Warning:
This wiki has been archived and is now read-only.

Teleconference.2007.12.12/Minutes

From OWL
Jump to: navigation, search

See also: IRC log

Admin

Ian Horrocks: agenda amendments; Peter may have suggested one in email?

Peter Patel-Schneider: I don't remember asking, but I can look at my mail logs

Ian Horrocks: no agenda amendments

Ian Horrocks: PROPOSED: Accept Previous Minutes

Ian Horrocks: Hearing no objection

RESOLVED: Accept Previous Minutes

Ian Horrocks: Review of action items status

Ian Horrocks: ACTION-4 Sandro not on phone; postpone until next week.

Ian Horrocks: ACTION-22 Deborah not on the phone; postpone until next week.

Jeremy Carroll: ACTION-23 not resolved; would like to close as not done.

Jeremy Carroll: Jeremy Carroll: action-30 continues

Ian Horrocks: ACTION-31 status, not done yet

Bijan Parsia: lost the track of what is the purpose of this action

Ian Horrocks: we'll abandon the remaining actions from UFDTF

Bijan Parsia: I believe boris did action 36 for me at the f2f

Ian Horrocks: ...because neither Evan dor Deborah are here

Rinke Hoekstra: tomorrow

Bijan Parsia: I didn't do that yet (I yet have to describe imports)

Ian Horrocks and Boris: ACTION-35 done at F2F, action closed

Discussion

Ian Horrocks: What else do we need to do to publish the working draft?

Ian Horrocks: we decided that the attibutions on the documents will be extended to include the editors that will do the document clean-up Ian Horrocks: Sandro needs to run an exportatino script to produce the HTML document

Ivan Herman: We have to wait for Sandro to come back to resolve the issue of short names in the URI of the documents

Alan Ruttenberg: two: things - scribe re: attribution. Notes not naming issues.
Bijan Parsia: I wonder if we can't export the documents before the short name is resolved?
Bijan Parsia: Presumably there's other clean up that needs to be done?

Ivan Herman: Questionable whether the documents will be published before Christmas

Alan Ruttenberg: (Correction) It was decided at the F2F that the attributions will be left as-is, and the authors will be responsible for producing the final documents

Ian Horrocks: Does anything else need to be done? Ian Horrocks: {{{what}}}


Peter Patel-Schneider: Sandro should push the button so that we see what needs to be done at the technical level

Ian Horrocks: good idea

Peter Patel-Schneider: We should try with the syntax document

Ian Horrocks: Boris Motik and Bernardo should do one each for the semantics and RDF mapping

Peter Patel-Schneider: let's do a *quick* and dirty - otherwise we may not get anything done until summer

Ian Horrocks: Sandro will produce the actual html document, and the editors (i.e., authors) will need to do some manual tweaking to the html to make if W3C OK

Ian Horrocks: This might put a lot of work on Sandro

Rinke Hoekstra: I remember Boris agreeing to checking the HTML for any complIan Horrocksce bugs

Alan Ruttenberg: Let's leave this open

Ian Horrocks: Let's leave this item to Sandro and editors (= authors)

Ian Horrocks: we cannot fix a date, because there aer many variables open

Rinke Hoekstra: perhaps it would be good if the publishable docs are available before the meeting so that we can discuss any late minute changes/(de)bugs

Alan Ruttenberg: I've been looking into rdf:type mapping in OWL 1.0

Rinke Hoekstra: wonders too
Peter Patel-Schneider: it is *not* related!

Alan Ruttenberg: removing one mapping rules will allow rdf:type to be used as declarations

Peter Patel-Schneider: this is issue 89, I think, which should be *much* later

Alan Ruttenberg: This was issue 89

Alan Ruttenberg and Ivan Herman: What is the operational definition of a fragment that Jim wants to define?

Alan Ruttenberg: <a href="http://en.wikipedia.org/wiki/Operational_definition">http://en.wikipedia.org/wiki/Operational_definition</a>
Bijan Parsia: notes that in programming language theory (PLT), Op Semantics has a *very* specific meaning

Ian Horrocks: We might look up the definition of operational semantics in Wikipedia, but we need a clearer proposal for a fragment

Bijan Parsia: The current OWL Prime/RDFS 3.0 doesn't give a *grammar*
Peter Patel-Schneider: operational semantics is *very* different from axiomatic semantics

Ian Horrocks: Jim should make a clear proposal that we can all understand

Bernardo Cuenca Grau: The proposal by Jim on RDFS 3.0 is a proper fragment of Horn-SHIQ

Bijan Parsia: I would like to know how bernardo determined that it was a fragment of horn-shiq

Bernardo Cuenca Grau: I determined that it was Horn-SHIQ by a bit of guessing

Peter Patel-Schneider: +1, me too, the semantics could be just about anything, even non-computable
Carsten Lutz: thinks there is a lot of guessing required
Bijan Parsia: Ok, so a bernardo version of RDFS 3.0 is in horn-shiq
Alan Ruttenberg: to clarify - I would think that the operational semantics would define "conformance" not fragments
Peter Patel-Schneider: the WG is *not* in the business of guessing about other's proposals - the proposer is responsible for clarifying

Bernardo Cuenca Grau: I'm not clear about what "operational" means

Zhe Wu: What kind of grammar are you looking for exactly? Zhe Wu: We'd need a forward chaining rules for OWL 1.1


Alan Ruttenberg: My attempt to understand Jim is that his proposal is contrary to his requirement that the language should not depend on the reasoning system

Ian Horrocks: Zhe, will you take an action to come up with a more specific proposal of the syntax and semantics?

Zhe Wu: Yes

Zhe Wu: ACTION: Provide a syntax and a more comprehensive definition for OWLPrime

Jeremy Carroll: We should not limit subsets to syntactic subsets only; we should allow for semantic subsets as well

Jeremy Carroll: the WG is responding to Jim's proposal with antagonism

Alan Ruttenberg: I hear passion, not antagonism, fwiw

Zhe Wu: I'm waiting for Carsten to give me some paper

Peter Patel-Schneider: the pD* paper is readable (actually better than many) the semantic rules are very hard to wade through because of the abbreviations
Ivan Herman: wonders whether a similar rule set is also doable for Horn-SHIQ

Boris Motik: pD* is closely related to DLP - should look at that

Alan Ruttenberg: What do you mean by "support SPARQL query"?

Zhe Wu: I would like the user to be able to query the original query and the inferred information

Bijan Parsia: SPARQL-DL paper: <a href="http://webont.org/owled/2007/PapersPDF/submission_23.pdf">http://webont.org/owled/2007/PapersPDF/submission_23.pdf</a>

Jeremy Carroll: Maybe we should wait for next year when Jim will be on call and then continue with the discussion

Alan Ruttenberg: Can we all agree that pD* is a well-defined fragment?

Peter Patel-Schneider: I don't argue that pD* is ill-defined; who is doing that?

Jeremy Carroll: I heard people suggesting that OWL Prime/RDFS 3.0 was ill-defined
Alan Ruttenberg: no one. I hadn't heard the positive. Now I'm happy
Bijan Parsia: Which it isn't. E.g., there is no grammar
Jeremy Carroll: If pD* is not ill-defined, then maybe Alan or Jim should be linking back to pD* more
Bijan Parsia: But this is true of lots of propsoals

Ian Horrocks: We can continue this discussion via email (accessible to Jim) and next year when Jim will be on call

Issues

Ian Horrocks: We'll look at newly reported issues (we didin't have a mechanism for accepting issues)

Ian Horrocks: Each person raising an issue should explain what an issue is about

Uli Sattler: I do have an idea what ISSUE-85 could mean Uli Sattler: Alan Rector thinks that this could be handled using annotation properties

Ian Horrocks: does this mean we should close the issue?

Alan Ruttenberg: no to closing issue.
Peter Patel-Schneider: yes, I responded to the initial message to the effect that the facility is best handled in tools, with the help of an annotation property

Uli Sattler: What Alan Rector wants is to have a special annotation property that gives hints to the reasoner how to handle certain classes

Alan Ruttenberg: Bijan Parsia, was asking only about pd*. not Jim,Alan's stuff.
Bijan Parsia: And Jeremy Carroll seems to think that asking for clarification is rude or hostile.

Uli Sattler: This property does not change the model theory; it is used only for addictional syntactical checks

Alan Ruttenberg: +1 to accept and discuss

Ian Horrocks: what is the propoal for the issue? Should we accept it and discuss it?

Ian Horrocks: Issue accepted, Uli gets an action to explain the issue more clearly

Uli Sattler: ACTION: Provide more detail about ISSUE-85

Ian Horrocks: ISSUE-86 - that is this about?

Alan Ruttenberg: Need to accept it

Ian Horrocks: ISSUE-87: We agreed at F2F to have rationals, so we'll accept it

Rinke Hoekstra: Mike Smith raised the issue, but he had to leave early
Alan Ruttenberg: Accept on rational.

Uli Sattler: We'll work out a proposal for it

Alan Ruttenberg: do we need an action on that?
Bijan Parsia: We have a draft proposal for the type, but we are to work out e.g., the facets involved

Uli Sattler: doens't mind to accept ISSUE-87 as an action item

Uli Sattler: ACTION: Provide a proposal for the resolution of ISSUE-87

Alan Ruttenberg: We should accept the issue and resolve it when we discuss declarations

Ian Horrocks: When people edit the documents and close issues, they should send chairs an e-mail informing them about it

Ian Horrocks: ISSUE-35 is an examlpe of such an issue

Peter Patel-Schneider: the question was how to quote
Rinke Hoekstra: solution was double quotes for quotes
Bijan Parsia: Could we use e.g., Turtle quoting?
Rinke Hoekstra: Or Java string escapes?
Bijan Parsia: How about SPARQL encoding?
Rinke Hoekstra: link to turtle?
Bijan Parsia: points rinke to: <a href="http://www.dajobe.org/2004/01/turtle/">http://www.dajobe.org/2004/01/turtle/</a>

Ian Horrocks: There was some confusion about the resolution of ISSUE-13, not accept the resolution at this point

Jeremy Carroll: (If we are using turtle syntax elsewhere, then this resolution is inappropriate)

Ian Horrocks: ISSUE-13 will be left on the agenda for next week Ian Horrocks: ISSUE-83


Alan Ruttenberg: We can have this in OWL 1.1 Full. If the editor of OWL 1.1 wants to do this, it is at his discretion

Ivan Herman: Alan Ruttenberg, are you proposing to put a feature into OWL Full without giving it semantics?

Alan Ruttenberg: so, consider issue RESOLVED. Not in OWL DL. In OWL Full by virtue of all syntax. Giving OWL Full semantics not necessary, by precedent of OWL 1.0. However at discretion of OWL Full editor as to whether define semantics.

Ian Horrocks: There is lots of stuff that you can have in OWL 1.1 Full

Uli Sattler: Vipul gave an example, but implementation?

Ivan Herman: There should be some requirement that the features included into OWL 1.1 Full should have some use case requirement, and also some implementation evidence

Alan Ruttenberg: If OWL Full editor decides to give semantics, then yes, give example etc.

Ian Horrocks: We don't suggest to put this into OWL 1.1 Full; we are just not prohibiting it

Bijan Parsia: +1 to people won't implement chains on both sides of a subproperty axiom...sounds like work for nothing for a spec writer

Jeremy Carroll: Let me summarize: Vipul means that this would be useful; OWL DL will not implement it; OWL Full will give it desired meaning, but nobody is likely to implement it

Ian Horrocks: We can't resolve ISSUE-83; Alan Ruttenberg should provide some wording in the spec

Alan Ruttenberg: We should come with a policy for dealing similar issues in future

Alan Ruttenberg: ACTION: Alan and Ian Horrocks to define wording for resolution of issue 83 and similar

Ian Horrocks: ISSUE-29

Jeremy Carroll: we have two concepts and we don't know when to use which

Peter Patel-Schneider: to my mind Jeremy Carroll's message indicates that we can go to rdfs:Datatype

Alan Ruttenberg: Why do we have DataRanges in OWL 1.0?

Peter Patel-Schneider: hmm, I'm not sure that we can't cons up a mapping - consider XML Schema union datatypes

Peter Patel-Schneider: We might use rdfs:Datatype throughout the spec

Jeremy Carroll: maybe maybe - but why make such hard work?
Peter Patel-Schneider: to reduce vocabulary - which some people seem to like
Jeremy Carroll: yes I am one of them ...
Peter Patel-Schneider: OK, I could support owl:datarange being a subcategory of rdfs:Datatype

Boris Motik: we should stick to DataRanges because they fit well with the object model

Jeremy Carroll: We can have anonymous datatypes in RDF; they are awkward but legal

Peter Patel-Schneider: yes, but the primitive datatypes have URIs and this is all that is needed

Ian Horrocks: please continue this discussion via email and try to find a resolution. In view of time, let's move on to AOB.

AOB

Ian Horrocks: report from RIF meeting from Bijan Parsia

Jeremy Carroll: Undef AOB - I could mention POWDER

Ian Horrocks: no we don't have the report

Ivan Herman: Did we agree on next meeting?

Bijan Parsia: wonders if he means F2F?

Ian Horrocks: Let's go to any other business

Ivan Herman: When should we have the next telco?

Ian Horrocks: Next week we should have a meeting

Ian Horrocks: We should have a meeting next week (this is too early for Christmas)

Alan Ruttenberg: But no meeting on the 26th

Jeremy Carroll: Some POWDER issues are related to OWL

Bijan Parsia: POWDER: <a href="http://www.w3.org/2007/powder/">http://www.w3.org/2007/powder/</a>
Bijan Parsia: wonders if POWDER is on our dependancie list

Ian Horrocks: POWDER is not on the dependecy list, but we might want to use some features in OWL 1.1

Jeremy Carroll: POWDER people want annotation on axioms; we might look for inspiration over there

Ivan Herman: POWDER is not on the dependency list because the charter of POWDER says that everything as the level of RDF

Peter Patel-Schneider: OWL uses POWDER features? surely it is POWDER that uses OWL

Jeremy Carroll: Will report next week to OWL-WG on progress

Jeremy Carroll: yes correct peter
Ivan Herman: peter, yes: POWDER uses OWL


Summary of Action Items

[NEW] ACTION: Zhe to provide a syntax and a more comprehensive definition for OWLPrime
 

[NEW] ACTION: Uli to provide more detail about ISSUE-85
 

[NEW] ACTION: Uli to provide a proposal for the resolution of ISSUE-87
 

[NEW] ACTION: Alan and Ian Horrocks to define wording for resolution of issue 83 and similar