W3C Web Ontology Working Group IRC Chat Logs for 2002-07-02

00:00:05 <jhendler> deadline for raising editorial comments with respect to the documents is July 15.

00:00:53 <DeborahMc> editors prefer suggestions to have the form of an operational suggestion:

00:01:10 <DeborahMc> i.e., instead of saying "xxx" I suggest saying "yyy"

00:09:10 <GuusS> Raphael: should we define our "owl" version of Dublin Core features?

00:09:49 <GuusS> DanC: strongly agianst, tools like Adobe may have dificulties interpreting owl:title

00:10:13 <GuusS> Deb: confirms that her current wording about DC is OK

00:10:23 <DanC> an RCS ontology I happend to need enough to make it up: http://www.w3.org/2001/03swell/rcs.n3

00:10:41 <GuusS> Raphael: use CVS format for versioning information

00:24:54 <GuusS> Jeff: ability to state backward compatibility

00:27:46 <GuusS> discussion about the logical implications

00:32:02 <GuusS> no normative lagnauge features for ontology metadata

00:36:59 <GuusS> Jim: are the internationalization solved?

00:37:10 <GuusS> DanC: done by RDF Core

00:37:59 <GuusS> Other issues that are solved: character model,, ....(scribe forgot)

00:45:22 <GuusS> discussion about extensibility mechanism

00:45:33 <GuusS> Jim: summary

00:45:54 <GuusS> time needs to spent on versioning and import

00:46:25 * DanC realizes the only thing in order at this point is a motion to adjourn or a proposal to go past 5:30

00:48:47 <GuusS> Req: attachment of info to statements

00:49:24 <GuusS> It will have to be clear tomorrow wether there is an issue here

16:20:03 Topic now webont f2f log at http://www.w3.org/2002/07/01-webont-irc

16:20:03 Users on #webont: logger_1 DeborahMc Chris GuusS heflin jhendler JonB Zakim RRSAgent herman las

16:22:46 * las thinks Peter should say what he's concerned with

16:28:07 * DanC wonders who's scribing this session

16:30:18 <DeborahMc>http://www.ksl.stanford.edu/people/dlm/webont/OWLFeatureSynopsis.htm

16:30:19 <jhendler> Recap of discussion of document

16:30:24 <DeborahMc> this is my working version

16:30:37 <heflin> PPS expresses concern about whether docs will be ready for publication

16:30:41 <DanC> supplement to yesterday's record: * a few WebOnt notes from 1Jul connolly@jammer.dm93.org (Tue, Jul 02 2002) http://lists.w3.org/Archives/Public/www-archive/2002Jul/0003.html

16:30:42 <DeborahMc> it took into account real time updates from yesterday

16:31:30 <heflin> PPS: Okay if people can make objections after new drafts are available

16:32:36 <heflin> Hendler: Re reference description - mostly editorial changes

16:33:32 <heflin> Ian: doesn't think enough time to make editorial changes

16:36:13 <heflin> A few group members express desire for an extra week before release

16:38:19 <heflin> Sufficient consensus for extra week

16:38:47 <DanC> [I don't recall any 11July editorial deadline]

16:40:08 <heflin> PPS proposal: 7/11 versions, comments by 7/15, decision? by 7/18

16:40:35 <heflin> PPS: if no consensus on 7/18, then extend by 1 week to 7/25

16:40:43 <DanC> i.e. if we have consensus 18July, we can publish; else we take another week to decide.

16:40:56 <DanC> so RESOLVED.

16:41:24 <heflin> Jim: extra logical features

16:41:32 <heflin> Jim: versioning needs time

16:41:48 <heflin> Jim: internationalization go with RDF although not big fan of it

16:42:13 <heflin> Jim: imports has an issue that will be discussed

16:42:26 * DanC reviews records from yesterday to see if they say what JimH is saying http://www.w3.org/2002/07/01-webont-irc

16:42:57 <jhendler> summary on extra-logical:

16:43:04 <heflin> Jim: tagging of statements- must raise issue, remove from requirements, or decide it is already handled in some way

16:43:37 * DanC q+

16:43:39 * Zakim sees DanC on the speaker queue

16:45:07 <DanC> DanC explains mapping decision, recorded in http://lists.w3.org/Archives/Public/www-archive/2002Jul/0003.html

16:45:11 <heflin> DanC: other resolution from yesterday was a RESOLUTION that abstract syntax document provide a mapping to/from the language description

16:46:19 <heflin> PPS: has issue that we are adding lots of "syntactic fluff"

16:46:31 <heflin> Ian: thinks documents are mostly "PR"

16:48:47 <heflin> PPS: wants some current issues to be opened

16:48:56 <heflin> DanC: requests that 5.10 be opened

16:50:31 <heflin> PPS: Would like 5.3 and 5.10 referenced in language description

16:50:54 <heflin> DanC: maybe formal spec should be where these issues are referenced

16:52:11 <heflin> Jim: next conversation is about abstract syntax

16:52:35 <DanC> (I think peter just noticed the lack of mention of 5.3 and 5.10 in the language description; he didn't say they belonged; I think he said that was sorta the point; those issues aren't touched by that document)

16:53:32 <heflin> Jim: last telecon discussed abstract syntax, sent e-mail to mailing list

16:54:06 <heflin> one issue addressed by mapping resolution

16:54:37 <heflin> PPS: semantics have been out already

16:54:56 <heflin> comment that many semantics are out, which should be published?

16:55:05 <DanC> # SEM: semantics for the abstract syntax Peter F. Patel-Schneider (Tue, Jun 11 2002) http://lists.w3.org/Archives/Public/www-webont-wg/2002Jun/0082.html

16:55:06 <JonB> (JonB)

16:55:29 <DanC> # layering (5.3,5.10): a first-order same-syntax model theory Dan Connolly (Wed, Jun 19 2002) http://lists.w3.org/Archives/Public/www-webont-wg/2002Jun/0152.html

16:56:58 <heflin> PPS: prefers that semantics go out with other documents

16:57:36 <heflin> DanC: good to point out what positions docs take on open issues

16:58:04 <heflin> Jim: proposal to take abstract sytanx and turn into semantics doc

16:59:22 <heflin> DanC: has an alternative proposal he would prefer to publish

17:00:03 * DanC wonders what time expectations telcon participants have

17:00:16 <JonB>http://lists.w3.org/Archives/Public/www-webont-wg/2002Jun/0152.html

17:00:35 <JonB> DanC's proposal from June 19

17:02:38 <heflin> 705-756-3029 number to call PPS at

17:02:40 * las needs to leave....

17:02:46 <Zakim> -Stein

17:03:00 * las wishes all luck for this afternoon's discussion....

17:03:01 <DanC> oops; I cited the wrong model thoery

17:03:31 <Zakim> -??P3

17:04:05 <DanC> Re: layering (5.3,5.10): a first-order same-syntax model theory Dan Connolly (Fri, Jun 28 2002) http://lists.w3.org/Archives/Public/www-webont-wg/2002Jun/0209.html

17:04:12 <Zakim> -Welty

17:04:13 <Zakim> -??P4

17:04:14 <Zakim> - +1.705.756.aaaa

17:04:15 <Zakim> -JonB

17:04:17 <Zakim> SW_WebOnt()12:00PM has ended

17:04:23 <DanC> -> http://www.w3.org/2002/06/owlsem55.txt

17:09:26 <heflin> I'm still scribing - whoopee!

17:11:19 <heflin> Beginning call with PPS

17:12:30 <heflin> Ian: suggests that abstract syntax and mapping published before adding semantics

17:12:58 <heflin> DanC: objects to position abstract syntax takes on layering

17:15:01 <heflin> Deb: wants documents to separate out OWL lite from full OWL

17:15:20 <DanC> I think the relevant issues here are 5.1 Uniform treatment of literal data values, 5.3 Semantic Layering, 5.10 DAML+OIL semantics is too weak

17:16:05 <DanC> cf http://www.w3.org/2001/sw/WebOnt/webont-issues.html

17:17:04 <heflin> Straw poll: add mapping to abstract syntax but no model theory

17:17:13 <heflin> release on same schedule as other documents

17:19:05 * DanC thinks it's traditional to treat ammendments before the main question, but isn't sure

17:19:23 <heflin> IN FAVOR: 11

17:19:32 <heflin> OPPOSED: 1

17:20:09 <heflin> OPPOSED (DanC) can live with

17:21:35 <heflin> RESOLVED: see Straw poll. Abstain Danc, Jos, Jeff, who else?

17:21:46 <DeborahMc> the followon to my proposal was a separate section for owl lite with the additions pulled out for full owl

17:21:59 <DanC> "no model theory" doesn't preclude adding one in this meeting

17:22:29 <heflin> Proposal: have abstract syntax have specific section for OWL lite

17:22:34 * DanC realizes peter can't see the IRC log

17:23:10 <heflin> Revised Proposal: pull out section for OWL lite, distinguish differences for OWL

17:31:55 <heflin> DanC: Asks for preference of editor

17:32:47 <heflin> PPS: prefers defining OWL lite by substraction

17:33:21 * DanC prefers to let the WG give advice and then leave it up to the editors

17:34:44 <heflin> PROPOSAL: define OWL by addition from OWL lite

17:35:10 <heflin> In favor: 11

17:35:30 <heflin> Opposed: 3

17:36:28 <heflin> This was a straw-poll

17:38:22 <heflin> RESOLVED: with 7 abstains

17:41:14 <heflin> Jim: move to discussion of model theory

17:42:45 <heflin> "Formal Spec" is misleading, at least until it has a model theory

17:43:45 <DanC> "An OWL model theory layered on RDF" http://www.w3.org/2002/06/owlsem55.txt

17:43:54 <DanC> $Id: owlsem55.txt,v 1.2 2002/06/28 17:41:12 connolly Exp $

17:49:55 <heflin> DanC: has no comprehension axioms

17:50:18 <heflin> i.e, x in A intersect B does not imply x in B intersect A

17:58:21 <heflin> Issue is how we can have integration with this

17:59:40 <heflin> DanC says he could add an axiom to conclude that classes defined in this way are equal

18:07:03 * DanC would suggest a break

18:23:00 <heflin> Straw Poll: Take Peter's model theory and releasing subject to addressing Dan's concerns

18:23:51 <heflin> + to be done by August 1

18:25:33 <heflin> New Straw Poll: Put Peter's model theory through the standard editorial process

18:27:18 <heflin> (this means official reviewers must be solicited)

18:27:36 <heflin> In favor: 8

18:27:39 <heflin> Opposed: 5

18:27:46 <heflin> Abstain: 3

18:28:42 <GuusS> q+

18:28:45 * Zakim sees GuusS on the speaker queue

18:52:23 <jhendler> JimH scribes - session on Guide

18:53:08 <jhendler>http://www.w3.org/Archives/Public/www-webont-wg/2002Apr/0266.html

18:53:23 <jhendler> Guus' proposal for the outcomes of the GUIDE activity

18:53:44 <jhendler> 1) Presentation syntaxes: XML, UML

18:53:55 <jhendler> 2) Language primer/walkthrough

18:54:12 <jhendler> main change- should have more realistic ontology examples

18:54:39 * DanC waves from the coffe shopt

18:54:42 * DanC waves from the coffe shop

18:55:00 <jhendler> we should use examples from our use cases

18:55:08 * jhendler groups laughs at Dan's message

18:56:04 <jhendler> 3) How to do it document -- guidelines on how we might handle things in the objectives (example part/whole, schemas, defined classes, etc.)

18:56:14 <DeborahMc> gates 167 has been signed up as a breakout room

18:56:16 <jhendler> possible walkthrough and guideline merge

19:01:11 <jhendler> Guus: we need to produce these documents by Bristol (walkthru/primer; how-to-do-it )

19:01:31 <jhendler> Guus: what can we realistically achieve?

19:02:24 <jhendler> JimH: perhaps the articles http://www.xml.com/pub/a/2002/01/30/daml1.html -- by Roxane Ouellet, Uche Ogbuji - would be of use

19:02:50 <jhendler> JimH: possibility of doing the second as a FAQ - Deb: says she may have a starting place on that document

19:11:58 <jhendler> suggestion - maybe walkthru should be a document, how-to-do-it might be a web-based FAQ or other non-document

19:12:46 <DeborahMc>http://www.ksl.stanford.edu/people/dlm/papers/living-with-classic-abstract.html

19:12:52 <DeborahMc> this is a pointer to the paper i mentioned

19:16:43 <GuusS> datatypes -> walktru doc

19:16:52 <jhendler> discussion of what goes in each document

19:18:26 <jhendler> Mike Smith volunteers to lead the walkthru activity - Ora, Deb, Lynn have previously mentioned interest. Guus volunteers to help w/examples

19:18:34 <jhendler> Chairs accept Mike's offer

19:20:28 <jhendler> Evan: suggestion - let's collect examples and etc. and put on our web page

19:20:40 <jhendler> Guus agrees to maintain if Dan can work out access issues.

19:23:36 <jhendler> Deb - suggestion, let's convert the old walkthru

19:24:04 <DeborahMc> i also posted the how and when document to webont

19:24:20 <DeborahMc> on examples, i also suggest a set of wines examples i did for a recent article

19:26:11 <jhendler> Some people volunteering for pieces of "How to do document"

19:26:53 <jhendler> ACTION: Guus will generate a structure in which the examples should appear by July 11

19:26:54 * RRSAgent records action 1

19:27:09 <jhendler> this will also include one example

19:27:31 <jhendler> Larry is willing to write part/whole example

19:29:47 <jhendler> ACTION: Jim will work with Dan to set up structure for this. Guus will be the contact person for sending these things to.

19:29:49 * RRSAgent records action 2

19:31:48 <jhendler> Guus: presentation syntaxes

19:40:52 * jhendler notes that lunch is here in case the logicians are hungry...

19:43:43 <jhendler> discussion of how to do UML - Evan notes that DAML+OIL based UML tools exist, meaning implementation is important

19:45:44 <jhendler> ACTION: Evan will writeup a description of a recent OMG meeting that concerned UML and OWL, and the process he is running at OMG, and will post that to the WG

19:45:46 * RRSAgent records action 3

19:47:43 <jhendler> LUNCH!!!!

19:47:49 <DeborahMc> lunch until 2

20:32:33 * Zakim excuses himself; his presence no longer seems to be needed

21:06:17 <mdean> meeting resumed at 14:00

21:08:34 <mdean> discussing requirement 4.1 Ontologies as distinct objects

21:08:44 <mdean> Raphael expressed concern with using rdf:about=""

21:09:05 <mdean> changes if you move the ontology

21:11:27 <mdean> could use any URI in rdf:about="", but that loses binding with relative IDs or about's in that ontology

21:12:47 <mdean> xml:base would help here

21:13:31 <mdean> can't really have multiple ontologies (with common prefixes to Classes and properties) on the same page

21:14:27 <mdean> this could relate to imports

21:14:46 <mdean> Jim asked if anyone wanted to open an issue for this

21:17:34 <mdean> Raphael: propose that we use xml:base in documents and our .owl files

21:19:17 <mdean> ACTION (Raphael): send Guus paragraph suggesting preferred usage for owl:Ontology

21:19:38 <mdean> discussing unambiguous term referencing with URIs

21:19:42 <mdean> no known problems

21:19:53 <mdean> requirement: explicit ontology extension

21:20:36 <mdean> probably depends upon import issue

21:20:59 <mdean> for transitivity

21:22:25 <mdean> currently no strong notion of ontology extension

21:23:03 <mdean> can add restrictions to existing classes

21:23:17 <mdean> ontologies are not really first class objects

21:31:19 <mdean> plan to approve release of this Requirements document at end of this session

21:32:39 <mdean> could change a requirement to an objective

21:32:58 <mdean> WD updates include lists of substantive and editorial changes

21:33:06 <mdean> requirement: commitment to ontologies

21:34:51 <mdean> confusion over resources

21:35:38 <mdean> requirement: ontology metadata

21:35:46 <mdean> discussed

21:35:55 <mdean> currently can put anything in ontology headers

21:36:05 <mdean> we should provide some examples

21:36:59 <mdean> ACTION (Mike): use DC attributes in owl.owl

21:37:04 <mdean> (Mike Dean)

21:37:16 <mdean> requirement: versioning

21:37:23 <mdean> open issue, discussed yesterday

21:37:32 <mdean> requirement: class definition primitives

21:37:42 <mdean> addressed

21:37:49 <mdean> requirement: property definition primitives

21:37:53 <mdean> addressed

21:37:59 <mdean> requirement: data types

21:38:00 <mdean> open issue

21:38:12 <mdean> requirement: class and property equivalance

21:38:14 <mdean> addressed

21:38:21 <mdean> open issue to mix classes and properties

21:38:32 <mdean> requirement: individual equivalance

21:38:35 <mdean> owl:sameIndividualAs

21:38:49 <mdean> requirement: local unique names

21:39:41 <mdean> motivates owl:differentIndividualFrom

21:40:57 <mdean> abstract syntax provides macro function for many individuals

21:44:04 <mdean> role of UnambiguousProperty?

21:44:48 <mdean> this is currently being treated as a solved requirement -- otherwise needs a new issue

21:46:24 <mdean> ACTION (Deb): open issue

21:46:35 <mdean> could be resolved by guidelines

21:47:24 <mdean> when all issues are postponed or closed, we're done

21:48:30 <jhendler> ACTION: (Deb) write up an issue with respect to the unique names assumption requirement

21:48:31 * RRSAgent records action 4

21:49:21 <mdean> requirement: attaching information to statements

21:50:19 <mdean> currently only mechanism is RDF reification

22:04:09 <mdean> Mike showed example of using RDF statementIDs to show that Deb's hair was red on Tuesday

22:05:35 <mdean> not addressed by OWL model theory

22:16:34 <mdean> Jim suggests using owl:tag as an uninterpreted standard property

22:17:16 <mdean> see issue 4.4 extra-logical feature set

22:17:20 <mdean> issue not yet opened

22:17:52 <mdean> requirement: classes as instances

22:18:11 <mdean> Guus has a very nice use case for interoperability

22:18:22 <mdean> posted a couple days ago to www-rdf-interest

22:18:46 <mdean> Sergey Melnik's WordNet implementation

22:19:05 <mdean> hierarchy is hidden in instances and properties

22:19:45 <mdean> everything is just a Word, with properties like hyponymOf

22:19:57 <mdean> want to treat these instances as classes

22:20:39 <mdean> Michael Sintek: DLs will not be decidable with this feature

22:21:36 <mdean>http://lists.w3.org/Archives/Public/www-rdf-interest/2002Jun/0275.html

22:22:34 <mdean> Evan: do we need another issue (besides equivalentTo)?

22:31:39 <mdean> possible open new issue and subsume equivalentTo into it

22:31:46 <mdean> ACTION: (Raphael) raise issue

22:31:48 * RRSAgent records action 5

22:32:23 <mdean> ACTION: (Raphael) raise superissue to subsume equivalentTo

22:32:25 * RRSAgent records action 6

22:33:26 <mdean> RESOLVED: close issue 4.6 equivalentTo to be subsumed by the new issue Raphael will raise

22:34:42 <mdean> requirement: complex data types

22:34:45 <mdean> pending issues

22:34:49 <mdean> requirement: cardinality constraints

22:34:53 <mdean> satisfied

22:35:10 <mdean> various requirements satisfied

22:35:33 <mdean> charmod, etc: satisfied with RDF solutions

23:02:01 <jhendler> Resolved: Release the new draft of the requirements document as is

23:02:12 <jhendler> (way to go Jeff Heflin!)

23:02:33 <jhendler> --------------

23:02:59 <jhendler> report back of the Model Theory team

23:03:12 <jhendler> Pat: may be light at the end of the tunnel

23:03:45 <jhendler> Doing a model theory from scratch is dangerous -- should stick to conventionally understood techniques

23:04:05 <jhendler> Some of Dan's proposal maybe problematic with respect to that

23:04:39 <jhendler> Way in which Peter's model theory is phrased may help make relation to RDF (and RDF MT) more clear

23:04:43 <jhendler> major problems:

23:07:48 <jhendler> 1) what happens when you use a RDFS notion with a primitive from Owl - can you take "rdf:subproperty of owl:intersection"? should it be legal (probably yes), should the RDF inference apply to the owl vocabulary (three views - (a) NO, (b) sure, but if you do you are outside the semantic domain of owl, (c) syntactically restrict those - i.e. owl graph could be simply "illegal" by owl)

23:09:33 <jhendler> possibility - put these together - i.e. something could say "if you are owl legal graph" you will work w/a tool like fact, but if you use RDF you may not get right inferences - but you'll still be legal owl.

23:10:59 <jhendler> i.e. consider a well-formed RDF graph that is "not sanctioned" by the OWL model - we could say it is legal, but not "owl coherent" or something like that

23:13:32 <jhendler> for example - some tools could say "if you are syntactically in owl (plus RDF etc.)" then my tool will work. Someone else could say "my tool is 'smarter' but you must use a restricted graph if you want guaranteed results

23:14:24 <jhendler> the hard part -- unclear exactly where we must "darken" what to acieve this. There is consensus we should strive to make this set as small as possible.

23:15:53 <jhendler> RDFS vs OWL issue w/respect to what are classes and etc - so there may be things in RDFS that OWL cannot handle "properly" -- open question - what do you say about these?

23:16:33 <jhendler> should we sanction the mixing or restrict it?

23:31:47 <jhendler> (several examples - photo will be taken for record)

23:34:48 <jhendler> Discussion of various programs with respect to how they treat various things

23:43:53 <jhendler> discussion of various tools and their needs, and use cases

23:44:01 <jhendler> miked - did you discuss Lbase?

23:44:34 <jhendler> Pat - yes, describes it (a proposal w/Guha to map RDF, RDFS, DAML into a common specification lang - FOL plus a bit of XML and a couple other nice things)

23:45:16 <jhendler> then each tool can map to that logic - provides an axiomatic semantics and to provide a way of relating content in the different languages

23:46:02 <jhendler> RDF Core will have a non-normative mapping into Lbase as part of their semantic document

23:46:50 <jhendler> document about Lbase will be a W3C NOTE

23:47:43 <jhendler> makes it clearer what the differences between the languages is and how to map them

23:51:18 <DeborahMc> pat -this wont fix the layering problems - they just come from needing to represent the same things in two languages

23:52:02 <jhendler> MikeD - should we do Lbase for OWL? Pat - sure, probably in same way as RDF Core (as an appendix to the MT)

23:53:07 <jhendler> Pat in fact, mapping to Lbase and creating a model theory are very similar -- a good way to be clear and to get the issues resolved

23:53:15 <jhendler> (but doesn't solve the issues necessarily)

23:57:57 <jhendler> ACTION: Pat will attempt to take abstract syntax, and Peter's MT and the mapping into RDF and will write a model theory in the Connolly style (i.e. as an extension to RDF MT) and see if he can identify the exact issues.

23:57:58 * RRSAgent records action 7

23:59:23 <jhendler> RRSAgent, show action items

23:59:24 <RRSAgent> I see 7 open action items:

23:59:25 <RRSAgent> ACTION: Guus will generate a structure in which the examples should appear by July 11 [1]

23:59:27 <RRSAgent> recorded in http://www.w3.org/2002/07/02-webont-irc#T19-26-53

23:59:29 <RRSAgent> ACTION: Jim will work with Dan to set up structure for this. Guus will be the contact person for sending these things to. [2]

23:59:32 <RRSAgent> recorded in http://www.w3.org/2002/07/02-webont-irc#T19-29-47

23:59:34 <RRSAgent> ACTION: Evan will writeup a description of a recent OMG meeting that concerned UML and OWL, and the process he is running at OMG, and will post that to the WG [3]

23:59:37 <RRSAgent> recorded in http://www.w3.org/2002/07/02-webont-irc#T19-45-44

23:59:39 <RRSAgent> ACTION: (Deb) write up an issue with respect to the unique names assumption requirement [4]

23:59:41 <RRSAgent> recorded in http://www.w3.org/2002/07/02-webont-irc#T21-48-30

23:59:43 <RRSAgent> ACTION: (Raphael) raise issue [5]

23:59:44 <RRSAgent> recorded in http://www.w3.org/2002/07/02-webont-irc#T22-31-46

23:59:46 <RRSAgent> ACTION: (Raphael) raise superissue to subsume equivalentTo [6]

23:59:48 <RRSAgent> recorded in http://www.w3.org/2002/07/02-webont-irc#T22-32-23

23:59:50 <RRSAgent> ACTION: Pat will attempt to take abstract syntax, and Peter's MT and the mapping into RDF and will write a model theory in the Connolly style (i.e. as an extension to RDF MT) and see if he can identify the exact issues. [7]

23:59:53 <RRSAgent> recorded in http://www.w3.org/2002/07/02-webont-irc#T23-57-57