Teleconference.2007.11.07/Minutes

From OWL
Revision as of 09:37, 7 January 2008 by AlanRuttenberg (Talk | contribs)

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

See also: IRC log

Previous Minutes

PROPOSED: Accept the previous minutes

Peter Patel-Schneider: +1 to accept minutes
Ivan Herman: +1
Achille Fokoue: +1
Rinke Hoekstra: +1
Ratnesh Sahay: +1

RESOLUTION: minutes accepted

Action item status

Action 4: Send email reminding people to make wiki account (Continued)

Action 5: Sandro to migration of documents [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action01]

Rinke Hoekstra: except for the images
Peter Patel-Schneider: sandro sent out a message concerning where to put the images

Continued, because the figures from the syntax document still need to be moved over

Peter Patel-Schneider: there are *two* documents
Peter Patel-Schneider: both http://www.w3.org/2007/OWL/wiki/Issues and http://www.w3.org/2007/OWL/wiki/HowIssuesAreProcessed tell how issues are processed
Peter Patel-Schneider: should the link to the duplicate page be removed?
Rinke Hoekstra: +1 to removing the (link to the) duplicate page
Rinke Hoekstra: perhaps turn HowIssuesAreProcessed into a redirect to Issues
Achille Fokoue: DONE (maybe after the meeting)
Jeremy Carroll: alan said about reminder email that the naming conventions were unclear, and Sandro is clarifying

Issues

Issue 2 No syntax for AllDisjoint in RDF mapping

Evan Wallace: +1 to Bijan's viewpoint on this

Ian Horrocks: do new properties affect backward compatibility

Evan Wallace: Cardinality Q was also turned down
Bijan Parsia: ooo, excellent point ew
Bijan Parsia: Does peter actually object? Can't we just disagree with the old group?
Carsten Lutz: I disagree. There are usually *many* disjoint classes. Only ontology designers overlook that and
Carsten Lutz: usually fail to state it.
Bijan Parsia: I see peter's point...this is a owl full thing
Peter Patel-Schneider: many large disjointness sets (a disjointness set with thousands of elements)
Peter Patel-Schneider: there can be lots of small disjointness sets

Alan Ruttenberg: the assumption that there will not be a lot of disjoint is not correct

Bijan Parsia: +1 to alan
Carsten Lutz: peter: if you have thousands of classes, why not?
Peter Patel-Schneider: if your disjointness sets are small then you don't get many extra triples
Bijan Parsia: we get this request all the time
Bijan Parsia: NO NO NO NO
Boris Motik: It is not just the disjointness. In the functional spec, you can have equivalences between n classes, but in RDF you can have only pair-wise equivalences. This is inherited from OWL 1.0 RDF mapping.
Peter Patel-Schneider: but it is easy and direct to put a large equivalence set into RDF with no size increase
Boris Motik: The same thing also holds for properties (disjointness and equivalences). It is probably a good idea to come up with the same solution for all of these constructs, not just for disjointness.
Bijan Parsia: +1 to bmotik
Jeremy Carroll: disagree with bmotik
Boris Motik: And there are also sameAs and differentFrom on individuals.
Peter Patel-Schneider: Alan's case seems to be quite compelling

Ian Horrocks: AllDisjoint does not necessary have a negative impact on implementations

Boris Motik: OK, I'll just add this to the issue.
Peter Patel-Schneider: i never had an objection, just a caution
Zakim (Teleconferencing System): jjc, you wanted to mention back chat
Alan Ruttenberg: +1
Peter Patel-Schneider: ok by me
Evan Wallace: +1
Doug Lenat: sounds good to me

Ian Horrocks: should we make the change in the docs for AllDisjoint

Jeremy Carroll: +1
Ivan Herman: not against
Achille Fokoue: +1
Rinke Hoekstra: +1
Zhe Wu: +1
Carsten Lutz: +1
Fabian Neuhaus: +1
Uli Sattler: +1
Michael Smith: +1 for AllDisjoint
Bijan Parsia: +1 for AllDisjoint

RESOLUTION: Issue 2 AllDisjoint will be added in the docs (modified later)

Peter Patel-Schneider: do it the same way that allDifferent is handled in OWL 1.0
Alan Ruttenberg: presume that the rdf mapping will be O(n)
Peter Patel-Schneider: it is in the FS already, right?
Bijan Parsia: pfps, no! it seems!
Bijan Parsia: ah it is
Bijan Parsia: disjointClasses
Bijan Parsia: Here: http://www.w3.org/2007/OWL/wiki/Syntax#Class_or_Description_Axioms
Peter Patel-Schneider: how about: treat DisjointClasses like DifferentIndividuals?
Bijan Parsia: """The disjointClasses axiom takes a set of descriptions and states that all descriptions from the set are pair-wise disjoint."""
Bijan Parsia: It is in
Alan Ruttenberg: RESOLVED: Issue 2 A O(n) rdf mapping of disjointClasses will be added
Peter Patel-Schneider: votes?
Peter Patel-Schneider: we changed the resolution, so it is probably best to confirm the change
Alan Ruttenberg: +1
Peter Patel-Schneider: +1
Bijan Parsia: +1 for current resolution
Achille Fokoue: +1
Doug Lenat: +1
Rinke Hoekstra: +1
Michael Smith: +1 for current resolution
Ratnesh Sahay: +1
Martin Dzbor: +1
Uli Sattler: +1
Jeremy Carroll: +1
Bernardo Cuenca Grau: +1
Ivan Herman: +1
Bijan Parsia: it taeks an arbitrary number
Bijan Parsia: disjointClasses�:= 'DisjointClasses' '(' { annotation } description description { description } ')'
Bijan Parsia: Brackets!

RESOLUTION: AllDisjoint will be added to the functional syntax and the RDF mapping

Issue 3 (anonymous individuals)

Alan Ruttenberg: Individual(type(owl:Thing)) legal in 1.0
Uli Sattler: but this doesn't help us to refer to it?
Alan Ruttenberg: "overzealous"
Jeremy Carroll: Jeremy: was the lack of this a bug? or deliberate?

Jeremy Carroll: no anonymous individual in owl 1.1. Is it a bug?

Peter Patel-Schneider: is a bug

Alan Ruttenberg: tree-like
Alan Ruttenberg: also see http://www.w3.org/2007/OWL/tracker/issues/23
Peter Patel-Schneider: I would like to see a proposal
Bijan Parsia: Ooo, interesting!
Bernardo Cuenca Grau: we don't have the universal role in Owl 1.1

Carsten Lutz: related to having a universal property. should we consider adding universal property?

Bernardo Cuenca Grau: but, it was included in SROIQ
Uli Sattler: which is why we mention this here

Carsten Lutz: it will allow anonymous individuals

Jeremy Carroll: [I would like non-tree like stuff, as well as tree stuff, and this issue is only the first step but ...]

Carsten Lutz: Universal property is more general. It will make anonymous individuals a special case

Carsten Lutz: correct
Uli Sattler: yes
Alan Ruttenberg: in above we are talking about "tree-like" networks of anonymous individuals

Boris Motik: arbitrary anonymous individuals could yield to undecideability

Bijan Parsia: How about close this with tree ones and opening a new issue
Uli Sattler: but there aren't any anon. inds. in the owl1.1? Can you clarify, Boris?
Boris Motik: To be more precise: nontree like anonymous individuals (in an ABox) easily make ontology entailment undecidable
Alan Ruttenberg: Ian and I had a discussion about this a while ago concerning when we could distinguish skolems from bnodes
??: jjc, you wanted to propose resolution Issue 3 is a bug report. Action pfps to fix
Alan Ruttenberg: I will look it up
Jeremy Carroll: jeremy bows to the chair
Alan Ruttenberg: negated property values 1 issue
Bijan Parsia: Action to boris to start the discussion?
Bijan Parsia: Or someone?

Ian Horrocks: we will not be able to resolve it now ... we should continue on emails

Peter Patel-Schneider: make sure the emails include Issue 3 or Issue 23
Jeremy Carroll: +1 to bijan
Carsten Lutz: I am not on the mailinglist
Carsten Lutz: But working on it :)
Jeremy Carroll: (neither am I, but can participate nevertheless
Bijan Parsia: Yes!
Boris Motik: Sire
Bijan Parsia: +1 to boris starting it
Boris Motik: Sure

Action 11: Boris to send an email on issue 3 [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action03]

Issue 11 and Issue 28 (datatype facets)

Bijan Parsia: Is this only for the XML sytnax?
Bijan Parsia: If so, shouldn't we defer until we've decided about the XML sytnax?
Evan Wallace: Its part of the structural spec
Bijan Parsia: http://lists.w3.org/Archives/Public/public-owl-dev/2007JanMar/0127.html
Boris Motik: The reason why we have no several facts is rather arbitrary: we just didn't think of it. I believe we can easily extend the language to be more practicable.
Bijan Parsia: I don't undersatnd this issue ;)
Jeremy Carroll: Why can't we use user names in this syntax ?

Peter Patel-Schneider: I thought the issue was about the XML serialization

Bijan Parsia: We could ask Evren to come on
Bijan Parsia: next week

Ian Horrocks: let's move on since we do not understand it

Boris Motik: The XML schema is now different from what is reporeted in the issue. It was likely a bug that I just fixed later.
Boris Motik: I was talking about issue 11
Bijan Parsia: They both go back to evren's email
Bijan Parsia: Both 28 and 11

Ian Horrocks: Let's nominate someone to start an email discussion

Boris Motik: it is about having multiple facets.

Bijan Parsia: That sounds promising!
Peter Patel-Schneider: suggest sending a message to Evrin to ask him if he thinks 11 is resolved
Jeremy Carroll: move to email

Boris Motik: we should add them it was a bug in the XML Schema ... no problem either for issue 28

Bijan Parsia: close 11 and resolut 28 with action to liberalize the syntax

Action 10: Boris to send an email about issue 11 fixed and how to fix issue 28 [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action04]

Ivan Herman: see Issue 28 notes

Issue 13 and 14 (quotation and CURIES)

Bijan Parsia: This relates to whether we can use the XML Schema syntax
Boris Motik: Where is the XML Schema stored in the working drafts? I see the XML Ser. document, but it has no pointer to the actual schema.
Boris Motik: I tried uploading the XML schema, but the system said that .xsd is not a supported extension
Bijan Parsia: Ok, send an email to sandro
Peter Patel-Schneider: jeremy is probably the closest thing we have to an expert here
Bijan Parsia: I propose to reuse the Turtle string quoting conventions
Peter Patel-Schneider: which are?
Bijan Parsia: Common, comprehensive, easy to lift
Markus Krötzsch: http://www.dajobe.org/2004/01/turtle/#sec-strings
Bijan Parsia: I feel pretty sure that the Turtle spec covers everything RDF can handle
Peter Patel-Schneider: these are roughly the same as RDF quoting
Bijan Parsia: Wait! are we resolving on that?
Peter Patel-Schneider: Yes, wait.
Bijan Parsia: Yes, alanr, curies can represent properites that rdf/xml cannot
Alan Ruttenberg: so can abstract syntax
Bijan Parsia: So can turtle
Peter Patel-Schneider: pointer to rdfa?
Rinke Hoekstra: http://www.w3.org/TR/xhtml-rdfa-primer/
Ian Horrocks: Sorry, but I have to run off
Rinke Hoekstra: http://www.w3.org/TR/rdfa-syntax/
Bijan Parsia: We'd need that for more arbitrary patterns of bnodes
Bijan Parsia: But that's delegated to email discussion
Alan Ruttenberg: PROPOSED: Resolved issue 13 by adopting quotation from turtle (\). Resolve 14 by adopting SPARQL syntax for extended qnames.
Markus Krötzsch: +1
Alan Ruttenberg: +1
Bijan Parsia: +1
Ivan Herman: +1
Elisa Kendall: +1
Rinke Hoekstra: +1
Michael Smith: +1 resolution for issue 13
Uli Sattler: +1
Michael Smith: +1 to resolution for issue 14
Giorgos Stoilos: +1
Peter Patel-Schneider: +1 on the straw poll, but I would like to take a look to finalize
Ratnesh Sahay: +1
Martin Dzbor: +1
Peter Patel-Schneider: I guess that we can invoke the one week rule if necessary

RESOLVED: Resolved issue 13 by adopting quotation from turtle (\). Resolve 14 by adopting SPARQL syntax for extended qnames. (with checkin one written into the spec)

Second F2F

Ian Horrocks: let's talk about the second f2f

Jeremy Carroll: peter: the second f2f seems a long way away but it's not - Jeremy wonders whether it's on the moon.

Peter Patel-Schneider: collocate the 2nd f2f with OWLED

Bijan Parsia: +1
Carsten Lutz: is OWLED collocated with any conference?
Bijan Parsia: Carsten, no
Jeremy Carroll: dates of OWLED again?
Jeremy Carroll: dates: 1-4 April 2 days

Peter Patel-Schneider: Proposed date 1,2,3,4 of April 2008

Peter Patel-Schneider: this is "interim" OWLED which will not be colocated

Ian Horrocks: online poll should be done

Peter Patel-Schneider: any objections for the date?

Bijan Parsia: Who would host?
Ivan Herman: for info: april 21-25 are the dates for WWW2008
Uli Sattler: possibly
Ivan Herman: i will be in china
Alan Ruttenberg: +1
Bernardo Cuenca Grau: possibly
Markus Krötzsch: possibly
Carsten Lutz: -1
Jeremy Carroll: possibly
Peter Patel-Schneider: maybe me, but not too likely
Bijan Parsia: possibly
Fabian Neuhaus: -1
Achille Fokoue: -1
Giorgos Stoilos: possible
Elisa Kendall: -1
Evan Wallace: -1
Ratnesh Sahay: maybe
Rinke Hoekstra: -1
Michael Smith: -1 to www2008
Zhe Wu: -1
Bijan Parsia: So, w3c would host?
Jeremy Carroll: ivan: w3c china could help host

Ivan Herman: would be happy to help with hosting it in China
... if the group decides to do it in China, we can do it

Peter Patel-Schneider: we also need time to do publicity

Ivan Herman: f2f meeting page
Peter Patel-Schneider: have Ian send out an email about this

Bijan Parsia: please add your name in http://www.w3.org/2007/OWL/wiki/Manchester_F2F f2f if you are coming to the first f2f

Action 12: pfps to send an email about your proposal collocated with OWLED [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action05]

Action 13: ianh to send an email about your proposal collocated with OWLED [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action06]


Summary of Action Items

[NEW] ACTION: Boris to send an email about issue 11 fixed and how to fix issue 28 [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action04]
[NEW] ACTION: Boris to send an email on issue 3 [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action03]
[NEW] ACTION: ianh to send an email about your proposal collocated with OWLED [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action06]
[NEW] ACTION: pfps to send an email about your proposal collocated with OWLED [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action05]
[DONE] ACTION: 6 to [recorded in http://www.w3.org/2007/11/07-owl-minutes.html#action02]

[End of minutes]