] * tim-bos has his hand up [14:31] The plishing work is very different from the starting work. Different phases. Maybe different groups. Lots of littel groups, wit the same people, can work. [14:31] * tim-bos wonderes whether there is a projector near the phone or something with a fan. [14:31] [see charter thiny in notepad] [14:32] * tim-bos rrgrgrrngrnggrnrggrgnrggrg rgnrngrgrngr grnrg aah [14:32] btw... [[[[[[[ [14:32] PROPOSED: [14:32] 1. A working group will be constituted under the W3C to develop a spec [14:32] for ontolog sorts of stuff (cardinality, functional, ...) [14:32] a. ACTION JAH with DC: draft a charter. [14:32] 2. An interest group, probably www-rdf-logic, will be constituted to [14:32] play with logic stuff: rules, variables, proof, ... [14:32] b. ACTION Dan: draft a charter for this IG for review by JAH et. al. [14:32] 3. A joint US/EU committee intends to develop a proposal to www-rdf-logic. [14:32] The present company is likely to participate in [14:33] this. No charter necessary; [14:33] let's take the existing documentation and call it just a "description" or [14:33] something. [14:33] a. But we are re-constituting ourselves. While we're at it, let's state [14:33] expectations while we're inviting everybody to rejoin. [14:33] RESOLVED. [14:33] ]]] [14:33] ============ onward, technical issues. [14:33] agenda revisit... [14:35] "_presentation_ syntax"? [14:36] I dunno what that is either. [14:36] "something more suitable for emailing around" -- Frank H [14:36] * DanC_DC recalls notation3 [14:40] presentation by Ian H. [@@link to slides] [14:41] MikeD: when did you start? [14:41] IanH: we just had our 1st birthday; 1 year old. [14:46] slides: http://www.cs.man.ac.uk/~horrocks/Secret/LogicWorks/LuxembourgSemWebStuder.ppt [14:46] PDF: http://www.cs.man.ac.uk/~horrocks/Secret/LogicWorks/DAML-washington-00.pdf [14:47] Ian H: this reasoner isn't integral to oil. [14:47] DanC: but the reasoner is complete? i.e. the underlying logic is limited? [14:47] Ian H: yes. [14:47] Ian H: reasoning is very important to support ontology design. [14:48] ... at design time, you can afford to spend time on reasoning that you might not want to spend at query time. [14:48] *** tim-lap has joined #daml-dc [14:49] * tim-lap missed the URI if you pasted it - tracking it crashed the desktop [14:49] PDF: http://www.cs.man.ac.uk/~horrocks/Secret/LogicWorks/DAML-washington-00.pdf [14:50] *** tim-bos has quit IRC (Ping timeout) [14:50] Ian H: this "kludge" around individuals is necessary to support complete/efficient reasoning. [14:59] JAH: Mike, we should let these (OIL etc.) folks look at the tools wish list. [15:02] another EU research project: WonderWeb http://www.cs.man.ac.uk/~horrocks/WonderWeb/ [15:06] OilEd demo... [15:06] Dean: is this based on the protoge editor? [15:07] Ian H: it borrows a lot of interface, but the code is our own. [15:07] * tim-lap assumes that this isn't a web thing. [15:08] nope; java on a laptop here. [15:08] Jeff H: are the users you're working with knowledge engineers, or domain experts? [15:08] Ian H: some of each. [15:08] Jeff H: we had similar experience with the FDA; there was education in both directions. [15:08] Ian/Frank: yeah.. [15:35] * tim-lap resolves to snd a laptop wiht a camera for the jokes [15:38] {{ x workson [ a researchProject] } implies { x a researcher } } log:forall x. [15:40] tim, how did you know what was on our screen? [15:41] Aaaahh. I took the liberty of sending a small automaton ... no, it was also discussed. [15:41] ah. I see. [15:42] But i still wonder about the trademark amusement [15:42] ! [15:42] The tool did this "FaCT in side" logo complete with earcon. [15:43] :) [15:43] wrapping up this demo, chosing the next agenda item... [15:46] * tim-lap realises he has a sort of Dan's-ear view of this meeting. [15:46] ======= presentation syntax [15:47] http://www.cs.vu.nl/~frankh/spool/DAML-OIL-syntax/ [15:47] http://www.cs.vu.nl/~frankh/spool/DAML-OIL-syntax/presentation-syntax.pdf [15:49] Aaaagh ... it i smostly text but I can't cut and paste it! [15:49] I wanted to convert the examples into N3 [15:50] there are example files in that ...syntax/ directory [15:51] african animal example... [15:51] I takie that back -- can cut out of .pdf [15:51] paste buffer BTW contained: :granpa :ancestor :pa, [15:51] :bill . [15:51] no can't get it... [15:53] [15:53] Class "elephant" [15:53] subClassOf "# animal" [15:53] RestrictedBy Restriction "# eats" toClass "# plant� [15:53] RestrictedBy Restriction "# colour" hasValue EQUAL ``grey�� [15:53] Class "elephant" [15:53] subClassOf "# animal" [15:53] RestrictedBy Restriction "# eats" toClass "# plant� [15:53] RestrictedBy Restriction "# colour" hasValue EQUAL ``grey�� [15:53] This example is not in that [15:53] directortu [15:53] oops [15:56] <#elephant> a rdf:Class; rdfs:subClassOf <#animal>; daml:restrictedBy [ a daml:Restriction; daml:onProperty <#eats>; daml:toClass <#plant> ]; [16:00] tim, do you have daml-ex in .n3 handy? [16:00] now we're looking at daml-ex in this proposed syntax: http://www.cs.vu.nl/~frankh/spool/DAML-OIL-syntax/daml-oil-ex.html [16:02] Yes ... [16:02] I have daml-ont.n3 [16:04] http://www.w3.org/2000/10/swap/test/daml-ont.n3 [16:05] Try: http://www.w3.org/2000/10/swap/Primer [16:08] This is Ian back again - I tried that URL but it was "not found" [16:09] 404 at http://www.w3.org/SyleSheets/base.css [16:09] I know that one... [16:09] S^tyle [16:09] ta [16:10] * DanC_DC turns stylesheets off... [16:10] http://www.w3.org/2000/10/swap/Primer [16:11] ^R on http://www.w3.org/2000/10/swap/Primer [16:12] It doesn't do bags. [16:13] tim, we're looking at http://www.w3.org/2000/10/swap/Primer now [16:17] http://www.w3.org/DesignIssues/Notation3.html [16:17] is the bNF [16:26] * tim-lap can't hear - everybody talking [16:26] * tim-lap that's ok but questions have to be in irc [16:30] [[[ [16:30] bind : ...daml... . [16:30] bind rdf: <...rdf...> . [16:30] <#AfricanAnimal> restrictedBy [16:30] [ rdf:type Restriction; [16:30] onProperty <#comes_from>; [16:30] hasValue [ unionOf [16:30] ([ oneOf <#Africa>], [16:30] [ rdf:type Restriction; [16:30] onProperty <#is_part_of>; [16:30] hasValue <#Africa>; ] ) ]. [16:30] ]]] [16:31] # Base URI of process is file:///devel/WWW/2000/10/swap/test/ [16:31] # Input from file:/devel/WWW/2000/10/swap/test/daml-ex-1.n3 [16:31] # Notation3 generation by [16:31] # $Id: 2000-12-07-irc.txt,v 1.2 2000/12/12 18:11:20 mdean Exp $ [16:31] @prefix : . [16:31] a :Ontology; [16:31] :versionInfo """$Id: daml-ex.daml,v 1.2 [16:31] 1:17 connolly Exp $"""; [16:31] :comment "An example ontology"; [16:31] :imports . [16:31] ... a :Class; [16:31] :label "Animal"; [16:31] :comment """This class of animals is i [16:31] number of [16:31] ontological idioms.""" . [16:31] a :Class; [16:31] :subClassOf . [16:31] a :Class; [16:31] :subClassOf ; [16:31] :disjointFrom . [16:31] a :Property; [16:31] :domain ; [16:31] :cardinality "2" . [16:32] [[[ [16:32] <#AfricanAnimal> restrictedBy [16:32] [ onProperty <#comes_from>; [16:32] hasValue [ unionOf [16:32] ([ oneOf <#Africa>], [16:32] [ onProperty <#is_part_of>; [16:32] hasValue <#Africa>; ] ) ]. [16:32] ]]] [16:33] it didn't make into a qname. [16:33] ===== [16:33] It will if you give it a prefix. It doesn't invent one [16:34] Frank: top two items for tomorrow: language release, open issues (concrete types, ...) [16:34] @prefix f: . [16:52] My parsing of daml-ont came up against the implicit parsetype difference between RDF and DAML [16:56] you mean parseType="daml:collection"? [16:56] ==== Deb M chimera demo [16:57] Chimaera, actually [17:02] Given that I can't see the demo, and it is getting late I will drop off. I may revisit IRC later. [17:02] I meant that if it hits a start tag when expecting a plain value it stops. [17:03] there is an implict parseType="Resource" in DAML? [17:03] bye, time. [17:04] bye, Tim. [17:54] Logging off - bye guys [17:55] *** IanH has left #daml-dc (IanH)