Teleconference.2008.01.09/Minutes

From OWL
Jump to: navigation, search

These minutes have been approved by the Working Group and are now protected from editing. (See IRC log of approval discussion.)

See also: IRC log



Vojtech Svatek: new to W3C, interested in ontology patters/matching

Doug Lenat: welcome, Vojtech.

... worked some on the rich annotations

Alan Ruttenberg: agenda ammendments? [none received]


PROPOSED: accept minutes of 2007-12-19

Doug Lenat: +1
Peter Patel-Schneider: +1 to 12/19 minutes

RESOLVED: accept minutes of 2007-12-19

PROPOSED: accept minutes of 2008-01-02

Peter Patel-Schneider: +1 to 1/2 minutes

RESOLVED: accept minutes of 2008-01-02

Bijan Parsia: +1 to 1/2 minutes

PROPOSED: accept minutes of manchester face to face

Jim Hendler: +0
Peter Patel-Schneider: +0 to F2F (there are still missing pointers to presentations)
Boris Motik: +1
Doug Lenat: Jim, why +0?
Jim Hendler: abstaining because I did not attend the f2f
Jeremy Carroll: ahh, I didn't understand that
Elisa Kendall: +1 - for the part I attended by phone
Alan Ruttenberg: -1
Bijan Parsia: Do we have a list of the presentations?
Bijan Parsia: q+

general discussion of minutes and presentations

Alan Ruttenberg: let's postpone to next week and we have time to make sure presentations are linked right

Alan Ruttenberg: ok, please fix this up during the week - I could put presentation pointers in

Jeremy Carroll: http://lists.w3.org/Archives/Public/public-owl-wg/2007Dec/att-0097/dl-and-full.pdf
Jeremy Carroll: were my slides
Peter Patel-Schneider: PDF is OK for now

ACTION: alanr will try to get links as best as possible, and to get slides right

ACTION: Alan will try to get links updated and link slides int

ACTION: Alan will try to get links updated and link slides int

Boris Motik: bmotik_ has joined #owl

working drafts are out, Alan thanks everyone

Evan Wallace: Thanks to editors.

and there is much rejoicing

Boris Motik: bmotik_ has left #owl

Alan Ruttenberg: review actions - I'd like to consider these closed

Boris Motik: bmotik_ has joined #owl

ACTION: Alan to add test guidelines to working group to agenda for next week

RESOLVED: all pending review actions on the WIKI are closed [Alan will close]

scribe notes these are actions 51,52, and 53

Alan Ruttenberg: Adding a datatype to represent rational numbers

Action Review

Michael Smith: http://www.w3.org/2007/OWL/wiki/OWL_Rational

Uli Sattler: working on this, problem is we still are slightly unclear about

Bijan Parsia: And algebreic numbers

... do we want rationals, rationals and reals, or what?

Alan Ruttenberg: issue x^2=2 unsatisfiable for rationals.

... we're currently looking at a way to go that is a little simpler

... I think we can go further, but I would like some feedback


Alan Ruttenberg: we need a proposal so we can discuss it

Jeremy Carroll: q+ to propose accepting action is done
Michael Smith: +1 to closing the *action*


Bijan Parsia: I'm unclear on where things stand, and there's some related issues - I'd like time to review

Alan Ruttenberg: action is postponed, Bijan will work w/Uli


Evan Wallace: I like rationals

Bijan Parsia: need to discuss in context of n-ary. Can we discuss with that aproach

Bijan will lead discussion on WG email, action 56 is postponed

Alan Ruttenberg: who would like to be on a task force on imports

Jonathan Rees: action58

Peter Patel-Schneider: +1 import - no more evenings CET please
Alan Ruttenberg: +1
Boris Motik: +1(no restrctions to time)
Jeremy Carroll: +1
Rinke Hoekstra: +1 (no evenings)
Ian Horrocks: +1 and agree about CET evenings
Michael Smith: +1 for import TF

Alan Ruttenberg: will set up TF

Bijan Parsia: +1 (but oh I'm sad to do so)

Sandro Hawke: suggests use of WBS for surveying times

alanr and sandro will discuss

action58 postponed

Bijan Parsia: -1 to being on imports task force (changed my mine)
Jeremy Carroll: please continue Action 48 - I have finally scheduled time to do it

action46 continued

FPWDs should be announced widely

Alan Ruttenberg: what relevant lists should we use, who will post

Ian Horrocks: we should do as soon as possible

Alan Ruttenberg: what lists, please let us know on irc:

Sandro Hawke: semantics-web@w3.org, I'm willing to do it.


Sandro Hawke: http://www.w3.org/2007/OWL/wiki/Publicity
Rinke Hoekstra: Should there be some standard announcement text?
Jeremy Carroll: -1 to standard text
Rinke Hoekstra: ok

Ian Horrocks: set up a page in Wiki, if you see a list that isn't included, add the list to the wiki page, and send the announcement - you can use the standard announce from W3C or something more community relevant

Jeremy Carroll: +1 to sandro, customized announcement for community
Sandro Hawke: http://www.w3.org/2007/OWL/wiki/Publicity now created, with one entry

Alan Ruttenberg: what should we do with UFDTF

Elisa Kendall: I'm also interested in UFDTF, and would like to be included if possible
Evan Wallace: i will be available next week

James Hendler: Deb, Vipul, me and a couple of others were also on the list, weren't we?

Elisa Kendall: There should be a list in the wiki - Jeremy had chaired a number of calls
Peter Patel-Schneider: membership is supposed to be AlanRuttenberg BijanParsia DebMcGuinness EvanWallace JeremyCarroll JimHendler VipulKashyap MartinDzbor Peter F. Patel-Schneider Elisa Kendall

Alan Ruttenberg: I will poll folks, look for a time we can do it

Peter Patel-Schneider: from http://www.w3.org/2007/OWL/wiki/UFDTF


Jeremy Carroll: page has list of people, my recollection is we were going to move earlier - but same day (Mon)

Alan Ruttenberg: I'll make sure to get more people in

Jeremy Carroll: two hours earlier

issue proposals

Alan Ruttenberg: issue83 - propose to close as resolved with text as written

Alan Ruttenberg: the issue has to make some words w/respect to the relation between DL and Full

Evan Wallace: +1
Jeremy Carroll: +1 (noting OWL Full semantics is likely)

PROPOSED: close (as RESOLVED) Issue 83 (Property Chain Axiom: P1 o P2 => P2 o P1 ) as per email

Peter Patel-Schneider: +1 to resolve Issue 83 as proposed
Rinke Hoekstra: +1
Jim Hendler: +1
Boris Motik: +1
Uli Sattler: +1
Carsten Lutz: +1
Zhe Wu: +1
Alan Ruttenberg: http://lists.w3.org/Archives/Public/public-owl-wg/2008Jan/0022.html
Bijan Parsia: +1
Alan Ruttenberg: +1
Sandro Hawke: +0
Jeremy Carroll: q+ to comment before closure
Michael Smith: +1 to closing Issue 83 as proposed
Doug Lenat: +1
Elisa Kendall: Elisa +1
Markus Krötzsch: +1 to close
Jeff Pan: +0

Jeremy Carroll: there may be some issues with the OWL Full 1.1 semantics which might cause us to reopen later

Alan Ruttenberg: new info can always let us reopen an issue

discussion of http://lists.w3.org/Archives/Public/public-owl-wg/2008Jan/0022.html

Jeremy Carroll: just nervous that OWL Full semantics could mean we change things later

Jeremy Carroll: My assumptions in voting for this, is that an OWL Full semantics will be quite easy to produce for this


Jeremy Carroll: I doubt there will be any OWL Full issues

Ian Horrocks: the email says the WG might or might not give the OWL 1.1 full semantics, so closing text doesn't require this

Alan Ruttenberg: we have general approval with a couple of abstentions

Jeremy Carroll: I am expecting to make a good effort at an OWL Full semantics and expect that to be adequate for this

RESOLVED: close (as RESOLVED) Issue 83 (Property Chain Axiom: P1 o P2 => P2 o P1 ) as per email

Alan Ruttenberg: discussion of Issue 55

Alan Ruttenberg: we propose to close this as postponed

PROPOSED: close (as POSTPONED) Issue 55 (owl:class v. rdfs:class) per email

Boris Motik: +1
Jeremy Carroll: +1 to postpone
Bullwinkle - err I mean Hendler: +1
Alan Ruttenberg: http://lists.w3.org/Archives/Public/public-owl-wg/2008Jan/0023.html
Uli Sattler: +1
Doug Lenat: +1
Jim Hendler: oops, I meant -0
Zhe Wu: +1
Bernardo Cuenca Grau: +1
Michael Smith: +1

PROPOSED: close (as POSTPONED) Issue 55 (owl:class v. rdfs:class) per http://lists.w3.org/Archives/Public/public-owl-wg/2008Jan/0023.html

Jeff Pan: +1
Doug Lenat: +1
Peter Patel-Schneider: -1 to postpone as I see neither technical nor resource issues to support postponement over resolution
Bijan Parsia: +1 but I'll go on record here that I think "POSTPONED" is substantively meaningless; I don't see that a closed issue is "more" closed than a POSTPONED; but if people feel better with "POSTPONED" I don't care as long as it *is* meaningless - and anyway, if I keep arguing the scribe might put words in my mouth
Evan Wallace: +1


Markus Krötzsch: +1
Rinke Hoekstra: +1 with bijan's sidenote
Elisa Kendall: +1
Alan Ruttenberg: Following a discussion with Ian, in which we acknowledge Peter's
Alan Ruttenberg: comment below and subsequent discussion on the mailing list, and
Alan Ruttenberg: Jim's desire to postpone this issue, Ian and I propose that we close
Alan Ruttenberg: the issue by postponing it, noting Peter's comment.


Peter Patel-Schneider: I will not be registering a formal objection

Peter Patel-Schneider: votes against, but does not request further discussion


Sandro Hawke: (I am very confused -- I thought -1 was a formal objection. But this doesn't seem like the issue on which to get into that.)
Sandro Hawke: (basically, Peter, I think what you mean to say is "-0")

RESOLVED: close (as POSTPONED) Issue 55 (owl:class v. rdfs:class) per http://lists.w3.org/Archives/Public/public-owl-wg/2008Jan/0023.html


Should we keep or abandon data/object property punning

Alan Ruttenberg: several issues have come up linked together with respect to punning - esp around cardinality restrictions

Alan Ruttenberg: so some people said maybe we should drop this kind of punning

Boris Motik: note that IMO this issue is not just about datatypes, but is general to our approach to typing (cf emails on Dec 15 2007)

Boris Motik: http://lists.w3.org/Archives/Public/public-owl-wg/2007Dec/0184.html


Boris Motik: http://lists.w3.org/Archives/Public/public-owl-wg/2007Dec/0185.html


Boris Motik: http://lists.w3.org/Archives/Public/public-owl-wg/2007Dec/0186.html

Peter Patel-Schneider: in some sense I echo Boris' comment, I don't see how abandoning this kind of punning fixes Issue 65

Bijan Parsia: I don't see either

Alan Ruttenberg: as I understand it - there's a relation between vocabulary for easy local typing, as opposed to just the punning

Alan Ruttenberg: (notes Jon Reese is sitting in on this portion of the call)

Alan Ruttenberg: asks about declarations

Peter Patel-Schneider: not local - each use of a term has to be typed

Bijan Parsia: describes issue w/respect to a restriction - how do I know if this is being used as a data or object type

Bijan Parsia: so problem is we cannot "contextualize" use of a URI

Bijan Parsia: with typed vocabulary = "Data somevaluesfrom" ...

Ian Horrocks: will chair so alan will discuss

Ian Horrocks: yes


Jeremy Carroll: Jim - I muted you because you seemed to be noisy
Jim Hendler: heck, this is the quietest I've been in a long time :-)

alan and ian and peter send time discussing how long to discuss the issue instead of discussing it

ian eventually says "Carry on"

Boris Motik: I believe it is slightly misleading to say this is just Issue 65 - this comes up a lot

Ian Horrocks: boris, can you summarize quickly?

Boris Motik: our general approach to typing things in ontologies is complex w/respect to a number of issues (parsing RDF, punning, etc.)

note (guest): The intention was to discuss the punning issue here, with 65 as support. I would drop 65 as rationale in order to keep on topic [Scribe assist by Alan Ruttenberg]

... these are related, and what we do and don't allow relates to that

... in my emails, I discussed typing triples and the compatibilities


Alan Ruttenberg: I meant primarily to discuss the particular punning issues, so maybe identifying 65 was the issue... I think there are specific problems w/this kind of punning

... seems if we don't use this we need annotation properties back in

Alan Ruttenberg: we didn't take annotations out. We took out annotation properties

Bijan Parsia: one reason we introduced punning was to let more OWL RDF graphs be in DL - punning made some of this easy

Alan Ruttenberg: there is a different mechanisms
Ian Horrocks: no parallel conversations on IRC please!

... this way, more things out in the word would not have to worry about DL vs. Full issues

Boris Motik: Bijan, even without punning, in OWL 1.0 DL we'd have problems releted to parsing RDF and declarations, particularly when you also have imports

... there are ways we can look at this - things are either always one or the other, or things could be typed, or (etc)

... so the problem is we fix the problem (of rejecting graphs) by coming up wth a solution that also rejects graphs (as it were)

... so not sure what to do about it

Zakim: jjc, you wanted to note alternative design on typing triples/declarations


Jeremy Carroll: thanks Bijan, that was helpful - my thought is perhaps we could come up with a different way to handle typing triples


Jeremy Carroll: thinks that the issue of wider RDF compatibility important

Jeremy Carroll: it may be possible to have SHOULD force statements to put typing triples early, to make DL parsing of RDF/XML easier [Scribe assist by Jeremy Carroll]

Alan Ruttenberg: question - to take "second half of this" (not typing issue) - i.e. we would like for the Web architecture for instance names to mean different things in different contexts

... relates this to punning and entailments

Jeremy Carroll: this could be done perhaps with putting the typing triples in a separate file that is imported first [Scribe assist by Jeremy Carroll]


Alan Ruttenberg: so what about cardinality restrictions?

Alan Ruttenberg: cardinality 3 data, cardinality 4 object. When this "means" the same property is the cardinality of that 3, 4 or 7?

Bijan Parsia: Alan, I think that doesn't happen - untyped quantifiers (somevaluesfrom) and a couple of constructs... (and then bijan loses scribe)...

Bijan Parsia: but this is like qualified cardinality

Bijan Parsia: so doable

Bijan Parsia: Alan, I think that doesn't happen - untyped quantifiers (somevaluesfrom) over punned data/object properties can be thought of as typed quantifiers over a single property which can be thought of as a qualified cardinality over a single property where the qualification defaults to owl:Thing for the object oriented restriction and rdfs:Literal for the data oriented one. So it would be =3 P.rdfs:Literal and =4 P.owl:Thing. [Scribe assist by Bijan Parsia]

Bijan Parsia: problem is when we cannot know what they are - it's lack of information, not too much, that causes the problem


summary (guest): Think of typed quantifiers as qualified cardinality constraints [Scribe assist by Alan Ruttenberg]

James Hendler: thinks if we can find way to make this "optional" as opposed to required it would make a lot of OWL/RDF graphs be in OWL DL -- this vocabulary hurts this -- which is why I think the issue is so important

Bijan Parsia: so maybe we could add this as a way to do things, and maybe in some "how to handle RDF graphs" (which may be slightly heuristic) we could suggest how to fix these things

Bijan Parsia: essentially defaults that would help with the coercion

Ian Horrocks: global issue - I think we're going to need an email discussion on this, so maybe someone should lead email discussion on this following on from Boris' email?

Zakim: alanr, you wanted to ask whether we want data/object property punning because we *can*, or whether we think it will be useful and asked for by the community

Alan Ruttenberg: I find this discussion helpful, glad we had it, but -- is this a case of us doing something because technology says we can, or is this something somebody has asked for -- do we have use cases?

Alan Ruttenberg: if this adds complexity, and isn't called for, maybe we should consider whether it is worth the trouble and potential incompatibilities

Boris Motik: I don't know whether people have asked for this explicitly, but it did come up in annotation discussion - punning helps


... when URI is used two ways --- but the question is "does this add complexity to the spec"?

... I mean we do already have this split.

Zakim: jjc, you wanted to comment on e-mail discussion


Jeremy Carroll: moving things to email is not always useful - but chairs focusing some issues to the email does help -

Ian Horrocks: good idea

Ian Horrocks: let's focus on this issue for this week

Peter Patel-Schneider: I agree w/Jeremy, but if we're going to do this sort of thing, we need more lead time -- Ian: I agree

Jeremy Carroll: +1 to peter: the chairs should give a call-to-discuss an issue shortly after telecon
Evan Wallace: +1 on Peter's proposed schedule for email issue focus

perhaps the wiki irc agent could have some way we could note - discussions to be moved to email - so that people could more easily see those in the logs

Ian Horrocks: back to alan

Jeremy Carroll: suggested action ian to modify WG process to include 'issues to discuss this week'

raised issue

issue92

Boris Motik: +1 to take up the issue
Rinke Hoekstra: +1
Peter Patel-Schneider: +1 this issue already has a "resolution-in-waiting"

+1 to take up this issue (and the realated issues Rinke brought up w/other owl ontology declaration vocabularies)

Alan Ruttenberg: issue92 is opened for discussion

issue 29 and issue 74

Alan Ruttenberg: bijan are you there?
Michael Smith: q+ to ask why discussed together
Ian Horrocks: mike smith might be the right person
Zakim: msmith, you wanted to ask why discussed together

Michael Smith: I don't think this issues are the same

Michael Smith: they were split into two issues because they weren't really related.


Michael Smith: on Issue 74, I sent some email

Jeremy Carroll: I care but have forgotten

Michael Smith: not clear if anyone cares

Peter Patel-Schneider: In some sense, I don't care, as this is all RDF-compatability stuff.
Jeremy Carroll: it's important to me that we don't step on XS WG's toes

Ian Horrocks: suggest in line w/previous discussion, it might be case that we suggest this as mailing list point

Ian Horrocks: we could revisit next week


Michael Smith: hendler, there is not an email linked from you at the issue
lightning bolt from above strikes Mike: don't criticize the scribe!!


Bijan Parsia: I thought this included interaction w/XS WG

Peter Patel-Schneider: I sent an email, was told it is under "heated" discussion

Michael Smith: +1 to require someone to advocate for the change

Uli Sattler: I don't want to drop this, because I agree with Bijan it would be nice if we could just use xsd: - so why should we not wait and see?

Jeff Pan: +1 bijan and uli

Jeremy Carroll: procedurally, I suggest we write it into our spec and ask them to review it

Peter Patel-Schneider: +1 it seems to me that the XML Schema specs allow the usage we want
Evan Wallace: +1 to "just making the change"

Bijan Parsia: advocate we close this making the change, and reopen if we discover an issue

Peter Patel-Schneider: the proposal is part of http://lists.w3.org/Archives/Public/public-owl-wg/2008Jan/0017.html

ACTION: jeremy to write a proposal to close issue-74 XSD URIs for facets

ACTION: bijan to draft proposal to close ISSUE-29


Alan Ruttenberg: propose to adjourn

ADJOURNED