ADMIN: minutes for Mar 7 telecon

Minutes of the WebOnt teleconf on 7 March 2002, 12:00 to 13:25 EST
By Peter F. Patel-Schneider


ITEM 1) Join call/attendance/admin (10 min)


ALT	James Barnette
Regrets	Jonathan Borden
Absent	Einar Breen
ALT	Frederik Brysse
YES	Stephen Buswell
YES	Jeremy Carroll
YES	Dan Connolly
Regrets	Jonathan Dale
YES	Jos De Roo
Regrets	Mike Dean
Absent	Stefan Decker
ALT	D.C. DeRoure
YES	Larry Eshelman
Absent	Dieter Fensel
Late	Tim Finin
Regrets Nicholas Gibbins
Absent	Patrick Hayes
YES	Jeff Heflin
Late	Ziv Hellman
Regrets James Hendler
YES	Ian Horrocks
Absent	Oisen Hurley
Regrets	Francesco Iannuzzelli
ALT	Mario Jeckle
YES	Ruediger Klein
ALT	Michael Kohlhase
ALT	Natasha Kravtsova
YES	Ora Lassila
ALT	Alexander Maedche
YES	Deborah McGuinness
YES	Libby Miller
Regrets	Enrico Motta
YES	Leo Obrst
Regrets	Laurent Olivry
YES	Peter Patel-Schneider
ALT	Martin Pike
Late	Marwan Sabbouh
YES	Guus Schreiber
Late	Noboru Shimizu
YES	Michael Sintek
YES	Michael Smith
ALT	Ned Smith
YES	John Stanton
YES	Lynn Andrea Stein
ALT	Patrick Stickler
Yes	Said Tabet
ALT	Warner ten Kate
YES	Herman ter Horst
Late	Lynne R. Thompson
YES	David Trastour
YES	Frank van Harmelen
	Laxman Venigalla
Late	Raphael Volz
YES	Evan Wallace

ACTION Chairs: update ftf schedule on group home page.
	CONTINUED: fix errors

ACTION ALL: register for A'dam F2F meeting (also in case of regrets)
	CONTINUED: URGENT: only 16 registered as of noon EST
			   only 24 registered as of 13:30 EST
ACTION Dan Connolly: set up this mailing list

ACTION on editors of reqdoc: include mailing list in doc
	DONE: by Dan Connolly

ACTION on editors of reqdoc: verify W3C guidelines on syntax & structure

ACTION on Mike Smith: setup repository of "open issues"

ACTION ALL: register interest in focus groups:
	CONTINUED: only 8 missing

ACTION: Jeremy Carroll to propose labels for focus areas

ITEM 4) Format of discussions in focus areas (chair, 5 min)

 - Each telecon will have agenda item for each area
 - Focus in each telecon on 1-2 areas (depending on issues arising)
 - Only additional telecons if absolutely needed, on ad-hoc
   basis. These will be open to all WOWG members

Some discussion, particularly that telecons for the focus groups may allow
better participation, but basically this was accepted as a way to go, at
least for now.

ITEM 2) LANG: OWL lite (van Harmelen, 30 min)


 - short overview of the proposal, including rationale
 - Q&A session
 - discussion (only initial, bulk expected after telecon on mailing

Frank van Harmelen presented the proposal.  Discussion on what was hard /
easy, what comes out of frames.

ACTION Frank van Harmelen: initial proposal for frame part of language in 2

ITEM 3) TEST: Introduction to the role of test cases (Carroll/De Roo, 20 min)

Relevant input to the discussion about test cases at today's telecon is
found in Jos's and Dan's recent messages:

 - role of test cases in RDF (see the links provided by Carroll and De Roo)
 - discussion about role of test cases for OWL

Discussion on what kinds of tests to use, what they are good for, and how
to proceed.

Sources of test suites:
Deborah McGuinness -
Ian Horrocks - FaCT tests
Peter F. Patel-Schneider - CLASSIC - 

Systems that may want to try out test suites:
FaCT, RACER, DLP, [Tim Finin?], [Dan Connolly], JTP

ACTION Dan Connolly: produce several tests in DAML+OIL syntax
ACTION Jos De Roo, Ian Horrocks, Deborah McGuinness, Tim Finin: run these tests
ACTION Jeremy Carroll: look up different test suites and categorize
ACTION Guus Schreiber: turn CM examples into tests

Deadline for the above action items was not precisely given, but all agreed
to aim to be done by next week.

ACTION: John Stanton: report back on different kinds of tests (very short) - next week

5) A.O.B. (5 min)

Agenda items for next week:

SEMANTICS - layering possibilities - Peter Patel-Schneider (agreed) and Pat
	  Hayes (will be contacted by Guus Schreiber)

IRC Log:

16:01:04 <logger_2> logger_2 has joined #webont
16:01:04 <> Users on #webont: logger_2 @las 
16:45:02 <connolly> connolly has joined #webont
16:45:16 <connolly> connolly has changed the topic to: WebOnt 7Mar
16:45:20 <connolly> connolly has changed the topic to: WebOnt 7Mar; scribe?
16:55:03 <IanHorrocks> IanHorrocks has joined #webont
16:58:33 <ora> ora has joined #webont
17:04:30 <pfps> pfps has joined #webont
17:04:41 <DeborahMc> DeborahMc has joined #webont
17:05:04 <libby> libby has joined #webont
17:05:37 <JosD> JosD has joined #webont
17:05:52 <connolly> * connolly finishes up some tech report publication business...
17:05:56 <connolly> connolly is now known as DanC
17:08:49 <DanC> agenda:
17:08:49 <JosD> JosD has quit
17:09:16 <pfps> chat for now is open projects
17:09:19 <pfps> actions:
17:10:40 <DanC> 16 registered (+2 regrets) per
17:11:21 <DanC> done:
17:12:17 <timfinin> timfinin has joined #webont
17:12:22 <las> * las hadn't realized that regrets were to register as well....apologies (and regrets, but those are registered formally).
17:12:54 <DanC> pfps, I gather you're scribing today?
17:13:22 <pfps> yes, but the actions are hard to scribe into the chat
17:13:41 <JosD> JosD has joined #webont
17:13:42 <pfps> meeting will be slightly out of order
17:13:43 <DanC> very well. I don't mind how you do it. I just wondered who was doing it.
17:14:42 <DanC> DanC has changed the topic to: WebOnt 7 Mar; chair: Guus; scribe: PeterPS.
17:15:35 <DanC> * DanC wonders who just joined
17:17:37 <pfps> focus groups and telecons
17:18:11 <pfps> lynn stein believes that focus groups with their own telecons will allow WG members
17:18:21 <pfps> to participate better in the areas that they are interested in
17:19:02 <DanC> note that W3C process requires 7 day notice of telcon times, meanwhile.
17:19:55 <pfps> Guus proposes to leave this issue as in the agenda
17:20:40 <pfps> Agenda item LANG OWL
17:22:07 <pfps> fvh: rationale for doing things as in the document: not to have to redo stuff
17:22:49 <pfps> fvh: logic vs ontology modelling language - so want lower entry cost
17:23:20 <pfps> fvh: proposal - keep axiomatic approach as basis
17:25:49 <pfps> fvh: proposal - add back in frame-like idiom (like in OIL)
17:26:11 <pfps> fvh: tests can be informal, then use frame syntax
17:26:26 <raphael> raphael has joined #webont
17:26:35 <pfps> fvh: semantics can use axiomatic part which will be like DAML+OIL ??
17:26:46 <raphael> sorry for coming late. Raphael Volz (FZI) joined.
17:27:43 <pfps> fvh: ontology stuff is still open
17:27:58 <pfps> guus: what does easy mean?
17:28:43 <pfps> fvh: easy is what is listed for frame idiom
17:28:58 <DanC> for reference: "OWL Language features: general line of attack & a first sketch" --
17:29:01 <pfps> strike previous line
17:29:16 <pfps> fvh: frame idiom is what is listed
17:29:59 <pfps> fvh: easy is easy to write/comprehend
17:31:02 <pfps> fvh: easy - unambiguous property (makes DanC happy)
17:31:26 <libby> * libby too
17:31:54 <DanC> re 7 day notice earlier; cf
17:32:00 <pfps> mike dean: inverse is problematic (hard or easy?)
17:32:59 <pfps> ian horrocks: inverse looks easy, but is hard to understand (doesn't match intuitions)
17:33:25 <pfps> fvh: how to fix - move to hard list, or provide some other construct
17:33:51 <pfps> deborah mcguinness: some things are easy in special cases, but hard in general
17:34:09 <pfps> deborah mcguinness: other things are hard in special cases, but easy in general  ?
17:34:24 <pfps> fvh: Q: idioms are only in examples?
17:34:28 <DanC> teaching people to think formally is a challenge. I hope we aren't aiming *too* high there. Anything more than giving standardized syntax to folks that already understand the underlying concepts is gravy, to me. If you can't set the time on your VCR, I don't expect you'll be able to use owl:inverse.
17:35:00 <pfps> deborah mcguinness: no, idiomatic syntax is also good, but examples are great
17:35:23 <pfps> pfps: what should be done next?
17:36:00 <pfps> fvh: work on frame system, collect idioms
17:36:17 <pfps> pfps: what about the requirements document
17:36:39 <pfps> fvh: need to select language features to satisfy requirements
17:37:00 <pfps> ian horrocks: need to work fast
17:37:30 <pfps> fvh: timescale for frame idioms / syntax - 1 to 2 weeks
17:37:47 <pfps> danc: for each requirement an idiom / construct ?
17:37:52 <pfps> fvh: very ambitious
17:38:09 <pfps> guus: get the other focus groups to help?
17:39:38 <pfps> ian horrocks: getting the right combination may not be trivial
17:39:55 <pfps> ian horrocks: getting the right syntax for the idiom may also be necessary (?)
17:40:22 <pfps> danc: copy and paste from examples may work
17:41:04 <pfps> fvh: only major difference from D+O is the frame part
17:41:29 <pfps> ian horrocks: maybe the D+O part can be reduced to basics
17:41:57 <pfps> ter horst: what are frames and slots as opposed to classes and properties?
17:42:18 <pfps> fvh: frames are restrictive, classes are not
17:43:04 <pfps> danc: frame definitions are encapsulated, RDF classes are not
17:43:32 <pfps> ian horrocks: true to an extent, but D+O can do this sort of encapslation via conjunction
17:44:11 <pfps> ian horrocks: in D+O you have quantification, which is hard for users to comprehend
17:44:28 <pfps> ian horrocks: frames hides this
17:45:04 <pfps> fvh: yes, in the standard patterns (required multi-value slot)
17:45:45 <pfps> carroll: restrictive reading of frames?
17:46:20 <pfps> fvh: yes I did *say* that
17:46:21 <DanC> we're up to 18registered to attend, 5 regrets, per . Real-time collaboration is fun. ;-)
17:46:44 <pfps> ian horrocks: worrisome, handled differently by different systems
17:46:58 <pfps> guus: related to modelling languages (e.g., UML)
17:48:04 <pfps> ian horrocks: we need a language that doesn't have surprises
17:49:35 <pfps> ACTION: produce an idiom for frame/UML stuff - frank - 2 weeks
17:50:21 <pfps> item 3) TEST
17:51:05 <pfps> jeremy: test cases help to iron out problems
17:51:25 <pfps> jeremy: emails give examples of how this is supposed to work
17:52:01 <pfps> jeremy: e.g. what does rdf:li mean and how does it work
17:52:21 <pfps> jeremy: don't have to provide precise *wording* for this
17:52:40 <pfps> jeremy: can be changed
17:53:32 <DanC> jeremy is discussing his message:
17:54:58 <pfps> DanC: the value of tests appears to be self-evident to WG
17:55:48 <pfps> DanC: e.g. showing entailment - e.g. unambiguous properties can tie together resource names
17:56:17 <pfps> DanC: can anyone handle this sort of test?
17:57:21 <pfps> ian horrocks: more or less - turn it into a test with a T/F answer
17:59:10 <pfps> jeremy: what kinds of tests will be considered
17:59:52 <DanC> I'd like somebody to answer "is our language closed under negation" briefly here and then in substance offline.
18:00:19 <pfps> no, the question may not even make sense
18:00:53 <las> * las just beeped as she lost the line and rejoined....
18:01:03 <pfps> DanC: where are lists of test cases?
18:01:10 <DanC> sound like actions on Ian and Deb.
18:01:21 <pfps> various: FaCT, DAML+OIL, CLASSIC
18:02:33 <DanC> I don't think we're committing to anything. But for my purpose, I'm convinced by Ian... I'm going to make one-file tests for the next week or so.
18:02:35 <pfps> jeremy: don't require single-input test cases
18:03:19 <DeborahMc>  is the pointer that i also sent to the mailing  list
18:03:28 <DeborahMc> this is just testing cardinalities
18:03:42 <DanC> "This is a small testing ontology. It is includes as an example of how to test a theorem prover implementing inferences required for complete processing of DAML+OIL."
18:05:00 <pfps> tim finin: lots of incomplete reasoners, so difficult tests are of interest
18:05:42 <pfps> ian horrocks: FaCT test set will be a good start there
18:06:00 <pfps> tim finin: probably need more kinds of tests
18:06:48 <pfps> deborah mcguinness: CLASSIC test suite is a regression suite
18:08:15 <pfps> stanton: lots of uses for test cases
18:08:52 <pfps> jeremy: tests cases may have to be fixed very late, maybe
18:11:01 <pfps> DanC: which systems are available - FaCT, RACER, DLP, finin?, DanC, JTP
18:11:19 <pfps> DanC: which systems will participate - 
18:13:03 <libby> * libby would like to see if plausible to see if query-based code can be used for the tests
18:13:41 <pfps> ACTION: DanC will produce several tests in DAML+OIL syntax
18:14:37 <pfps> ACTION: run these tests - Jos, Ian Horrocks, Deborah McGuinness, Tim Finin
18:15:21 <pfps> ACTION: Jeremy - look up different test suites and categorize
18:16:15 <pfps> ACTIOM timing - report back next week
18:16:49 <DanC> good question, libby... please do send some mail about how query systems could be tested.
18:16:56 <pfps> ACTION: stanton - report back on different kinds of tests (very short) - next week
18:18:42 <pfps> ACTION: Guus - turn CM examples into tests
18:19:15 <pfps> AOB
18:19:40 <pfps> Jos: layering?
18:20:50 <pfps> Guus: synthesis of layering discussion for next week
18:22:02 <ora> ora has left #webont
18:22:34 <raphael> raphael has left #webont
18:22:35 <DanC> hmm... Jos? I'm actually not sure how to get cwm to do these one-file tests. I'll have to think about it.
18:22:36 <JosD> JosD has quit

Received on Thursday, 7 March 2002 16:13:27 UTC