Teleconference.2007.11.28/Minutes

From OWL
Revision as of 11:21, 13 December 2007 by IanHorrocks (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

See also: IRC log

Bijan Parsia: I could report on rif meeting

ADMIN

comments on the minutes?

minutes accepted

Alan Ruttenberg: action items ...

Action 4 continued

Bijan Parsia: alanr, I could take the imports action ifyou'd like
Bijan Parsia: ok
Jeremy Carroll: yes - I claim victory on action
Jeremy Carroll: (I put the action number in the e-mail so the tracker tracks)

Action 27 is close

Michael Smith: yes

Action 29 is close

Alan Ruttenberg: any further suggestions of the f2f agenda?

Conrad Bock: Where is the draft agenda?
Alan Ruttenberg: http://www.w3.org/2007/OWL/wiki/F2F.2007.12-Manchester/Agenda

Alan Ruttenberg: we might need to move the import and rich annotation to first day

Bijan Parsia: Can we swap fragments and imports and RAs?

... since alan won't be in day 2

Ian Horrocks: we could discuss it offline, there are other constraints

Jeremy Carroll: +1

Alan Ruttenberg: we should move the user facing part to later part of the day

Alan Ruttenberg: http://www.w3.org/2007/OWL/wiki/Manchester_F2F
Jeremy Carroll: Elisa and Deborah are remote participants in UFDTF

Teleconference 2007.12.05 ?

Jeremy Carroll: -1

Alan Ruttenberg: I cannot [make it]

Peter Patel-Schneider: there is already a conflicting event on the 5th
Deborah McGuinness: i think jim h wants to participate in the ufdtf as well
Alan Ruttenberg: -1
Bijan Parsia: It's traditional not to have such telecons
Jeff Pan: -1
Bijan Parsia: -1
Michael Smith: -1 on telecon on 5th
Ivan Herman: -1
Bernardo Cuenca Grau: -1
Peter Patel-Schneider: -1

RESOLVED: no telecon on Dec 5th

Discussions (35 min)

Publication Schedule and First Working Drafts

Bijan Parsia: Functional Syntax and Structural Spec, Semantics, Rdf mapping
Jeremy Carroll: (for me - the issue list is in good shape for reflecting the lack of consensus)
Bijan Parsia: Were the original proposal
Jeremy Carroll: q+ to suggest shading ...
Zakim: jeremy, you wanted to suggest shading ...
Bijan Parsia: We have some issues that touch everything so "shady" per se won't work

Jeremy Carroll: editors might want to relate part of the documents with issues in the issue list

Bijan Parsia: E.g., Typed constructs

Ian Horrocks: we need to see whether it is practical

Jeremy Carroll: +1 to sandro as an alternative
Alan Ruttenberg: +1
Jeremy Carroll: to my suggestion!
Jeremy Carroll: for me yes
Jeremy Carroll: (yes to RDF Mapping with such riders)
Peter Patel-Schneider: how about putting the disclaimer in the document status, pointing to the issue list?
Peter Patel-Schneider: the mapping between sections and issues is not particularly easy

Alan Ruttenberg: should issues be written in the beginning part of the sections?

Boris Motik: some issues are more meta issues

Jeremy Carroll: A list of other issues might be good

Alan Ruttenberg: we meant the specific ones

Bijan Parsia: I think we're in editor's choice space
Bijan Parsia: some issues are minor, don't need huge highlighting
Jeremy Carroll: I agree with bijan
Alan Ruttenberg: we don't have editors yet
Bijan Parsia: Some are important and controversial

Boris Motik: I meant more issues are meta ones

Bijan Parsia: And need to be highlighted
Ian Horrocks: We don't need to insist that every issue is mentioned in some document
Alan Ruttenberg: +1 to Ian
Jeremy Carroll: I agree with bijan again!
Bijan Parsia: How about we task someone(s) to add some issue texts to a document?

Boris Motik: and who should do this?

Bijan Parsia: And experiment a bit

Alan Ruttenberg: we could assign them to some people, including me

Ian Horrocks: which document?

Alan Ruttenberg: does not matter

Sandro Hawke: you still need to be familiar with the documents enough to do the job properly

Bijan Parsia: Yes
Bijan Parsia: On behalf of carsten

Ian Horrocks: do we need to have such fine-grain issule list mechanism

Deborah McGuinness: +1 to grouping

Deborah McGuinness: it is also hard to follow


Bijan Parsia: I'm big +1 on grouping issues loosely
Jeremy Carroll: -1 to grouping
Boris Motik: +1 to grouping (not all, but some should be grouped)
Peter Patel-Schneider: we are supposed to be spending 8 hours a week on this, minimum
Jeremy Carroll: (peter's days might be longer than mine! 7.5 hrs)

Sandro Hawke: do we want a hierarchical issue list?

Bijan Parsia: I tried this on rich annotations...addding all the issues related to annotations
Bijan Parsia: Peter complained about this :)
Jeremy Carroll: We can group issues by 'product'

Sandro Hawke: we should take a simplified approach

Bijan Parsia: How about soem web pages for this

Ian Horrocks: we just started but we have 81 issues already

Zakim: jeremy, you wanted to explain a separation
Bijan Parsia: Issues are a tool for the chairs
Uli Sattler: ha ha
Bijan Parsia: The chairs should figure out what works for them
Jeremy Carroll: lots of small issues are easier to make progress on than a few big issues [Scribe assist by Jeremy Carroll]

Alan Ruttenberg: a quick poll to see if the solution is acceptable by everyone

Alan Ruttenberg: three documents: syntax, semantics and mapping to RDF


Jeremy Carroll: no we cannot decide to publish docs until they are ready
Jeremy Carroll: I can confirm informally the plan
Bijan Parsia: actually, we can

PROPOSED: Annotate the Syntax, Semantics, and Mapping-to-RDF documents with links to all issues, as they might affect the text....

Peter Patel-Schneider: we can publish subject to editorial changes
Bijan Parsia: We can say, "We will publish these docuemnts pace soem editoral mod...

PROPOSED: Annotate the Syntax, Semantics, and Mapping-to-RDF documents with links to all issues, as they might affect the text, and then (give or take editorial changes) we will publish.

Jeremy Carroll: i guess i bow to pfps and bijan here, but some rider
Peter Patel-Schneider: +1, as long as I'm not doing the work!
Alan Ruttenberg: +1
Ian Horrocks: would prefer to say "issues" rather than all issues
Jeremy Carroll: ok
Achille Fokoue: +1
Uli Sattler: +1
Deborah McGuinness: +1
Ivan Herman: +1
Bijan Parsia: +1 with "issues"
Zhe Wu: +1
Conrad Bock: -1
Martin Dzbor: +1
Ratnesh Sahay: +1
Michael Smith: +1
Ian Horrocks: +1 otherwise
Giorgos Stoilos: +1
Jeff Pan: +1
Sandro Hawke: -0 concerned about deciding before deciding, but it seems vaguely okay.
Carsten Lutz: Regarding my issues discussed here:
Carsten Lutz: - I saw lots small issues before, so I was just trying to follow common practice
Bernardo Cuenca Grau: +1
Carsten Lutz: - Several of my issues I wouldn't call "small"
Carsten Lutz: - I can live with grouping as long as decisions are taken individually (this
Carsten Lutz: is why I separated them)
Carsten Lutz: +1
Boris Motik: -0 This might be a lot of work
Evan Wallace: 0
Uli Sattler: -0 is the slightly more negative vote than the neutral o?
Evan Wallace: +1 with metamodel issue included
Carsten Lutz: -o ?

Sandro Hawke: I don't think we should have a resolution here

Jeremy Carroll: (I find Sandro compelling, but can't say why)

... but it seems that it is good enough to proceed?

Peter Patel-Schneider: that's an action, not a resolution
Bijan Parsia: It's his air of mystery

Sandro Hawke: I think people need to review how it is done

Bijan Parsia: I think the difference between a blessing and vote is not good
Jeremy Carroll: (Personally I would like to see the post-Wiki doc before the vote)
Sandro Hawke: So the only future discussion will be about issue links.
Jeff Pan: action or resolution?
Evan Wallace: I'm confused now
Evan Wallace: Are we to do an up or down vote later?

Alan Ruttenberg: no resolution but general agreement to the direction

Evan Wallace: yes
Jeremy Carroll: The chairs appoint editors

Ian Horrocks: chairs may delegate the tasks

Bijan Parsia: I hate the idea, but I think UManchester can mobilize some resources to this end
Bijan Parsia: (er.. the idea of workign on it :)
Jeremy Carroll: I feel the discussion needs booze
Conrad Bock: Alan, I filed the metamodel issue on Syntax at http://www.w3.org/2007/OWL/tracker/issues/82.
Alan Ruttenberg: thanks, conrad
Bijan Parsia: Hmm. I think it is a fragment

Mapping to/from RDF

Boris Motik: owl dl is not 100% sub-language of owl full

Bijan Parsia: I'm looking at 5.3...what's the difference?
Ian Horrocks: Also, even to the extent that we believe that it is a fragment, this is a belief/conjecture and not a proven fact
Bijan Parsia: to reply to Boris

Bijan Parsia: it seems that part of the issue what counts as compatibility

Bijan Parsia: your idea is a stronger version of compatibility

Alan Ruttenberg: There is a strictly increasing chain of sets of entailments
Alan Ruttenberg: >
Alan Ruttenberg: > Simple <= RDF <= RDFS <= D- <= OWL DL <= OWL Full
Alan Ruttenberg: >
Alan Ruttenberg: > It was a design goal that the last strict inclusion should be an equality; see
Alan Ruttenberg: >
Alan Ruttenberg: > http://www.w3.org/2001/sw/WebOnt/webont-issues.html#I5.3-Semantic-Layering
Alan Ruttenberg: > http://lists.w3.org/Archives/Public/www-archive/2002Oct/0022.html
Alan Ruttenberg: >
Alan Ruttenberg: > In other words (subject to the syntactic constraints of OWL DL) the
Alan Ruttenberg: > entailments of OWL DL are exactly those of OWL Full. This goal
Alan Ruttenberg: > slipped a little: see
Alan Ruttenberg: >
Alan Ruttenberg: > http://www.w3.org/TR/owl-test/byFunction#function-Thing
Deborah McGuinness: +1 to getting clear on a working defn of compatibility
Jeremy Carroll: OWL 1.0

Ian Horrocks: we had similar discussion in WebOnt

Uli Sattler: but the entailments aren't the same.
Jeremy Carroll: The two known exceptions are size of universe and certain entailments involving annotations

Alan Ruttenberg: I want more compatibility than what is between RDF and OWL

Peter Patel-Schneider: where are the new sources of "incompatability"?
Bijan Parsia: I agree with alanr...the corner case semantic alignment is much less important than bringing in more wffs

Deborah McGuinness: it is a good idea to clarify compatibility

Bijan Parsia: That was a goal of, e.g., adding punnings
Bijan Parsia: I.e., to align to owl full
Peter Patel-Schneider: my understanding is that semantic-full annotations are the result of not having black triples in RDF
Peter Patel-Schneider: s/black/dark/
Alan Ruttenberg: s/dark/graphite/
Bijan Parsia: I thought we *did* implement anntotions per the spec...I would love a pointer to where this is not the case
Peter Patel-Schneider: there is a solution to backward compatability wrt annotations (but I don't like it)
Bijan Parsia: I would love there to be a unity OWL

Boris Motik: one design principle of OWL 1.1 is that ...

Boris Motik: annotations have no semantics

Jeremy Carroll: who are users???

Boris Motik: comments are simply comments

Ian Horrocks: in the charter we say to refine and extend OWL

Alan Ruttenberg: three cases of disparity: annotations, punning, (forget third)
Alan Ruttenberg: third= bnodes
Jeremy Carroll: wrong - strong disagreement
Alan Ruttenberg: wrong what, jeremy?

Ian Horrocks: an OWL document should be one that could be translated from the one written in abstract syntax

... according to the spec

Boris Motik: Yes, there is
Bijan Parsia: There seem to be test cases
Ian Horrocks: Sean's work was *excellent*, but it shouldn't have been *needed*
Boris Motik: Ontology O1 does not entail O2 even if O1 and O2 are the same save for annotations.

Jeremy Carroll: the semantics document is hard to get interpretation

Ian Horrocks: this is completely different order

... the semantics document is technical

... and understood by many people

Zakim: alanr, you wanted to ask how we can tell whether annotations are correctly implemented
Alan Ruttenberg: Does any tool implement ontology entailment?
Bijan Parsia: Yes
Bijan Parsia: Pellet
Boris Motik: We don't know even if it decidable
Alan Ruttenberg: Pellet has it commented out last time I checked
Bijan Parsia: ?
Bijan Parsia: No
Bijan Parsia: could do more on this
Boris Motik: No need to continue now. We should have a strategy to address this issue soon.
Bijan Parsia: $ Usage: java Pellet OPTIONS
Bijan Parsia: {-if,-inputFile} <file URI> Input ontology URI
Bijan Parsia: {-is,-inputString} string A string representation of the input file
Bijan Parsia: -cf,-conclusionsFile} <URI> Check if all the triples in this ontology is
Bijan Parsia: entailed by the input ontology
Bijan Parsia: {-cs,-conclusionsString} str A string representation of the conclusions fil

Issues (35 min)

PROPOSAL: Base abbreviated URIs in Functional-style Syntax on CURIES not QNAMES (see Issue 14)

PROPOSED: Base abbreviated URIs in Functional-style Syntax on CURIES not QNAMES (see Issue 14)

Alan Ruttenberg: +1
Ian Horrocks: +1
Peter Patel-Schneider: +1 Issue 14 proposal
Michael Smith: +1
Jeremy Carroll: HP abstain (but no desire for further discussion)
Ivan Herman: +1
Evan Wallace: +1
Bernardo Cuenca Grau: +1
Ratnesh Sahay: +1
Achille Fokoue: +1
Jeff Pan: 0
Boris Motik: 0
Deborah McGuinness: 0
Martin Dzbor: 0
Sandro Hawke: 0
Boris Motik: Could someone please point out the main benefit of using CURIES?

Ivan Herman: it seems that we could do that without too much problems

Peter Patel-Schneider: curies allow abbreviating all IRIs
Peter Patel-Schneider: qnames don't
Jeremy Carroll: The main benefit is ids like http://ex.og/id/9324

RESOLVED: Base abbreviated URIs in Functional-style Syntax on CURIES not QNAMES (see Issue 14)

PROPOSED: Structural specification, XML Schema, and RDF mapping be extended to allow for multiple facet--value pairs in data ranges (see Issue 28)

Alan Ruttenberg: +1
Jeff Pan: +1
Jeremy Carroll: +1
Evan Wallace: +1
Uli Sattler: +1
Peter Patel-Schneider: +1
Michael Smith: +1
Ian Horrocks: +1
Deborah McGuinness: +1
Bijan Parsia: +1
Martin Dzbor: +1
Giorgos Stoilos: +1
Ivan Herman: what I said is (for the minutes) we can normatively refer to the CURIE document (http://www.w3.org/TR/curie) [Scribe assist by Ivan Herman]
Bernardo Cuenca Grau: +1
Sandro Hawke: 0
Ivan Herman: +1
Achille Fokoue: 0
Boris Motik: +1

RESOLVED: PROPOSAL: Structural specification, XML Schema, and RDF mapping be extended to allow for multiple facet--value pairs in data ranges (see Issue 28)

PROPOSED: Issue 64 be resolved as per this email

Michael Smith: the email http://lists.w3.org/Archives/Public/public-owl-wg/2007Nov/0391.html

object property chains in triples: confusion of list with property

Alan Ruttenberg: +1
Boris Motik: +1
Ian Horrocks: +1
Peter Patel-Schneider: +1
Sandro Hawke: 0
Achille Fokoue: +1
Michael Smith: +1
Bernardo Cuenca Grau: +1
Deborah McGuinness: +1
Evan Wallace: 0
Martin Dzbor: +1
Uli Sattler: +1
Jeff Pan: 0
Ivan Herman: 0 (do not remember the detail, but no reason for discussion for me)
Jeremy Carroll: +1
Alan Ruttenberg: q?
Bijan Parsia: +1

RESOLVED: Issue 64 be resolved as per [this http://lists.w3.org/Archives/Public/public-owl-wg/2007Nov/0391.html] email

Jeremy Carroll: q+ to ask about publications
Alan Ruttenberg: ack jeremy
Zakim: jeremy, you wanted to ask about publications
Jeremy Carroll: OK
Alan Ruttenberg: Issue 2 revisited: RDF syntax for other "n-ary constructs"? (See [1] and thread.)
Jeremy Carroll: do the issues resolved go into about to be published docs [Scribe assist by Jeremy Carroll]
Alan Ruttenberg: yes [Scribe assist by Jeremy Carroll]
Boris Motik: Yes
Boris Motik: Yes
Alan Ruttenberg: q+ alanr
Alan Ruttenberg: ack pfps
Boris Motik: q+
Jeremy Carroll: I am against
Uli Sattler: against what
Uli Sattler: ?

Peter Patel-Schneider: I am not conviced that it would solve the round-tripping problem

Jeremy Carroll: against more vocab
Peter Patel-Schneider: what about the case for n=2?
Bijan Parsia: With suitable annotations, we could round trip without new vocab!
Peter Patel-Schneider: agreed
Ian Horrocks: against adding explicit rdf syntax for all n-ary constructs

Boris Motik: the spec should be more implementable

Bijan Parsia: need to make things easier

Bijan Parsia: Oh, I have users who really like nice serializations...that was a big deal
Bijan Parsia: So I would like to reduce *parser* freedom
Bijan Parsia: (sorry, illness making me think slow...)
Bijan Parsia: IT's not just an implementator issue, but a user issue
Bijan Parsia: Not if you use the binary constructs

Alan Ruttenberg: if so, then an OWL DL ontology could become something not a proper OWL DlL ontology after some parsing

Bijan Parsia: I think Protege4 is using some aggressive pretty printing
Jeremy Carroll: (for me this is hitting the bigger issues about what are the goals ....)
Ian Horrocks: Yes, I guess that P4 "helps" you by turning multiple binary into an nary?
Alan Ruttenberg: this is not a help
Bijan Parsia: It may
Peter Patel-Schneider: but what if you are building an OWL 1.0 ontology in Protege? then you would generate OWL 1.1, not OWL 1.0, in the RDF
Bijan Parsia: Yep...this may be a bug in the serializer
Uli Sattler: ...but this is about P4 and not about "what is possible/should be done"
Ian Horrocks: but this is a tool issue

is this n-ary thing done by Protete using the well known design pattern for n-ary relations?

Ian Horrocks: P4 should allow you to explicitly save as 1.0
Bijan Parsia: It should respect how you construct
Ian Horrocks: See numerous Microsoft tools for details
Bijan Parsia: That's the intend of the roundtripping design

Alan Ruttenberg: annotation properties would be changed too

Uli Sattler: But Alan, these are still P4 issues -- or are any of them due to OWL1.1?
Jeff Pan: (maybe we should take this offline)

Ian: I think this is a tool issue rather than a language issue

... tools supporting OWL 1.1 will produce OWL1.1 files by default

Bijan Parsia: But you need to distinguish the tool issues from langauge issues
Jeremy Carroll: Please let me know when we are on topic
Ivan Herman: :-)
Jeremy Carroll: and I will rejoin
Achille Fokoue: +1 for Ian's view that it is a tool issue
Bijan Parsia: Round tripping, as boris described, should work fine. If protege does some munging in its serialization...that's different

rool call on Issue 2?

Alan Ruttenberg: roll call on Issue 2 revisited?
Jeremy Carroll: -1
Boris Motik: +1 to extend all n-ary constructs to n-ary

Issue 2 revisited: RDF syntax for other "n-ary constructs"? (See [1] and thread.)

Bijan Parsia: Yeah, what'st he resolution?
Uli Sattler: did we not already do this call?
Peter Patel-Schneider: no proposal was in the agenda
Ian Horrocks: +1 to either extend all n-ary constructs to n-ary or remove them from functional syntax
Bijan Parsia: IanH, really?
Bijan Parsia: Including AllDisjoint?
Ian Horrocks: yes -- I want the various syntaxes to be properly aligned
Jeremy Carroll: AllDisjoint was solved and agreed
Bijan Parsia: That was my understanding too, jeremy
Bijan Parsia: IanH do you propose reopening AllDisjoint? What about AllDifferent?
Uli Sattler: mine too, Jeremy
Ian Horrocks: My preference would be to extend the RDF, not emasculate the functional syntax
Alan Ruttenberg: alldifferent is already in this form in 1.0 iirc
Ian Horrocks: I don't ropose reopening AllDisjoint? What about AllDifferent? -- they are in alignment
Jeremy Carroll: yes
Peter Patel-Schneider: what about objectunionof, objectintersectionof?
Bijan Parsia: Can we move this to email?

Ian, I would like to see some alignment between the two syntax

Evan Wallace: agree with Ivan, but can we have an action to record where we are on this
Bijan Parsia: sure
Ivan Herman: :-)
Alan Ruttenberg: open to suggestions of which one might have been easier :)

RIF meeting

Bijan Parsia: there is a draft on compatibility between RIF and OWL

Alan Ruttenberg: apologies for taking liberty with chair
Peter Patel-Schneider: http://www.w3.org/2005/rules/wg/wiki/SWC/OWL-Compatibility

... people suggested RIF should talk to us on it

... proposed a joint TF

... between the two WGs

Alan Ruttenberg: do we need nominate someone?

Sandro Hawke: we didn't remember the exact number

Bijan Parsia: {{{what}}}