W3C Web Ontology Working Group IRC Chat Logs for 2001-12-06

16:00:05 Users on #webont: logger_2 @DanC @dajobe

16:56:02 <nmg> evening

17:03:49 <las> JAH: Need a scribe.

17:03:50 <DanC> 19 on the bridge per http://www.w3.org/1998/12/bridge/Zakim.html

17:03:55 * las volunteers

17:04:08 <las> We are waiting for quorum, then will take attendance.

17:04:19 <las> JAH: When you enter, please say your name as you enter...

17:04:38 <las> I am not recording everyone's names as they join, will do so in attendance.

17:05:42 <las> James Burnett - no

17:05:47 <francesco> hello, here are francesco iannuzzelli and frederick brysse from ivisgroup, london

17:06:09 <las> Burnett - excused

17:06:13 <las> Breen - absent

17:06:17 <las> Brisa - present

17:06:21 <las> Boswell - present

17:06:25 <las> Carroll - absent

17:06:29 <las> Connolly - present

17:06:32 <las> Dale - absent

17:06:36 <las> de Roos - present

17:06:44 <las> Deruhr - regrets

17:06:46 <las> Dean - present

17:06:49 <las> Decker -p[resent

17:06:51 <las> Finin - present

17:06:57 <las> Gibbons - present

17:07:02 <las> Hayes - present

17:07:06 <las> Heflin _present

17:07:08 <las> Hendler present

17:07:12 <las> Helman -absent

17:07:14 <las> horrocks present

17:07:17 <las> hurley absent

17:07:22 <las> inazelli present

17:07:26 <las> jeckyl absent

17:07:28 <las> klein present

17:07:37 <las> kohasa ? absent

17:07:41 <las> kratsova absent

17:07:44 <las> lassila regrets

17:07:48 <ora> ora is on IRC but cannot join on the phone yet

17:07:53 <las> menci absent

17:07:57 <las> mcguinness present

17:08:01 <las> miller regrets

17:08:05 <las> oberst present

17:08:10 <las> patel schneider present

17:08:15 <las> pike excused

17:08:17 <las> saboo present

17:08:28 <las> schreiber regrets also shimizu

17:08:37 <las> shintek absent

17:08:44 <las> mike smith ned smit both here

17:08:50 <las> ned needs to leavfe early

17:08:53 <las> stanton present

17:08:58 <las> stein present

17:09:04 <las> stickler absent

17:09:14 <las> werner ten keit and ten horst present

17:09:42 <las> thompson present

17:09:46 <las> trasper absent

17:09:53 <las> van harmeln present

17:09:57 <las> laxman benegale absent

17:10:02 <las> voles absent

17:10:09 <las> dieter fensel absent

17:10:33 * las promises next time (if she knows she's scribing) to have the list in front of her!

17:10:48 <las> JAH: agenda change (given soem people need to leave early)

17:11:17 <las> JAH: First agenda item is volunteers for document editing/use cases. These seem to be converging, so let's divvy up the job.

17:11:39 <las> JAH: We got a bunch of use cases, categorized them, .... Jeff Heflin came up with three that JAH has renamed:

17:12:01 <las> JAH: Resource management, interoperability, agents.

17:12:18 <las> JAH: THose cover most of the use cases, plus a fourth that cut across use cases/requirements.

17:13:21 <las> JAH: We need four volunteers to head each area (plus additional volunteers in each area) to get together and propose how to outline use cases, identify critical issues. Their action: within next 2 weeks, need to produce a 1-2page outline of each use case area, what they're going to concentrate on.

17:13:55 <DanC> for future reference, a handy way to do this is to copy the WG membership list from the web page into a text editor, then write Y/N/R/E at the beginning of each line

17:13:59 <las> Subsequent action item will be to produce descriptions of these use cases in more detail.

17:14:01 <DanC> (this = roll call)

17:14:23 <las> Roll call: thanks, but i didn't have time to call up the list of attendance!

17:14:43 <las> LAS: Timetable isn't aggressive enough to get it done in time.

17:15:04 <las> JAH: Will use whatever timetable we can, want final report for mid-Jan.

17:15:18 <las> JAH: Gus Schreiber has volunteered to coordinate collection management.

17:15:24 <las> DanC: DId he say by when?

17:15:31 <las> JAH: Not his job to produce, just to coordinate.

17:15:41 <las> DanC: Disagree....need to write up.

17:15:44 <las> +Ora

17:15:57 <las> JAH: No, need to coordinate among possible use cases.

17:16:06 <las> DanC: Can't produce draft on what's gone by?

17:16:22 * las missed what Pat says.

17:16:36 <las> JAH: Maybe. I suggested one week.

17:17:00 <las> Leo: I will do interoperability.

17:17:07 <las> JAH: Web services?

17:17:16 <las> Stephan: I will do web services.

17:17:21 <las> JAH: CRosscutting technologies.

17:17:29 <las> Pat Hayes: Say a bit more?

17:17:46 <las> JAH: In the email. Versioning, .... things all might want to take advantage of, but not requirements of a specific area.

17:17:50 <las> JAH: Jeff?

17:17:56 <las> Jeff Heflin: OK.

17:18:03 <las> Deborah McG: I'll do w/Jeff.

17:18:13 <las> JAH: Leads on all four.

17:18:19 <mdean_> I'll be speaking from slides at http://www.daml.org/2001/11/webont-tour/Overview.html

17:18:24 <las> ACTION: Leads will produce outline of area as above.

17:18:43 <las> ?: If we want to participate, should we just contact the leads?

17:19:53 <las> DanC: Yes, contact coordination people. There's a mailing list called www-archive@w3.org....If you cc to this, it creates a web-visible place that can also be made public...it's a way to have private conversations that can be made public (or you can cc www-webont-wg).

17:19:58 <las> JAH: Who wants to be involved?

17:20:06 <las> JAH: collections

17:20:12 <las> Mike Smith

17:20:17 <las> Steven Boswell

17:20:20 <las> Nick Gibbons

17:20:31 <las> JAH: interoperability of sources

17:20:40 <las> Pat Hayes

17:20:44 <las> Dan Connolly

17:20:45 <las> Mike Dean

17:20:52 <las> Herman ter Horst

17:20:58 <las> JAH will keep an eye on all.

17:21:04 <las> JAH: web services/agents

17:21:08 <las> Ora Lassila

17:21:10 <las> Tim Finin

17:21:14 <las> Lynn Thompson

17:21:17 <las> Mike Dean

17:21:32 <las> JAH: technology and...

17:21:49 <las> Jeff: general requirements/supporting requirements

17:22:02 <las> Deborah: I have a list of requirements for ontology based applications.

17:22:06 <las> Ned ?

17:22:12 <las> Jeremy Carroll

17:22:16 <las> Pat Hayes

17:22:21 <las> Dan Connolly

17:22:33 <las> DanC tries to volunteer Jos de Roos.

17:22:41 <las> assented?

17:22:45 <DanC> I think so.

17:22:55 <DanC> Jos?

17:23:02 <las> JAH: anyone else want to volunteer?

17:23:21 <las> Mike Dean takes over....

17:23:52 <las> Go to www.daml.org and do a search on webont (or go straight to http://www.daml.org/2001/11/webont-tour/Overview.html)

17:23:58 <las> Not yet in the briefings.

17:24:02 <JosD> yes Dan,great

17:24:17 <DanC> indeed: http://www.daml.org/2001/11/webont-tour/Overview.html

17:24:41 <DanC> Mike, keep in mind not everybody here speaks English. Slowly and clearly, please.

17:25:22 <DanC> slide 2: http://www.daml.org/2001/11/webont-tour/slide2-0.html

17:25:49 <DanC> Mike, is that right? anybody can join the joint-committee discussion?

17:26:24 <las> Darpa Agent Markup Language funded by US DOD for past 15 months....daml.org is website for program but open to everyone.

17:26:25 <las> From the overview, click on DAML+OIL....language specific stuff.

17:26:29 <las> OIL is largely european technology

17:26:29 <las> DAML+OIL currently maintained by the joint committee (homepage on daml.org)

17:26:29 <las> telecon logs, etc.

17:26:29 <las> one slight caveat....email list is limited to members (spam reduction) but non-member traffic can be sent with a moderator delay.

17:26:32 <las> Next slide details about darpa-funded research teams....(Slide 3-0, DAML program information)

17:27:06 <DanC> Jim? pls be sure pointers to presentation materials are in the agenda from now on

17:27:30 <DanC> "IOW"

17:27:31 <DanC> ?

17:27:39 <DanC> where is he?

17:27:48 <DanC> I speak English and I can hardly follow. Mike? hello? can you see this?

17:29:14 <las> Mike took us to the program participants page, a list of all the funded projects. Now we're back on the 3d slide with a list of info about different aspects of the project.

17:29:28 <las> Still on slide 3-0 (info about the DAML research program): DAML services.

17:29:44 <las> There is a new release of DAML-S as of yesterday. This will be relevant to those interested in agents.

17:29:59 <las> There is a list of information on new events in the DAML program.

17:30:16 <las> JAH: Webont participants probably want to subscribe to Hot DAML newsletter

17:30:37 <las> JAH: low volume (every 6 weeks). Also all welcome to contribute to the newsletter.

17:30:57 <las> mdean resumes: Next slide 4-0, Daml ontology library.

17:31:14 <las> Clearinghouse for ontologies people wish to public.

17:31:31 <las> A lot were seeded by participants int he DAML program, sometimes translated from other ontologies.

17:31:36 * las missed how many, how large

17:31:45 <las>http://www.daml.org/2001/11/webont-tour/slide4-0.html

17:32:03 <las> This slide has some examples, e.g., how many include info on professors.

17:32:17 <las> JAH: I've heard recently that some submitted ontologies aren't showing up. How do you add one?

17:32:34 <ora> unfortunately I have to leave the telecon now (short visit :-)

17:32:36 <las> Mdean: There is a submission page for the crawler separate from the ontology library....you need to do both.

17:32:40 <las> -ora

17:33:03 <las> mdean: 24 hour delay for a library to appear in the website.

17:33:06 <DanC> Hmm... is the stuff that gets collected from http://www.daml.org/ontologies/submit.html available in DAML? can I submit in DAML?

17:33:19 <DanC> "the bottom there"? where?

17:33:34 <las> mdean: we try to gather some statistics on the ontologies....http://www.daml.org/ontologies/features shows a summary.

17:33:43 <las> "the bottom" of slide 4.

17:33:50 <DanC> right..thx

17:34:09 <las> Moving on to 5-0 about the DAML crawler....sort of like a google for daml....collects daml content from sites (currently 45 sites)....

17:34:32 <las> If you submit to the ontology library, it's automatically harvested by the crawler (so you don't need to submit both)

17:35:06 <las> The collected results are available for other projects to use (pointed to from this slide, http://www.daml.org/2001/11/webont-tour/slide5-0.html)

17:35:15 <las> Also querying (some egs on slide)

17:35:32 <las> Next slide, DAML tools

17:35:43 <las> first link is a set of pages that list tools developed to deal with DAML.

17:35:55 <las>http://www.daml.org/tools#project

17:36:14 <las> Also a tools wishlist

17:36:46 * las thinks that this particular way of Mike preparing the slides in advance and me trying to retype them into the minutes is less than fully effective.

17:37:10 <las> Wishlist is tracked to know whether things are being implemented. Submissions welcome to tools wishlist.

17:37:22 <las> 7-0 DAML applications http://www.daml.org/2001/11/webont-tour/slide7-0.html

17:37:31 <las> If I have a language like this, what can I actually do with this?

17:37:43 <las> Ultimately want this to evolve into use cases, what worked and what didn't work, ....

17:37:54 <las> Design patterns wanted....

17:38:27 <las> Helps if it's a web-based app someone can actually go to, but internal apps are also useful if people can learn from the idea rather than playing with the implementation.

17:38:48 <las> http://www.daml.org/2001/11/webont-tour/slide8-0.html is about some of the specific tools that might be useful for folks in this group.

17:39:13 <las> DAML validator takes a page w/DAML content and checks for consistency w/the ontology it was developed in. Like XML validator, sort -of.

17:39:37 <las> Checks cardinality constraints, looks for class definitions, ....

17:40:15 <las> There are a couple of viewers for DAML contents listed on this slide including one in Java and one for the palm.

17:40:55 <las> A couple of tools from ontology library will show you what's in an ontology....dumpont will show you what's in an ontlogy, hyperDAML provides links to definitions for any instance. Examples on pages referenced here.

17:41:13 <las> While I have a captive audience, I thought I'd talk about a couple of tools that are work in progress (slide 9-0)

17:41:21 <las>http://www.daml.org/2001/11/webont-tour/slide9-0.html

17:42:00 <las> Dynamic models: most pages take DAML input and put it in memory or dump to knowledge base to query....here i am exploiting possiblity of dynamic retrieval...couple tools here based on that.

17:42:12 <las> very simple to do, hopefully of general use.

17:42:34 <las> another that's not on the page yet is something that uses monarch graph to build the graph structure by clickling on object dynamically.

17:42:43 <las> (alot of tools using jena)

17:43:09 <las> Jeremy Carroll from HP: Planning a new release in Jan, get in touch if you want your stuff integrated....

17:43:25 <las> JAH: New JENA release would be great to put into HotDAML

17:43:35 <las> Jeremy: Design review, stopgap

17:44:02 <las> mdean: found jena very useful. toolkit developed by hplabs. best java api for rdf/daml content. next release will be even better.

17:44:10 * DanC wonders if Hendler is here in #webont; nope. no wonder he was getting ahead of the scribe during roll call

17:44:24 <las> html front end is for showing what daml content will look like when it's integrated into web pages.

17:45:01 <las> going to nyse site, want to do airport codes.

17:45:39 <las> daml agenda is a very simple app to describe a meeting....big thing is uses xml schema data types to compute sequence of meeting, times for the presenters....

17:46:01 <las> first app I know of that does computation w/xml data types.

17:46:03 <las>http://www.daml.org/2001/11/webont-tour/slide10-0.html

17:46:24 <las> more information: search capability (should add search engines) and roadmaps for selected audiences.

17:46:50 <las>http://www.daml.org/2001/11/webont-tour/slide11-0.html

17:47:06 <las> JAH: It would be very useful to put this presentation within a single click of the homepage.

17:48:16 <las> LAS: Scribing for someone who's not looking at the channel is hard.

17:48:35 <las> DanC: focus on the Q&A; important ideas should be in the presentation.

17:48:40 * las wishes she'd known that.

17:48:50 <DanC> sorry, lynn

17:48:53 <las> mdean: shared screen technology?

17:48:58 <las> danc: not necessarily.

17:49:05 <las> JAH: Is this chat logged?

17:49:10 <las> DanC: yes.

17:49:26 <las> JAH: Those of you who want to see the chat afterwards can get to it from the webont home page.

17:49:43 * las also wishes she'd known she was scribing *before* the meeting as she'd have gotten some stuff ready.

17:50:17 <las> Frank VH: Go to google, type in daml+oil, get walkthrough (or go straight to ) http://www.daml.org/2001/03/daml+oil-walkthru.html

17:50:45 <las> (all Frank unless indicated): I will go through highlights, not line by line.

17:51:03 <las> Walkthrough is an english-annotated version of a daml-code file also available from the parent page.

17:51:22 <las> Jump to first section, setting up namespaces. What's intro'd there is a shorthand for long urls.

17:51:47 <las> Jump to housekeeping, meta-information about the ontology (name, version, ...)

17:52:03 <las> import statement includes other ontologies into this one....there's no formal definition yet (problem).

17:52:13 <las> Real Stuff starts in next section, called Classes.

17:52:23 <las> One of key modelling mechanisms is classes

17:52:48 <las> First statement defines a class called animal. Animal is really a local URI (i.e. a uri local to this document).

17:53:03 <las> Every class in a daml ontology has a uri on the web.

17:53:12 <las> comments have no formal meaning, but could be useful to human readers.

17:53:34 <las> ?: Are these labels human-readable?

17:53:45 <las> ?: How do I know that they are English?

17:53:54 <las> DanC: There's a mechanism, but it's problematic.

17:54:20 <las> Frank: First example just names the class; simplest way to define. Next example uses a subclass to define....uses a URI to refer to other classes.

17:54:43 <DanC> re language: cf rdfms-xmllang: Why isn't xml:lang information represented within the RDF data

17:54:43 <DanC> model? http://www.w3.org/2000/03/rdf-tracking/#rdfms-xmllang

17:54:46 <las> Third example (female) is similar to male, but also has a requiremtn that it's disjoitn with male.

17:55:00 <las> Class Man has two superclasses (multiple inheritance).

17:55:05 <DanC> lynn, you probably needn't minute when he reads from the document.

17:56:15 <las> JAH: Man = subclass of person, male vs. man = person whose sex is male.

17:56:23 <las> Ian: Standard KR issue.

17:56:36 <las> JAH: Just want to say that this is *a* way of defining male, not *the* way.

17:57:01 <las> Frank: Interesting to note that man is defined as a subclass of person without person being defined....that's fine.

17:57:15 <DanC> [which it says in the document ;-]

17:57:21 * las is having trouble telling when Frank's reading and when he's extrapolating.

17:57:34 <las> PJH: the word "defines" worries me....

17:58:13 <DanC> ----- section: Defining Properties

17:58:15 <las> Frank: Yes, we're stating properties about these classes, but no claim is made that these are the only properties or definitional properties....this is intentional so we can define some aspects here and other aspects elsewhere.

17:59:44 * las is trying not to type the parts of this presentation that are contained in what she modestly thinks is a pretty darn good document (http://www.daml.org/2001/03/daml+oil-walkthru.html)

18:01:18 <las> Frank: datatype properties separate from resource-valued properties.

18:01:24 <las> DanC: Some of us don't like this.

18:01:43 <las> ?: Is it the case that domain can be unspecified, eg. a house could have a shoe size?

18:01:47 <las> A: Yes.

18:02:19 <las> JAH: RDF Core group is grappling with datatype properties. When they've concluded, we need to wait until they've come up with their decision and see if we want to adopt it.

18:02:31 <las> PJH: IT's turning out to be a much bigger problem than anticipated.

18:02:41 <las> ?: Is unique property a subclass of datatype property?

18:02:52 * las apologizes for not catching names....

18:03:01 <las> Frank: In this case, shoe size is a unique property.

18:03:17 <las> Same ?er: I understand what it's saying, but not how it's saying it.

18:03:37 <DanC> ------ Defining property restrictions

18:03:37 <las> Frank: Want to take the question off-line. It's complicated.

18:07:01 <las> ...description logic idiom....defining restriction, implicitly defining unnamed class of all things that meet this restriction, then saying this (animal) is a subclass of the implicitly defined class.

18:07:11 <las> DanC: also an artifact of trying to wedge into XML syntax.

18:07:59 <las> Frank continues....some people have more than one shoe size, apparently.

18:08:38 <las> ?: Class definitions in DAML+OIL, are they defined as in description logics or simply called out?

18:08:48 <las> Frank: That's what Pat asked: NOT sufficient, just necessary.

18:09:51 <las> Ian: As well as subclassOf, there's sameClassAs....so you can either make it a subset or definitionally the same as. You can use as many as you want of either. In Description Logics, you can have only one subclass or sameclass.

18:10:17 <las> Frank: There's a similar example later on w/SameClassAs, which has the effect Rudiger (sp?) was alluding to.

18:10:26 * las apologizes again.

18:11:31 <IanH> What I said was that in "old fashioned" DLs you can only have one axiom per name - "modern" DLs are like DAML+OIL in this respect.

18:11:56 * las missed the question.....

18:12:57 <las> Pat: The language is defined carefully so that it's impossible to put constraints from the daml language on the objects in the domain and range.

18:13:13 <las> ?: To say that two things are symmetric, could you use inverseOf (property is its own inverse).

18:13:16 <las> Frank: Yes.

18:13:35 <las> Same ?: How would you say that a property is reflexive?

18:13:40 <las> Frank: I don't think you can.

18:14:35 <las> JAH: WOuld like to let Frank keep going, minimize questions. Next week we can take questions about the language for Frank/Ian. (Time issues).

18:14:51 <las> ------

18:14:52 <las> Notations for classes

18:14:59 <timfinin> The "missed question" was mine. I asked if uniqueProperty worked for both bjectproperties and dataproperties.

18:15:14 * las thanks timfinin :o)

18:16:02 <IanH> Answer to Tim's question is yes

18:16:27 * las thanks ianh, too!

18:17:06 <las> ?: Sometimes I'm confused about when to use about

18:17:44 <las> Frank: Important to realize that you can have statements about the same thing in many different places in the file or in multiple files. There is a notational convention that the "first" or "primary" mention uses "id=" and others use "about=".

18:17:56 <las> same ?er: So it's just for human readers?

18:18:07 <las> Frank: Yes. I'm not even sure that in the SW it makes sense to do.

18:18:17 <las> same ?: and it's not clear what primary means....

18:18:36 <las> ------ Using User-defined Datatypes

18:18:49 <las> Main point: You can use xml schema data types.

18:19:12 <las> Skipping details....

18:19:27 <las> -----

18:19:27 <las> Defining individuals

18:21:00 * las again missed the sense of the question....

18:21:13 <las> Frank: You could say that the class is built as the disjoint union of the other classes.

18:21:35 <las> DanC: There isn't any technical limitation to putting classes inside asClasses.

18:21:53 <las> Ian: Then you're stepping outside the defined semantics.

18:22:05 <las> Frank calls the question, continues with the individuals.

18:22:07 <DanC> wow. I didn't know that. There sure are a lot of restrictions in DAML+OIL that I've been blissfully unaware of.

18:22:55 <las> JAH: Let me point out that the language allows using things (properties, classes) across files and ontologies, but then we may not always be able to determine what it ought to mean. They're all just uris, after all.

18:23:23 <las> JAH: Looking in the data in the DAML website, people do this without using import, just by referencing other uris.

18:23:53 <las> Frank: Also my slide set succinctly summarizes the highlights (pointer in email, also http://www.cs.vu.nl/~frankh/spool/DAML+OIL-overview.pdf

18:24:37 <las> JAH: Chair's note....if you go to the DAML reference document, tehre are a number of constructs not in the walkthrough. Is the language in the walkthrough sufficient, or do our use cases require additional machinery?

18:24:59 <las> JAH: We need to say what we think is the right starting point for us.

18:25:15 <las> JAH: Thanks Frank, thinks we had main conversations we needed,

18:25:31 <las> JAH: Those in use case areas, start exchanging mail to figure out how you're proceeding.

18:25:45 <las> ACTION: CHairs begin conversations w/deliverable next week of draft of how you're proceeding.

18:25:51 <las> Guus sent to the mailing list a suggestion.

18:25:58 * las wonders whether the action item is accurate.

18:26:24 <las> JAH: All working group members: Walkthrough came from member request. Please request other things that'd be useful.

18:26:42 <las> Pat: I'd like to hear from the "newcomers" what they'd like to hear,

18:26:49 <las> JAH: We're all newcomers.

18:27:05 <las> ?: Are drafts on our website member confidential?

18:27:19 <las> DanC: The proceedings of this group are public.

18:27:38 <las> JAH: Next meeting one week hence. Adjourned.

18:27:45 <las> DanC: allocate scribe now?

18:27:54 <las> JAH: I remove adjournment until scribe volunteer.

18:28:00 <las> DanC: Please?

18:28:13 <las> Connolly volunteers in the absence of anyone else.

18:28:19 <las> JAH: Thanks, Dan. G;bye all,

18:30:24 * las wonders what her further duties are wrt the irc log/minutes....