W3C Web Ontology Working Group IRC Chat Logs for 2003-01-09

07:52:02 Users on #webont: @logger

09:05:16 * RRSAgent is logging

09:05:31 <DanC> DanC has changed the topic to: WebOnt: Manchester ftf

09:05:56 <DanC> Guus convenes the meeting, welcomes implementors

09:06:16 <DanC> DanC has changed the topic to: WebOnt: Manchester ftf http://www.w3.org/2001/sw/WebOnt/

09:06:43 <DanC> Ian: recall from ftf info page (@@link): dinner at 7p at the Yang Sing etc.

09:07:22 * DanC RRSAgent, pointer?

09:07:23 <RRSAgent> See http://www.w3.org/2003/01/09-webont-irc#T09-06-43

09:07:51 <DanC> Attendance: Schreiber.

09:08:03 <DanC> -- Furguson, invited from Stanford, Protoge developer

09:08:14 <DanC> -- Connolly

09:08:21 <DanC> -- Barsia, UMD, invited

09:08:29 <DanC> -- Weilemaker (sp?), invited

09:08:33 <DanC> -- Hefflin

09:08:37 <DanC> -- Dean

09:09:01 <DanC> -- Stanton, DISA

09:09:08 <DanC> -- Buswell, Stilo

09:09:12 <DanC> -- Carroll, HP

09:09:18 <DanC> -- De Roo, Afga

09:09:33 <DanC> -- Jeff Pan, u. Manchester

09:09:38 <DanC> -- Horrocks, Network Inference

09:09:48 <DanC> -- Patel-Schneider

09:09:55 <DanC> -- van Harmelen

09:10:01 <DanC> -- Bajet, INRIA

09:10:16 <DanC> -- ter Horst, Philips

09:10:23 <DanC> -- Horan, Sun

09:10:32 <DanC> -- Beckhofer (sp?), Network Inference

09:10:38 <DanC> -- Crowther, NI

09:10:50 <DanC> -- Berkowitz (sp?), NI

09:10:54 <DanC> -- Hendler

09:11:32 <DanC> ====== Announcement from Hendler

09:11:59 <DanC> JH: owl.owl award goes to Frank van Harmelen for brokering the semantic truce.

09:12:38 <DanC> @@link to photo from Bijan

09:13:03 <DanC> ===== Working Draft review points, around the table

09:14:35 * DanC wrestles with chacl for the IRC log

09:16:13 <DanC> Hendler: editor: pls note new copyright statement at http://www.w3.org/Guide/pubrules#head updated for 2003 and INRIA->ERCIM transition

09:16:25 <DanC> s/editor/editors/

09:16:55 <DanC> Horrocks: Guide is in good shape. Semantics in good shape. Feature synopsis and reference need work; I'd like to see them combined.

09:17:48 <DanC> [note guus is taking notes on flipcharts; we might want pictures of those@@]

09:19:10 <DanC> Horrocks: in some cases, the description of the language features are better in the synopsis.

09:20:06 <DanC> Horrocks: it's clear what the purpose of guide and semantics are; what exactly are the purposes of synopsis and ref?

09:20:14 <DanC> PFPS: what Horrocks said, 2nded.

09:20:57 <DanC> ... I'm concerned about the amount of changes in the semantics doc in the last week or so. e.g. in response to comments from McBride

09:21:23 <DanC> s/McBride/Beckett/

09:21:32 * DanC wonders if jimh could paste a pointer

09:22:11 <DanC> JJC: RDFCore intends to approve Last Call Candidates tomorrow...

09:22:32 <DanC> ... editor's drafts are in w3.org space; you could link there for a week and do a global search/replace later

09:22:42 <DanC> PFPS: I also think Guide is in pretty good shape

09:23:15 <DanC> van Harmelen: re purpose of Synopsis: I suggest it's an entry point, as a preface to more serious docs like Guide, semantics.

09:23:32 <DanC> ... I support the proposal from a recent telcon to make it non-normative.

09:23:36 <jhendler> Dave Beckett's review of Semantics: http://lists.w3.org/Archives/Public/public-webont-comments/2003Jan/0000.html

09:24:01 <DanC> ... I think it should be more explicit about the semantics of the sublanguages.

09:24:10 <jhendler> Brian McBride's review of Ref: http://lists.w3.org/Archives/Public/public-webont-comments/2003Jan/0001.html

09:24:59 <DanC> VH: I think we should pay more attention to relationships between the documents. Another review is in order there... things like links between them, checking for redundancy and consistency.

09:25:51 <DanC> Baget: re synopsis: the dialects (Full/DL/Lite) are not sufficiently motivated.

09:26:43 <DanC> ... the motivation that seems implicit is to make implementation easier. For that purpose, some features are like "macros"; identifying those vs. a kernel would help.

09:26:56 <DanC> ... not sure where to put that. maybe in the semantics doc.

09:28:20 <DanC> ter Horst: re Reference: I'd like to see more examples. it's uneven. some important sections don't have examples.

09:28:29 <DanC> [something about overlap that scribe didn't catch]

09:29:26 <DanC> JJC: test document provides examples; does that help?

09:29:39 <DanC> ter Horst: it serves a different function. it helps, but...

09:30:23 <DanC> ter Horst: re Semantics: it has come far, but re the owl full/owl dl correspondence, there is an open point.

09:31:21 <DanC> ... and there are parts that are unreviewed. There's a new, very welcome, >3page [RDF-oriented description of DL?].

09:31:47 <DanC> ter Horst: I 2nd VH's comments about cross-document review.

09:31:58 * DanC can't recall speaker's name

09:32:14 * DanC thinks Horan

09:32:26 <DanC> Horan: as a consumer, the purpose/audience isn't clear.

09:32:32 <DanC> ... esp. Guide

09:32:38 <DanC> ... implementors? adopters?

09:32:54 <DanC> ... lack of consistency on terms: individuals, instances, members, elements

09:33:20 <DanC> ... with a background in Object Oriented Programming, Class/Instance means something completely different

09:33:43 <DanC> ... whether Full/DL/Lite are different languages, dialects, vocabularies is incosnsistently documented

09:33:57 <DanC> Horan: Guide is pretty good, except [missed it?]

09:34:04 <DanC> [something on ref?]

09:34:48 <jhendler> Horan: on Guide - good but could use a better consistency w/others and who intended audience is

09:34:54 <DanC> Horan: re Requirements: it seems some things haven't been met. [clarification about objectives vs. requirements]

09:35:30 <DanC> Beckhofer: mostly read Guide; it looks good

09:35:54 <DanC> Crowther: quite satisfied with Guide.

09:36:14 <DanC> ... supportive of combining features/ref

09:36:37 <DanC> Crowther: what does it mean to be OWL compliant? is a one-line shell script that accepts everything OK?

09:37:18 <DanC> [skipped somebody]

09:37:41 <DanC> Hendler: I read all 6 documents one day; came away feeling much more positive than before I read them...

09:38:05 <DanC> ... read some other W3C specs... if we released them today, they'd be as good as some stuff that has been used a lot

09:38:47 <DanC> Hendler: re ref/synopsis: given a lot of time, I think it might be good to combine them, but I'm not sure it's worth the time.

09:39:18 <DanC> Hendler: re ref purpose: [2 conflicting statements of purpose] let's go with the 1st of those: an explanation if you know RDFS. not an exhaustive reference

09:40:29 <DanC> ... we worked with students on porting to owl, and we found ourselves using the ref doc a lot; it answered questions about which vocabulary terms can be used where better than other docs.

09:40:51 <DanC> vanH: are you suggesting to make ref shorter?

09:41:19 <DanC> Hendler: yes, shorter with more links. systematic short reference with pointers. [something else?]

09:41:57 <DanC> Hendler: re Guide: it was mostly finished before OWL Full was decided on, so there isn't really a guide on OWL Full.

09:42:17 <DanC> ... it took us hours to discover that :subj :pred :aClass isn't in OWL DL.

09:43:01 <jjc> ACTION: jjc - add suvj pred aClass test case in I5.3

09:43:02 * RRSAgent records action 1

09:43:17 <DanC> Hendler: on features: I think it's close; as in email I sent, I suggest a section on OWL DL [and something about features of Full]

09:43:50 <DanC> Hendler: re Semantics: glad we have new reviewers. Jeff Pan [sp?] reviewed Semantics. Bijan is reviewing it.

09:44:41 <DanC> Mike: how much did the students use owl.owl? Hendler: not much; we didn't think to look there. more emphasis on the RDF schema for OWL seems in order.

09:45:37 <DanC> Crowther: are we saying that Reference is misnamed? "Reference" suggests a longer document.

09:45:45 <DanC> Hendler: yes, a different name seems in order.

09:46:24 <DanC> Furguson: read mostly Synopsis and Guide: I only have detailed comments.

09:46:36 <DanC> ... synopsis easy to read. very nice.

09:46:56 <jjc> DanC: comments

09:47:25 <jhendler> DanC: most of my stuff on earlier drafts:

09:47:30 <jhendler> DanC: semantics "done"

09:47:48 <jhendler> DanC: Guide "done" (but needs some fixes for Full, can live w)

09:48:29 <jhendler> DanC: Purpose of synop v. references - ref sets expectations it doesn't meet - what is shortest path to fix

09:49:11 <jhendler> DanC: should figure out cheapest way to fix feature v. Ref

09:49:57 <jhendler> DanC: Test - betw. Semantics/Guide - we may be in good shape - do NOT need a software spec -- we have met our charter and should not expand it.

09:50:09 <jhendler> DanC: will take more time to do so than is cost effective

09:50:45 <DanC> Parsia: docs aren't "gentle" to folks from an object oriented programming, or even from RDFS.

09:51:07 <DanC> ... folks taking RDFS schemas and moving to owl are going to have to do work that isn't sufficiently motivated.

09:51:38 <DanC> ... e.g. classes as instances

09:52:30 <DanC> Jan W [surname spelling?]: been using RDF/S for a long time... read OWL docs recently.

09:52:53 <DanC> ... "what do I need to do to make an OWL system?". I guess the test cases doc will help.

09:53:22 <DanC> [?]: issue isn't so much for implementor as systems integrator. [not sure I got that]

09:53:37 <DanC> Jan W: better roadmap, pls.

09:54:02 <DanC> Hefflin: Guide very good overall, still needs:

09:54:19 * jhendler Jan Wielemaker

09:54:27 <DanC> ... how to use owl ontologies to describe data. example of separate ontology from instance document

09:54:50 <DanC> Jan W == Jan Wielemaker

09:55:28 <DanC> [attendance + Chris Welty]

09:56:12 <DanC> Hefflin: I think it would make sense to move versioning stuff from [where?] to [where?]

09:56:23 <DanC> ... take some stuff from requirements on versioning

09:56:49 <DanC> Hefflin: I'd rather expand the reference than the guide; keep the guide short-ish.

09:57:24 <DanC> [something about abstract syntax?]

09:57:28 <jhendler> heflin: move versioning from up near top of doc (where it appears in wine.owl file) to further towards end "now that you know how to do ontology

09:57:58 <DanC> Heflin: we need to me more clear about the difference between sameClassAs and sameAs, applied to classes.

09:58:38 <jjc> ACTION: jjc - check sameInstanceAs/sameClassAs tests for FULL clarify different semantics

09:58:39 * RRSAgent records action 2

09:58:44 <DanC> vanH acknowledges this as a problem in synopsis

09:59:20 <DanC> DanC: our requirements doc is exemplary, both for audience for the WG, I think.

09:59:34 <DanC> Dean: I read [all the owl and RDF?] docs on the plane.

09:59:43 <DanC> ... (I sent mail about dataypes, btw...)

09:59:55 <DanC> Dean: I sent mail about the latest version [pointer somebody, pls?]

10:00:30 <DanC> Dean: I'd like to re-think the idea of putting the presentation syntaxes and MIME type stuff in the reference; they more than double the size of it.

10:00:53 <jhendler> Mike D's email re: datatypes - http://lists.w3.org/Archives/Public/www-webont-wg/2003Jan/0156.html

10:01:02 <DanC> Dean: McBride serves as "series editor" in RDFCore; let's consider doing that. JJC: that's a lot of work.

10:01:49 <DanC> Dale: nothing to add.

10:02:28 <DanC> Stanton: I think the Guide is a good rationale.

10:02:56 <DanC> Stanton: on conformance, I'm not clear... conformance to what? to W3C process? [not sure I got that point?]

10:03:41 <DanC> ... in my work, "degress" of interoperability don't help. One point of non-interoperability can costs zillions.

10:03:49 <DanC> ... I'd like to talk about a model implementation.

10:03:57 <DanC> ... or reference implementation

10:04:30 <DanC> Stanton: I'd like to see an exclusion of additional features as [outside something?]

10:05:03 <DanC> Buswell: most of it's been said... motivation for 3 species is lacking... should be in Guide.

10:05:17 <DanC> ... Guide also needs more about imports, and the interaction with dialects.

10:05:50 <DanC> Buswell: re ref: with programmer's hat on, "Reference" suggests a formal and complete specification. If that's not what it is, rename it.

10:06:25 <DanC> Buswell: tool vendors will want to say their tool is "owl compliant". the MathML spec has one, for example.

10:06:39 <DanC> (has a software compliance section)

10:06:48 <DanC> JJC: in sum, they're not ready.

10:07:32 <DanC> JJC: there's a lot of churn, and there are some outstanding technical issues... e.g. syntax about OWL DL/Lite.

10:08:12 <DanC> ... specs say OWL DL (and hence OWL Lite) is the range of the mapping rules, but WG discussion suggests otherwise.

10:09:11 <DanC> JJC: a section of the Guide could discuss "what you need to do to an RDF document to make it an OWL data document". It's not nothing.

10:09:25 <DanC> [scribe: OWL DL data document, I think he means]

10:09:39 <DanC> JJC: [bug in abstract syntax or something? missed it?]

10:10:37 <DanC> JJC: semantics have a "list of values" in the abstract syntax... it's a real pain to write in RDF. we need at least one example somewhere of how to write it, if we're serious about the language being that way.

10:11:23 <DanC> JJC: another technical issue: imports and levels: if you import an OWL Lite doc into into [something?], is the result [something?]

10:11:30 * DanC wonders why JJC didn't send mail raising new issues

10:11:45 <mdean> http://lists.w3.org/Archives/Public/www-webont-wg/2003Jan/0071.html contains F2F 5 version and lists outstanding issues and comments

10:12:11 <mdean> ... F2F 5 version of OWL Reference

10:12:34 <DanC> JJC: re process: I'm not prepared to vote on [decide] last call until we have 2 weeks to review stable documents...

10:13:10 <DanC> ... the mapping rules have changed, for good reason, about every day for the last weeks. We [I] can't begin to review until those mapping rules stabilize.

10:13:34 <DanC> Hendler: that two weeks doesn't make sense without committed reviewers, yes? JJC: yes.

10:14:06 <DanC> JJC: conformance is an unresolved issue. Fairly crucial, I think.

10:14:36 <DanC> JJC: I sent detailed review comments; I stand by those; I'm not sure how they've been disposed of...

10:15:23 <DanC> [specific comment on semantics; PFPS says "fixed"]

10:16:37 <DanC> JJC: re comments from McBride and Beckett: what I take away is that the overall level of polish isn't there. e.g. hyperlinking; for each term, it should have exactly one home, with other docs pointing there.

10:17:07 <DanC> ... Guide's level of editorial polish is quite good, as an exception.

10:18:02 <DanC> JJC: re annotations on classes. In the abstract syntax, they are allowed, but with some odd-looking restrictions.

10:19:01 <DanC> JJC: re features: if it makes it look easy to go from RDF to OWL [DL], it's wrong/misleading.

10:19:19 <DanC> JJC: I made a comment on the list of editors for reference... not sure what the status of that is...

10:20:28 <DanC> JJC: re status of features/reference: recently we talked about features being informative and reference being normative... discussion here suggests reference should be informative. re path of least action: it's to drop them.

10:20:34 <DanC> JJC: let's drop them both.

10:21:08 <DanC> JJC: if we keep reference, [good comment; pls key it in jeremy?]

10:21:27 <DanC> De Roo: I very much enjoyed reading the Guide.

10:21:51 <DanC> De Roo: re reference: minor point: owl:item [isn't there? shouldn't be there?]

10:22:35 <DanC> ... owl:incompatibleWith is in the reference; not sure what it means [?]

10:22:52 <DanC> Guus: AllDifferent is later in the agenda...

10:23:05 <DanC> De Roo: I think owl:DataType isn't needed; rdfs:Datatype is enough

10:23:39 <DanC> De Roo: owl:Class, owl:Set ... merits discussion.

10:24:26 <DanC> De Roo: in owl Full, we don't need [some condition about owl:Class and rdfs:Class ?]

10:25:00 <DanC> De Roo: re presentation syntax, let's keep in mind that they're non-normative. Document it better, if it's not clear.

10:25:15 <DanC> De Roo: re conformance: I 2nd Dan's point: too early to document conformance [of software?]

10:25:28 <DanC> attendance + Ziv Hellman (sp?) from Unicorn Solutions

10:25:38 <jhendler> de roo: we need if, not iff, for relation betw owl:class and rdfs:class

10:25:49 <DanC> Hellman: I like Guide very much...

10:26:21 <jhendler> attendance + Chris Welty

10:26:32 <DanC> Hellman: more UML diagrams [or did he mean mappings?] please. for Full, Lite, everything, please.

10:27:11 <DanC> Welty: 2nd point by JJC re syntax... difficult to tell where to go to figure out syntax...

10:27:32 <DanC> ... figuring out which document to go for what has been a challenge, in developing the Guide.

10:27:57 <DanC> Welty: I'm nervous about the wine ontology; needs review from [folks that know what our syntax is?]

10:28:26 <DanC> Dean: does latest version separate instances out? Welty: not sure.

10:29:08 <DanC> Welty: I think we haven't... I think it's not hard... JJC: it is hard!

10:29:51 <jhendler> Frank - Dan, based on history, what do you think?

10:30:11 <jhendler> Dan - this WG is more critical of their documents than the web community is!

10:30:49 <jhendler> dan - I will argue we are done.

10:31:14 <DanC> Pan: I reviewed Semantics in detail...

10:31:23 <DanC> ... I made detailed comments on section [5.2?].

10:31:33 <DanC> Hendler: is it pretty much ready? Pan: yes.

10:32:01 <DanC> ..........

10:32:32 <DanC> Guus: after the coffee break, I'll spend 10 minutes summarizing general issues before we break.

10:32:55 <DanC> Guus: the test document has a separate plenary discussion.

10:33:44 <DanC> Welty to scribe next session.

10:51:23 * DanC tries to chacl again...

10:51:26 * DanC RRSAgent, pointer?

10:51:28 <RRSAgent> See http://www.w3.org/2003/01/09-webont-irc#T10-33-44

10:51:58 * DanC cheers... the sun never sets on the W3C systems team!

10:52:53 <ChrisW> ready to scribe second session

10:53:10 * DanC linked IRC log from WG homepage

10:53:45 <ChrisW> Guus and Jim want to propose a breakout session of four doc editors (not Test) during the implementation session this afternoon

10:54:16 <ChrisW> think about a strategy for solving the term consistency problem

10:54:30 <ChrisW> Mike wants to attend the implementation session

10:55:59 <ChrisW> Jim suggests Mike send a replacement to the editors meeting and release "write lock" on reference

10:56:19 <ChrisW> Jim and Guus feel it is important for editors to meet in a subgroup to address the general issues

10:56:38 * DanC is scanning WG mail archive for "here's the doc/issues for the ftf" messages from the editors, linking them from WG homepage

10:57:06 <ChrisW> Guus goes over general points -

10:57:26 <ChrisW> combine FS & REF (shoudl be a WG discussion)

10:57:48 <ChrisW> FvH wants to discuss this issue now before further work on them

10:58:28 <ChrisW> Discussion of issue (whether to combine FS & REF docs)

10:58:43 <ChrisW> FvH felt FS is the "entry point" document for OWL

10:59:11 <ChrisW> Need to discuss what the perceived role of REF is

10:59:30 <ChrisW> Jim agrees - need to decide what REF is supposed to be

11:00:23 <ChrisW> MikeD - view of FS as the main document to describe the additional layers

11:00:28 <jhendler> but suggests we do it in small group, and then come back and discuss in more detail

11:01:39 <ChrisW> MikeD - the REF document is a traditional reference with index of "capabilities" and descriptions of them. Feels two documents are needed

11:02:13 <ChrisW> Herman - Is it a syntax reference?

11:03:39 <ChrisW> MikeD - Yes for the exchange syntax - but owl.owl is the real one

11:03:53 <ChrisW> Jeremy discusses what he thinks "syntax" really means

11:04:20 <ChrisW> - a set of RDF triples with constraints that come from the syntax and the language levels

11:05:36 <ChrisW> - mappings are only discussed in AS&S, and that is the only syntax

11:06:04 <ChrisW> Jim - the AS&S is the only normative document on the language

11:06:24 <ChrisW> - other documents should present this in more human-readable form

11:06:51 <ChrisW> jjc - ok, the other documents are informal. How many informal documents do we need?

11:07:21 <ChrisW> Jim - doesn't bother me, as long as they are consistent (which isn't the case now), as long as

11:07:28 <ChrisW> - their purpose is clearly stated

11:08:28 <ChrisW> jjc - given that a lot of people want to finish, isn't it less work to just drop reference? Doc doens't seem to serve a purpose

11:09:19 <ChrisW> JIm - FS (short synopsis, entry point document) is important to get OWL through the process

11:09:49 <ChrisW> DanC - if REF was focused on media type registration and owl.owl, it would seem better to me

11:14:49 <ChrisW> Guus - proposes that FS document (perhaps merged with REF) must stay

11:15:44 * DanC checks the record...

11:16:04 * jhendler suggests "OWL Vocabulary Document" instead of reference.

11:16:23 <ChrisW> - not a formal proposal, no one seemed opposed, jjc and MikeD said they would have abstained

11:17:05 <ChrisW> straw poll - 5 abstains, no opposed

11:17:54 <ChrisW> Guus observes that there seem to be three proposals regarding REF:

11:17:59 <ChrisW> - drop it

11:18:09 <ChrisW> - make it correspond to media types and owl.owl

11:18:16 <ChrisW> - make a new doc on media types

11:19:59 <ChrisW> 5 vote for "fixed editors view"

11:20:00 * DanC discourages the scribe from using the word "vote"

11:20:13 * DanC suggests "supports" or some such

11:20:30 <ChrisW> 8 vote for new documents

11:20:58 <ChrisW> 3 for dropping it

11:21:11 <ChrisW> 6 for dropping document but moving useful material to other docs

11:21:38 <DanC> Guus: summry: straw poll largely inconclusive.

11:21:41 <ChrisW> informal vote just used as input to the editors breakout group

11:22:46 <ChrisW> guus describes breakout sessions: FS, REF, GUIDE, SEM groups

11:22:50 <ChrisW> make list of things to change

11:22:55 <ChrisW> plan how to change them

11:22:59 <ChrisW> even do the work in the session

11:23:26 <ChrisW> Jim suggests also an intended audience description for each doc

11:23:54 <ChrisW> Guus - consistency issues should be discussed in the editors breakout

11:24:00 <jhendler> Jim "suggests" = jim states as chair

11:24:15 <ChrisW> FvH - what about differences with DAML+OIL

11:24:23 <ChrisW> -should go in REF

11:24:43 <ChrisW> Each breakout shoudl have a scribe

11:24:54 <DanC> pls mail break-out records to connolly,guus,jimh,www-archive@w3.org

11:31:40 <ChrisW> s.buswell scribing Guide breakout group

11:32:10 <ChrisW> present - welty, dale, horan, schreiber, hefflin

11:35:04 <ChrisW> changes to intro definition of ontology

11:37:29 <ChrisW> GS has created an example with OWL-full

11:38:44 <ChrisW> bh - questions use of para on 'availability of tools' ?

11:39:44 <ChrisW> cw - use is generic tools understanding OWL, do not need specific app.

11:40:19 <ChrisW> also reduces level of business logic in app

11:41:41 <ChrisW> berkowitz also in breakout group

11:47:39 <ChrisW> cw - guide is aimed at owl ontology editors, not tool implementers

11:47:48 <ChrisW> sb - disagrees

11:48:21 <ChrisW> jb - supports cw, but need OWL-DL supports ...

11:49:16 <ChrisW> gs - tried 2 examples, need things not in OWL-Lite

11:49:56 <ChrisW> gs - ok as guide is based around OWL-DL

11:51:42 <ChrisW> bh - waht is 'complete OWL vocabulary' ?

11:52:25 <ChrisW> gs/cw - means language constructs

11:52:37 <ChrisW> also say that OWL lite does not have all

11:53:31 <ChrisW> gs - link to feature synopsis

11:54:44 <ChrisW> reword 'type separation' tet

11:56:24 <ChrisW> bh - ?reasoning support 'predictable' - about existence of tools or results ?

11:56:43 <ChrisW> cw - both readings intended

12:00:36 <ChrisW> structure - wine.owl is OWD-DL ontology with marked exceptions

12:01:02 <ChrisW> gs - link here to semantics doc

12:04:33 <ChrisW> ontology headers - states that doc is intended to be an ontology

12:04:57 <ChrisW> cw - analogy is HEAD tag in HTML: ignore tagname

12:05:16 <ChrisW> jf - can used rdf:description for this

12:06:29 <ChrisW> jh - ontology 'means' reusable vocab

12:06:49 <ChrisW> jh - product catalog is not ontology

12:07:03 <ChrisW> gs - cannot make distinction

12:09:26 <ChrisW> jh - would like to make distinction

12:10:55 <ChrisW> bh - may want to subclass instance in someone else's ontlogy

12:14:56 <ChrisW> gs - ?maybe rename tag

12:17:12 <ChrisW> cw - imports has been updated

12:22:32 <ChrisW> jh - move versioning into separate section: most users will not need this in header section

12:23:02 <ChrisW> cw - but is part of ontology element

12:23:20 <ChrisW> cw - put forward ref to new section

12:24:27 <ChrisW> defining data - ref ontology from instance file ?

12:25:22 <ChrisW> jh - ?need separate file of instances re-using wine.owl ontology ?

12:26:35 <ChrisW> how does one convert an RDF doc into an OWL doc ?

12:26:54 <ChrisW> need someone who knows to write this section

12:28:13 <ChrisW> cw - we need an example

12:29:45 <ChrisW> bh - ?identify target readers?

12:30:09 <ChrisW> cw - OWL users defining ontologies (not tool developers)

12:53:55 * bwm wonders if he may lurk?

13:15:43 * jhendler says yes, but it is lunchtime

13:33:26 <JosD> ============ plenary reporting session from the different breakout sessions

13:33:41 <JosD> 1. Semantics

13:34:04 * DanC requests recording those volunteers as ACTIONs

13:35:48 <JosD> [@@ pointer to PFPS's summary]

13:37:26 <JosD> jeremy: my email with comments on annotations

13:42:04 <JosD> jeremy: points out to explain/be aware of the consequence of syntactic constaints holding for LITE/DL graphs

13:42:50 * DanC suggests ACTION Welty: explain that "all names need types" in the guide

13:43:02 * DanC suggests ACTION Welty: explain that "in OWL Lite, all names need types" in the guide

13:43:41 <JosD> see very last paragraph in 4.1 of seamantics

13:43:50 <DanC> ACTION Welty: explain that "in OWL Lite, all names need types" in the guide

13:43:51 * RRSAgent records action 3

13:44:04 <DanC> .... and and in OWL DL.

13:47:15 <JosD> Peter closes his discussion: do the right thing

13:47:21 <JosD> 2. Guide

13:48:09 * DanC wonders if PFPS sent a "here's the semantics version for the ftf" message to the WG

13:48:17 <JosD> [@@ pointers to ChrisW's summary]

13:49:46 <DanC> found it: http://www.w3.org/2001/sw/WebOnt/guide-src/wine.owl

13:49:55 <DanC> under 1.2. Structure of the Document

13:52:50 * DanC looks at wine.owl thru N3 glasses

13:53:36 <JosD> Guus: rdf->owl full is straight ahead

13:54:46 <heflin> Chris, what were you asking me?

13:56:05 <JosD> ACTION: Jeremy writes a section about the constraints to go from RDF to OWL-DL/LITE

13:56:07 <ChrisW> You had said in the group you might be able to take that action (That jeremy took)

13:56:07 * RRSAgent records action 4

13:57:11 <JosD> 3. FS

13:57:38 <JosD> rename to Overview??!!!

13:58:10 <JosD> [@@ pointer to FrankVH summary]

13:59:43 * DanC finds something surprising about CabernetFranc in wine.owl, but can't find it in the text of the guide

14:02:04 <JosD> FVH's FS summary http://lists.w3.org/Archives/Public/www-archive/2003Jan/0021.html

14:02:53 <JosD> jeremy: it sound good

14:03:23 <JosD> frank: good to hear

14:04:05 <ChrisW> \me wonders what DanC found

14:04:09 * ChrisW wonders

14:04:52 <JosD> mikeD: is it the entry point?

14:05:05 <JosD> frank:yes

14:05:13 <JosD> 4. Reference

14:05:55 <JosD> JimH presents summary http://lists.w3.org/Archives/Public/www-archive/2003Jan/0022.html

14:12:14 * DanC creates a #webont-guide channel to discuss wine.owl, hopefully with ChrisW

14:15:02 <JosD> Jeremy: section 2 rather as an appendix

14:17:00 <JosD> JimH: things that are currently hooked into other documents

14:18:15 <JosD> Guus: main part is language structure and if you shorten that... more convinced about FS

14:19:35 <JosD> IanH: a quick reference would be a good thing to hava

14:22:30 <JosD> Jeremy: repeats his concern about timing specifically w.r.t. Reference

14:24:35 <JosD> JimH: pipeline (document) structure versus portal/dispatch structure and feels we are more going to the latter

14:25:02 <JosD> [Guus took a picture fron FimH drawing]

14:25:36 <JosD> JohnS: the "kapstan" document

14:25:42 <DanC> capstone

14:26:15 * JosD oops

14:29:11 <JosD> Guus: straw poll "option of quick reference: do we need such a document"

14:29:51 <JosD> CrisW: could be achieved as an index document fron reference

14:32:34 <JosD> DanC: better if all documents are uniform

14:32:45 <DanC> better if *each* document is uniform

14:36:48 <JosD> Guus: feeling confident to decide tomorrow

14:37:19 <JosD> ============ coffee break

14:50:46 <jjc> jjc is now known as jjcscribe

14:51:05 <jjcscribe> Afternoon session - implmentors

14:51:13 <jjcscribe> Ray Ferguson - Protege

14:51:28 <jjcscribe> Presentation to be available at www-archive

14:53:38 <DanC> RayF: when rule systems can't handle feature X, we advise our users to just don't use feature X; we get by that way.

14:53:41 <jjcscribe> Some systems have 100,000 classes with 50,000 instances

14:54:35 <jjcscribe> PSM = Problem Solving Methods

14:54:46 <jjcscribe> i.e. generic ways to solve problems, e.g. classification

14:56:34 <jjcscribe> Protege - impl OWL Lite is hard,

14:56:49 <jjcscribe> so we will probably implement a subset of OWL Full that is not OWL Lite or OWL DL

14:58:13 <jjcscribe> Protege and RDF history - we worked on an RDF backend

14:58:25 <jjcscribe> We released it in alpha

14:58:30 <jjcscribe> it didn't work

14:59:06 <jjcscribe> so we got the reputation in the RDF world that Protege did not work for RDF

14:59:49 <jjcscribe> eventually we got a system that worked with RDF, but people no longer believed in us.

15:02:05 <DanC> hmm... I wonder why "you don't want to use an RDF parser for OWL". We use RDF triple-stores for OWL all the time.

15:04:33 <jjcscribe> differentFrom problem N^2 problem.

15:06:03 <jjcscribe> OWL DL and OWL Lite have restrictions ...

15:06:25 <jjcscribe> range constraints ...

15:06:54 <jjcscribe> modellers want range constraints with two classes with disjunctive semantics

15:07:55 <DanC> ACTION: IanH explain how to say "range of P is A union B" in owl lite

15:07:56 * RRSAgent records action 5

15:08:10 <jjcscribe> Metaclasses are used by more than 50% of users

15:09:58 <jjcscribe> protege supports any integer for cardinalities ...

15:11:00 <jjcscribe> protege uses closed world assumption, but only partially enforced

15:11:30 <jjcscribe> End first presentation

15:11:34 <jjcscribe> Second presentation

15:14:53 <jjcscribe> Bijan from Mindswap

15:15:28 <seanb> quit

15:16:27 <jjcscribe> use multiple ontologies and pieces of ontologies

15:18:26 * DanC RRSAgent, pointer?

15:18:27 <RRSAgent> See http://www.w3.org/2003/01/09-webont-irc#T15-17-32

15:19:23 * DanC wonders who bh is

15:19:40 <bh> bh is Bernard

15:19:59 * DanC thx

15:20:12 <jjcscribe> demo

15:20:25 <jjcscribe> parser tries to do the best it can with broken material

15:20:57 <jjcscribe> reasoner is not sophisticated

15:21:15 <jjcscribe> follows superclass links

15:21:48 <jjcscribe> validating markup, not sophisticated reasoning

15:22:25 <jjcscribe> some tools provide crawling, and text searches of ontology

15:23:45 <jjcscribe> xsd support

15:25:32 <jjcscribe> end demo - back to slides (SMore)

15:25:38 <jjcscribe> (demo was RIC)

15:26:28 <jjcscribe> SMore creates RDF instance data with explicit knowledge derived by reasoning

15:29:02 * JosD very impressed by the reasoner based composer

15:30:02 <jjcscribe> End presentation

15:30:06 <jjcscribe> Next presentation

15:30:27 <jjcscribe> Network Inference - Sean B.

15:36:47 <jjcscribe> Grumb les + Issues

15:36:57 <jjcscribe> freames and roundtripping

15:37:24 <jjcscribe> compliance is very important

15:38:07 <jjcscribe> datatypes ignored

15:39:41 <jjcscribe> reasoning is classification and satisfiability

15:39:51 <jjcscribe> Peter Crowther continues ...

15:44:27 <jjcscribe> On real systems Query writers are not the ontology designers

15:44:42 <jjcscribe> contrast this with research situatuon where it is the same person

15:48:06 <DanC> ---- break

16:02:10 <jjcscribe> next presentation

16:02:22 <jjcscribe> Jan from SWI Prolog

16:02:40 <jjcscribe> Demo MIA tool

16:03:18 <jjcscribe> annotate photos and pictures

16:03:47 <jjcscribe> use various DB with info about terms from art world, and wordnet

16:04:09 <jjcscribe> tool does some inheritance based reasoning

16:05:17 <jjcscribe> both class and proeprty hierarchies shown from AAT and WOrdnet

16:05:25 <jjcscribe> translated into RDF

16:05:47 <jjcscribe> ontology browser

16:06:05 <jjcscribe> can view class as class or class as instance

16:06:16 <jjcscribe> browser not editor

16:06:28 <jjcscribe> end demo - start slides

16:08:32 <jjcscribe> currently based on RDFS if we move to OWL it will be OWL full

16:08:46 <jjcscribe> because we don't need the formal reasoning and we need metaclasses

16:09:40 <jjcscribe> no standards for translating ontology into RDFS

16:09:47 <jjcscribe> we did a home-brew

16:11:26 <jjcscribe> wordnet:hyponymOf rdfs:subpropertyOf rdfs:subClassOf .

16:11:38 <jjcscribe> allows reinterpreting of whole database

16:12:07 * DanC wonders whether that's just a neato trick or if it's a really good engineering technique

16:18:54 * jhendler having irc problems - does this show up?

16:20:02 <JosD> yes

16:20:50 <jjcscribe> end presentation

16:24:02 <jjcscribe> start presentation - Raphael

16:35:24 <jjcscribe> end presentation - see slides on archive

16:36:18 <jjcscribe> start presentation - Jeff Heflin

16:36:44 <jjcscribe> demo of D+O query tool

16:36:50 <jjcscribe> takes in ontologies and instance data

16:37:08 <jjcscribe> shows hierarchies of ontologies

16:37:27 <jjcscribe> filtues ontologies

16:37:34 <jjcscribe> filtures ontologies

16:38:25 <jjcscribe> MS Access underlying this,

16:38:36 <jjcscribe> does not have enough power,

16:38:47 <jjcscribe> use FACT at load tiome to compute subsumption hierarchy

16:40:22 <jjcscribe> database is pretty fast approach

16:41:56 <jjcscribe> end presentation

16:42:12 <jjcscribe> start presentation - Jos De Roo

16:48:59 <jjcscribe> live session showing passing of RDF entailement tests and many of the OWL tests.

16:49:37 <jjcscribe> proof is output ...

16:50:33 <jjcscribe> jos complains is test haven't finished during his short presentation!

16:51:44 <jjcscribe> concluding slides ...

16:52:00 <jjcscribe> shows medical application to do with leg length

16:53:53 <jjcscribe> proof finished end of presentation

16:54:21 <DanC> very nifty stuff, that Euler.

16:55:02 <jjcscribe> next presentation - Mike Dean

16:55:14 <jjcscribe> OWL Validator - update on DAML validator

16:55:34 <jjcscribe> anlagous to lint for C

16:58:00 <jjcscribe> RDF parser not quite up to date (no datatype support, Jena 1.6.0)

17:00:22 <jjcscribe> The validator "validates" against owl.owl

17:02:28 <jjcscribe> DAML programme is migrating to OWL ....

17:08:54 <jhendler> Jeremy - Jena update

17:09:01 <jjcscribe> jjc: jena 1 stream is complete

17:09:04 <jhendler> Jeremy - Jena 1 reahing end of its life

17:09:15 <jhendler> Jeremy - Jena going to

17:09:24 <jjcscribe> next - RDF support with new specs

17:09:34 <jhendler> ... support for RDF w/new specs

17:09:47 <jhendler> Jena 2 CVS available

17:09:59 <jhendler> i.e. pre-alpha

17:10:15 <jhendler> RDF 2003 compliant parser

17:10:21 <jhendler> N3 output

17:10:31 <jhendler> Daml support not yet migrated

17:10:54 <jhendler> expect to migrate DAML stuff, put OWL stuff "on top of it"

17:11:31 <jjcscribe> reasoning - will use other peoples reasoners

17:11:33 <jhendler> not clear what reasoning J2 will support - ideal is for Jena to use other people's reasoners with an API/connection to other Jena components

17:12:03 <jhendler> J2 will do better support for ontology stuff going to DB

17:12:14 <jhendler> pressure on us to move to OWL

17:12:29 <jhendler> pressure on us to become RDF Core reference architecture

17:12:57 <jhendler> datatype support in their now - Jos - but not complere - Jeremy - coming along

17:13:17 <jhendler> working out our "where to go next" in next couple of weeks

17:13:42 <jhendler> Plan release of parser very soon, and release of rest of stuff (pre-alpha) not too long from now

17:14:01 <jhendler> hope to have RDF 2003 w/in time of RDF 2003 move from LC to CR/PR

17:21:40 <jjcscribe> connolly shows some demos from SWAP

17:22:31 <jjcscribe> connolly tries to show some demos from SWAP - "I am losing hopelessly"

17:23:45 <jjcscribe> eventual usccess - showing png giving flight plan

17:24:35 <ChrisW> what's the email address to send scribe notes?

17:25:24 <jjcscribe> www-archive@w3.org

17:27:35 <jjcscribe> connolly also shows an OTTER proof of the statecode test

17:27:45 <jjcscribe> (see www-archive for Jan 2003)

17:28:11 <jjcscribe> end presentation

17:29:33 <jjcscribe> Chris Welty reports back from break out group

17:38:17 <jjcscribe> discussion of normative status of TEST - current status OK

17:38:23 <jjcscribe> discussion of owl.owl

17:38:41 <jjcscribe> current status not OK - technical problems with comments, - true in all interpreations?

17:39:16 <jjcscribe> close for day.

20:58:55 * RRSAgent excuses himself; his presence no longer seems to be needed