W3C Web Ontology Working Group IRC Chat Logs for 2002-04-09

07:39:42 <JosD> Peters's started his talk http://www-db-out.bell-labs.com/user/pfps/talks/webont-f2f2-approaches/Overview.html

07:46:48 * RRSAgent is logging

07:47:01 <jhendler> jhendler has changed the topic to: Web Ontology WG - f2f: - log -

07:47:11 <JosD> syntax discussion...

07:48:20 <jhendler> RRSAgent listen

07:48:27 <jhendler> RRSAgent, listen

07:48:28 * RRSAgent is already logging

07:48:39 <jhendler> RRSAgent, bookmark

07:48:40 <RRSAgent> See http://www.w3.org/2002/04/09-webont-irc#T07-48-39

07:49:04 <jhendler> jhendler has changed the topic to: Web Ontology WG - f2f: - log - http://www.w3.org/2002/04/09-webont-irc

07:50:04 <JosD> pfps: if we keep RDF/XML syntax be careful not to fall inthe RDF pitfall

07:52:40 <JosD> jimH: don't confuse syntactical and semantical pieces of the language

07:54:08 <JosD> jeremy: decided yesterday that the base syntax is the graph syntax, *base* syntax (not presentation)

07:54:29 <JosD> frankvh: RDFCore will tell us how to do

07:54:51 <JosD> Path: a way of conveying a graph over the wire, that's all

07:55:25 <JosD> massimo: RDF/XML transport syntax

07:55:55 <JosD> jeremy: presentation syntax is out of order in this mornings's discussion

07:57:19 <jono> jono is now known as jd

07:57:54 <JosD> (peter proposed) decisions: syntax - RDF graph; expressive power - D + O

07:58:08 <JosD> peter continues his talk ...

08:00:14 <JosD> slide http://www-db-out.bell-labs.com/user/pfps/talks/webont-f2f2-approaches/slide7-0.html

08:01:03 <JosD> frankvh: proposes the term (Pat corrects) --> serialization syntax

08:02:18 <JosD> frankvh also proposes ``presentation'' syntax, normative? Jeremy doesn't care-- StevenB does care

08:03:03 <JosD> ... the issue is roundtrip

08:03:38 <JosD> PatH: the term roundtrip has 2 meanings (interrupted by JimH)

08:04:32 <JosD> JimH: we will return to this point today (after the coffee break)

08:04:49 <JosD> Peter continues again... RDF compatibility\

08:05:18 <JosD> slide http://www-db-out.bell-labs.com/user/pfps/talks/webont-f2f2-approaches/slide8-0.html

08:06:16 <JosD> (further proposed decisions) RDF MT; Inferences (what follows from what)

08:07:24 <JosD> Jeremy: entailment to clarify the meaning

08:08:51 <JosD> jonathanB: concerned about violation of intuitions...

08:16:42 <JosD> Peter & Pat believe there is a way to go...

08:17:07 <JosD> there is a diagonal line asserted / darh

08:17:36 <JosD> i.e. John loves Mary / D + O syntax

08:18:21 <JosD> i.e. constrain interptretaions / mentions like D + O syntax

08:22:16 <JosD> PatH: RDF is a vocabulary defining toolkit

08:24:21 <JosD> frankvh: 2 things: *language* conformance level 0 versus vocabulary defining toolkit

08:26:07 <JosD> Ian: the whole trouble is that it is doing these 2 things

08:27:23 <JosD> JimH is again interrupting the discussion

08:28:15 <JosD> JimH takes ACTION (for CG) : relationship with RDFS has to become very clear

08:29:06 <JosD> Jimh: RDF(S) has to mean the same thing...

08:30:06 <JosD> Guus: identify the part of RDFS that may act as conformance level 0

08:30:57 <JosD> PatH: yes we can (it's not vocabulary, but certain combination patterns (if I understood that well???))

08:32:12 <JosD> PatH: maybe, maybe (Peter: be lord and master of the universe.......)

08:32:59 <JosD> Peter continues... how to make the distinction (betweeen asserted and unasserted)

08:34:29 <JosD> one syntactical way: XML -- RDF document (asserted ) / XML Schema -- OWL document (unasserted)

08:34:55 <JosD> is it nice???

08:37:38 <JosD> all the definition stuff goes in the unasserted side

08:38:08 <JosD> they are dard for RDF but they have a meaning

08:39:20 <JosD> Jeremy wonders about his proposal (the sollipsitic one)

08:40:46 <JosD> Jeremy: you have to *assert* the existence of the class

08:43:40 <JosD> Peter: ontologies *refer* to other ontologies, or you merge them

08:45:15 <JosD> Guus: conclusions: issue is RDFS clarification

08:46:38 <JosD> 2. the proposed dark triple proposal is *not* affecting RDF core (question)

08:48:38 <JosD> PatH: 1. WebOnt definitely needs dark triples

08:49:42 <JosD> PatH 2. several proposals are possible, some are affecting RDFCore

08:51:13 <JosD> e.g. Jonathan's proposal recursive rdf:RDF elements to represent contexts

08:54:12 <JosD> Guus: there is an urgent timing issue involved!!!

08:55:54 <JosD> Jeremy is showing what current parsers do...

09:31:31 <JosD> Guus: propose synthesis discussion

09:32:20 <JosD> RDFCore provides us a serialization syntax

09:33:04 <JosD> let's use RDF/XML

09:34:38 <JosD> (the term used by RDFCore)

09:35:12 <JosD> ... and use that to serialize OWL level1 and OWL level 2

09:37:25 <JosD> RDFCore: RDF serialization syntax -> RDF graph

09:38:20 <JosD> WebOnt requirements -- 2 issues 1) LIST 2) DARK TRIPLES

09:42:12 <JosD> JimH: proposed action: PeterPS + FrankVH + JeremyC write down a paragraph for RDFCore w.r.t. these issues

09:42:42 <JosD> (will happen today)

09:43:33 <JosD> JimH: if we have a preference, we should say that

09:44:26 <JosD> other side if Guus's picture: OWL presentation syntax

09:50:18 <JosD> issues: 1) readablility 2) level of roundtripping (same presentation syntax/ other PS) 3) preferred one (normative?)

09:52:30 <JosD> (discuss again the trick to keep a pointer to the original RDF/XML document)

09:53:32 <JosD> Jeremy: roudtripping argument is spurious, issue is keeping the original presentation syntax

09:54:51 <JosD> jb: in general rt in not solvable, so keep original...

09:56:40 <JosD> Enrico: good requirement would be to give back a piece which is semantically equivalent

10:02:48 <JosD> StephenB: you can't have a requirement of rt (e.g. for something that came in as serialization syntax)

10:04:19 <JosD> frankvh: giving example of Protege tool (presentation syntax), but it doesn't work across tools...

10:05:47 <JosD> JimH: all of us are interoperating if we don't have a strong presentation syntax

10:06:20 <JosD> ... dozens of different ways to do this

10:08:18 <JosD> Ora: what are the characteristics we are interested in to be preserved?

10:09:14 <JosD> FrankVH is really concerned about usability

10:10:40 <JosD> Guus: group definitions + editable in something like emacs

10:11:39 <JosD> Enrico: requrement is slightly more readable syntax

10:12:21 <JosD> MikeD: worry is wether tools should export both

10:12:53 <JosD> StephenB: not mandate, keep as local thing

10:13:41 <JosD> Jeremy remarks that there is no common agreed format for word processors

10:15:15 <JosD> JimH: what GUIDE can do is work out this stuff in N3, UML etc

10:16:20 <JosD> ... don't give it the level the AC (advisory commitee) has to approved

10:19:51 <JosD> Ora: write down an ontology for the description of the things to be preserved

10:21:46 <JosD> Guus: 2 forms: readable text form (e.g. N3) and graphical form (e.g. UML)

10:23:15 <JosD> Enrico: user want 1 syntax which is easy to understand

10:24:40 <JosD> MikeD: finds the graph notation very useful (even to explain derivations)

10:28:00 <JosD> JimH: being able to interoperate is crucial but it's fine to have several ps's in the nn appendices

10:29:41 <JosD> (ps == presentation syntax; nn == non normative)

10:32:28 <JosD> jb: nn because of all syntactical guarantees; not part of the formal language definition

10:34:03 <JosD> StephenB: explains what it means to be NN using MathML example

10:34:54 <JosD> Massimo: explains further along the lines of the W3C process

10:35:38 <JosD> ... it is completely optional, it *might* become a de facto practice

10:37:30 <JosD> --------------------------------- END Of SESSION / LUNCH BREAK

11:58:45 <JosD> discussing entailment testcase :Peter a :Student, :Employee . |= :Peter a [ owl:intersectionOf ( :Student :Employee ) ] .

11:59:39 <jhendler> question raised: does a user (of OWL) need to know this exists?

12:00:46 <JosD> Jeremy would have to add [ owl:intersectionOf ( :Student :Employee ) ] . at LHS

12:00:56 <jhendler> that is - for dark triples would the rdf/xml need to include dark - answer, yes.

12:01:58 <jhendler> if you do it via "solipsistic" logic, the user doesn't need to know

12:02:07 <jhendler> however, there may be other problems.

12:23:10 <JosD> strawpoll: 12 people support dark triple proposal; 3 support solipsistic proposal; 6 abstain

12:25:35 <JosD> finding agreement on the proposed paragraphs...

12:29:50 <JosD> 1 paragraph is request; other paragraphs are suggestions

12:37:08 <JosD> volunteers from this group to join conversation with RDFCore: Jeremy, Pat, Peter, Jos, Massimo

12:37:29 <JosD> + Jonathan (Borden)

12:38:57 <JosD> unanimous agreement to send message to RDFCore

12:40:23 <JosD> Action to the chairs Jim and Guus to report this issue to the SWCG

12:42:26 <JosD> ======= Ora and "Thoughts about the Guide"

12:43:53 <JosD> (this is a summary of their break-out session)

12:54:38 <JosD> ======= discussing (again) D+O Appendix: Index of all language elements

12:55:39 <jhendler>http://www.cs.umd.edu/~hendler/2002/DAML-OWL.html

12:56:26 <jhendler> http://www.cs.umd.edu/~hendler/2002/DAML-OWL.txt

12:56:45 <JosD> ? * cardinality

12:57:58 <JosD> actually should be ? * cardinality

12:59:05 <JosD> oops 1/2 * cardinality

13:00:00 <JosD> decided taking out all the ...Q things

13:01:23 <JosD> R == rdf(s) 1 == level 1 2 == level 2

13:03:48 <JosD> DatatypeProperty, DatatypeRestriction, Datatype value are in 1

13:04:58 <JosD> Disjoint is in 1

13:07:18 <JosD> disjointWith is X (i.e. gone)

13:08:10 <JosD> arrow is, it goes to the issue list, cardinality, Class, equivalentTo

13:09:44 <JosD> hasClass, hasValue level 1

13:10:49 <JosD> imports is an X

13:12:03 <JosD> oops, imports is an -> (issue list)

13:13:22 <JosD> intersectionOf, inverseOf also issues

13:16:13 <JosD> ObjectClass, ObjctProperty, ObjectRestriction are in 1 (however there is the issue Datatype disjoint from Class)

13:16:46 <JosD> oneOf, onProperty, Ontology are in 1

13:17:46 <JosD> Property, range, subClassOf, subPropertyOf fall in R(DF)

13:19:00 <JosD> UnambiguousProperty (new proposed name UniqueIdentifyingProperty) level 1

13:20:57 <JosD> UniqueProperty (new proposed name SinglevaluedProperty) level 1

13:22:05 <JosD> version level 1

13:25:41 <JosD> ---------------------------------------time for a break....

13:29:33 <JosD> remarkable thing: UnionOf is the only level 2

13:29:57 <JosD> let's talk about that during the coffee break (10 min)

13:44:06 <JosD> rdf issue list http://www.w3.org/2000/03/rdf-tracking/

13:45:42 <JosD> similar Issues list in DAML:

13:45:55 <JosD> -> 1/2 cardinality

13:46:09 <JosD> -> R Class

13:46:22 <JosD> -> ? equivalentTo

13:46:33 <JosD> -> 1 imports

13:46:47 <JosD> -> * intersectionOf

13:47:01 <JosD> -> * inverseOf

13:47:17 <JosD> -> 2 unionOf

13:47:31 <JosD> -> 1 UniqueProperty

13:47:51 <JosD> Where does Defined Class go?

13:48:23 <JosD> Where do complementOf, UnionOf, etc. go?

13:49:10 <JosD> Guus: propose to move all constructs using unnamed classes to go to level 2

13:52:33 <JosD> PatH: existential quantifiers are already in RDF

13:54:50 <JosD> Guus: make (some) people familiar with notion of implicit class names

13:55:36 <JosD> PatH: not understanding the distinction

13:58:23 <JosD> Jeremy drawing a graph of the example :John a _:1 . _:1 a rdfs:Class; owl:intersectionOf ( :Student :Employee ) .

14:00:05 <JosD> JimH: group seems to be in favor of 1/2, but go back to drawing board to draw the borderline

14:03:57 <JosD> Jeremy: keep 1/2 and ask LANG new proposal ==> 12 people

14:04:32 <JosD> ... versus 1 language ==> 6 people

14:06:21 <em> em is now known as em-scribe

14:08:19 <JosD> ACTION: revisit new level 1/2 proposal by MikeD, EnricoM, ZivH, RaphaelV, IanH and FrankVH

14:08:21 * RRSAgent records action 1

14:09:48 <JosD> (Jim's pinguin example...)

14:14:13 <JosD> RESOLUTIONS: The meaning of an OWL document is conveyed in the RDF graph ==> ALL in favour, NO opposed

14:25:22 <JosD> RESOLUTION: All RDF/XML documents that are equivalent under the RDF REC are equivalent OWL exchange documents ==> 14 in favor -- 3 opposed

14:33:05 <JosD> RESOLUTION: The exchange language for OWL is RDF/XML ==> 16 in favour

14:41:56 <JosD> RESOLUTION: We intend to produce non-normative presentation syntaxes and their mapping to the exchange syntax ==> 16 in favour -- 1 opposed

14:48:39 <DeborahMc> good morning

14:50:51 <JosD> RESOLUTION: The preference of the WG is to produce at least one XML and one frame presentation syntax ==> 11 in favour -- 2 opposed

14:51:58 <JosD> logs are at http://ilrt.org/discovery/chatlogs/webont/2002-04-09.html

14:52:28 <JosD> the last part is about the resolutions

14:53:23 <JosD> ======= how to move forward till next f2f

14:55:55 <heflin> I just joined the telecon. Sorry I'm late.

14:56:29 <JosD> FrankVH gives some explanation about these resolutions

14:59:05 <JosD> JimH: a set of issues will be collected, then ask the group, then go to REAL issue list

15:00:29 <JosD> JimH: issue drive process -- ACTION chairs + MikeS

15:02:34 <JosD> DebMG will join the "revisit new level 1/2 proposal" group

15:06:33 <JosD> discussion about "what should be preserved?"

15:09:16 <JosD> e.g. order, grouping, etc. (within and between presentation syntaxes)

15:11:53 <JosD> NickG: matter of toolbuilders + JimH: to allow some metadata format to describe that

15:17:31 <JosD> discussion about the proposed schedule: 5 in favor -- 0 abstain

15:19:24 <jhendler> Meeting is officially ended w/respect to rsolutions - conversation continures

15:24:44 <jhendler> ADJOURNED

16:17:13 <em-scribe> em-scribe is now known as em

18:46:13 * RRSAgent excuses himself; his presence no longer seems to be needed