Warning:
This wiki has been archived and is now read-only.
Teleconference.2008.03.26/Minutes
These minutes have been approved by the Working Group and are now protected from editing. (See IRC log of approval discussion.)
Contents
- Present
- Ian Horrocks, Markus Krötzsch, Ivan Herman, Diego Calvanese, Bernardo Cuenca Grau, Sandro Hawke, Boris Motik, Uli Sattler, Doug Lenat, Evan Wallace, Ratnesh Sahay, Michael Smith, Jeremy Carroll, Zhe Wu, Alan Ruttenberg, Peter Patel-Schneider, Achille Fokoue, Jie Bao, Rinke Hoekstra
- Regrets
- Michael Schneider, Elisa Kendall, Carsten Lutz, Bijan Parsia
- Chair
- Ian Horrocks
- Scribe
- Markus Krötzsch
Admin
Agenda amendments: amendment on RIF OWL compatibility issue
(see below under #Additional other business)
Telcos will start at 19:00 CET again starting next telco (after F2F)
Accept previous minutes?
Previous minutes accepted
F2F registration fees must be paid, see http://www.webont.org/owled/2008dc/f2fregistration.html
Cheques are also accepted on-site
Action items
Action 113 completed
Action 105 completed
Action 72 due tomorrow
Action 76 delayed, to be done soon (before F2F2)
Action 86 delayed
Action 90 delayed, discussion scheduled for F2F2
Action 100 remains open (James Hendler not on call)
Action 101 delayed, to be done for F2F2
Action 102 remains open (James Hendler not on call)
Action 104 completed
Action 109 completed
Action 93 completed
Action 114 completed
Issues
Issue 86
Issue 86 http://lists.w3.org/Archives/Public/public-owl-wg/2008Mar/0189.html
Alan Ruttenberg: new updated proposal is indeed editorial, minor correction only
Michael Smith: inverse properties could also be handled in RDF mappnig by swapping object and subject and adding an annotation for enabling round-tripping
Alan Ruttenberg: annotations should not have any logical meaning
Ian Horrocks: the annotation was just for round-tripping, no logical impact
Alan Ruttenberg: I doubt that exchanging subject and object always works
Ian Horrocks: yes, it is obvious only for facts, but not for restrictions
Michael Smith: I think facts are the only case where this applies, there is no problem with inverses in restrictions
Jeremy Carroll: I agree with msmith, it is only a problem in facts
Alan Ruttenberg: Jeremy referred to the problem of multiple (nested) inverses; I also think that this should be disallowed
Ian Horrocks: there seem to be two competing proposals now
Alan Ruttenberg: if it is just about facts, then Michael Smith's proposal might work as well
Ian Horrocks: does anybody doubt that Michael Smith's proposal works?
Jeremy Carroll: my problem with it is that the proposal requires annotating the triple for round-tripping
Ian Horrocks: Who thinks that Alan Ruttenberg's proposal is the way to go
PROPOSED: Alan Ruttenberg's proposal
PROPOSED: who thinks that Michael Smith's proposal is the way to go?
Michael Smith: could we have a strawpoll on whether we wish to support ObjectProperty assertions where properties are inverses at all
Michael Smith: it might be that we do not really need that
Michael Smith: I suggested to disallow this in an email
Boris Motik: support for inverse properties without explicitly naming them was requested from the DIG working group, and is used in many reasoners
Ian Horrocks: further discussion needed, to be taken to email
Jeremy Carroll: can we have a strawpoll on disallowing inverses as proposed by Michael Smith?
PROPOSED: disallow anonymous inverse properties in facts
Issue 12
Issue 12 http://www.w3.org/mid/20080318.213218.171431830.pfps@research.bell-labs.com
Alan Ruttenberg: I suggested a modification for improving efficiency (see http://lists.w3.org/Archives/Public/public-owl-wg/2008Mar/0254.html)
Peter Patel-Schneider: Alan's proposal might be problematic, since it adds totally new different individuals
Alan Ruttenberg: But it is only an existential (a bnode)
Peter Patel-Schneider: Yes, but it still might increase the size of the domain.
Peter Patel-Schneider: this creates a burdon on round-tripping
Peter Patel-Schneider: but the creation of new individuals as such is the main problem
Alan Ruttenberg: for equivalent class and equivalent properties it should not be such a problem
Ian Horrocks: not sure without having a closer look
Ian Horrocks: it seems we cannot resolve the issue with that solution without further investigation
Alan Ruttenberg: my proposal for possible optimisation might be a separate issue
Jeremy Carroll: The issue also touches other issues related to reification, and I would currently abstain on this issue until the relations are clearer
PROPOSED: should this issue by closed according to pfps's proposal http://lists.w3.org/Archives/Public/public-owl-wg/2008Mar/0205.html?
PROPOSED: close Issue 12 as in pfps's email provided that new information does not come in and noting that this does not close the general reification issue
RESOLVED: close Issue 12 as in pfps's email http://lists.w3.org/Archives/Public/public-owl-wg/2008Mar/0205.html provided that new information does not come in and noting that this does not close the general reification issue
Ian Horrocks: Time for issues is up, proceed with reviews of candidate working drafts first
ACTION: bmotik2 to Update the RDF mapping with the accepted resolution of Issue 12 as per Peter's suggestion
Reviews of candidate FPWDs
Fragments document
Achille Fokoue: I corrected minor problems, and raised issues where needed
Michael Smith: mostly editorial comments, not put into wiki page yet
Alan Ruttenberg: uncontroversial editorial changes can be made directly in the document
Michael Smith: OK, I will shortly incorporate my comments then
Sandro Hawke: we might want to change the document title (wiki page name currently is Fragments Proposal)
Alan Ruttenberg: "Fragments of OWL" seems to be a good name
Ian Horrocks: I would just say "OWL Fragments"
Ian Horrocks: Actually just "Fragments" would be good
Jeremy Carroll: the abstract of the fragments doc is insufficient
Jeremy Carroll: also, the term "fragment" may have a negative connotation
Alan Ruttenberg: "profiles"?
Ian Horrocks: this discussion should be continued offline
Achille Fokoue: I would like an additional review of the changes I made recently
Diego Calvanese: regarding the DL Lite fragment, I think major changes are still required in the document
Diego Calvanese: for example function properties should be there
Diego Calvanese: several other additions could be made
Peter Patel-Schneider: we can note in the current version that the presented DL Lite fragment is a conservative fragment that is likely to be extended later on
ACTION: bmotik to add a note to the fragments document regarding possible later extension of DL Lite
XML Serialisation
Achille Fokoue: my comments were already addressed by pfps
Achille Fokoue: the document is in a good shape overall
Sandro Hawke: I noticed inconsistent uses of namespaces
Sandro Hawke: maybe we should add a note that the namespaces might be changed later
Ian Horrocks: it would be good to have a sensible solution eaerly on, so as to prevent people using placeholders that are changed later
Alan Ruttenberg: minor comment to change the namespace prefix in the schema to the one used in the text
Alan Ruttenberg: there are still inconsitent namespace prefixes in the document
Alan Ruttenberg: my second comment is about "abstract" and "concrete" classes; the use of those terms is not fully clear yet
Ian Horrocks: Who has implemented OWL 1.1 software? Which namespaces would users prefer?
Boris Motik: I do not think users would care much; I have not received any user comments on that
Alan Ruttenberg: We still should not change the namespace if there are already OWL 1.1 ontologies
Boris Motik: I think there are not so many OWL 1.1 ontologies that would be affected there
Michael Smith: I posted the namespaces used in the OWL API into IRC
Michael Smith: I also think that there is no problem in changing the namespace now
Sandro Hawke: I would prefer a non-dated namespace now
Ian Horrocks: yes, I also would like to not have dates in the namespace
Sandro Hawke: yes, the dates were only used to generate unique names, the are not really needed
PROPOSED: use http://www.w3.org/ns/owl11-xml as a namespace
PROPOSED: use http://www.w3.org/ns/owl11-xml as the XML namespace name for OWL 1.1 XML serialization
RESOLVED: use http://www.w3.org/ns/owl11-xml as the XML namespace name for OWL 1.1 XML serialization
Sandro Hawke: A related question is where to publish the schema, but I need to do some more research on this
Zhe Wu: Does this namespace decision imply that the new standard is called OWL 1.1?
Ian Horrocks: Should we rather change the namespace to drop the 11?
Jeremy Carroll: We can still change our minds on the namespace later on, since this is just a working draft
Ian Horrocks: Indeed, we can go with the current resolution for the early working draft now
Primer
Jie Bao: There is still some work to do. The main question is who will read the Primer.
Jie Bao: I think the document needs to be restructured to become accessible for users of different levels of expertise.
Jie Bao: Advanced topics like Open World Assumption should go to later sections.
Ian Horrocks: Jie, could you and Deborah cooperate with the authors to fix the perceived problems with the Primer until F2F2?
Jie Bao: I think Section 3 and 4 are mostly our problem, but fixing that might not require much time
Alan Ruttenberg: I object that OWA is considered an advanced topic. It is in fact a major point in using OWL
Alan Ruttenberg: I therefore would object to moivng that to the appendix
Doug Lenat: The Primer is far from being ready to be published.
Doug Lenat: whole sections are mostly blank
Doug Lenat: e.g., the advanced features section is not containing advanced features
Doug Lenat: more people from this group should review the Primer, since it is important for the perception of OWL
Doug Lenat: the current content of the Primier is good, but more is needed
Jie Bao: I agree with DougL
Jie Bao: I think that an incremental presentation would be better
Peter Patel-Schneider: we have only two days left to make changes, which seems to preclude any substantial changes before F2F2
Peter Patel-Schneider: the missing sections are "Using OWL" and "Using OWL Tools"
Peter Patel-Schneider: not sure whether "Using OWL" is needed
Peter Patel-Schneider: I do not agree, though, that the Primer is not publishable as is
Alan Ruttenberg: there is a UFDTF meeting this monday where we might discuss this issue
Alan Ruttenberg: would it be harmful to not publish now but in two weeks?
Ian Horrocks: we can consider that at the F2F
Sandro Hawke: Publish early, publish often!
Sandro Hawke: we can publish an incomplete document as a working draft
Jeremy Carroll: It would be useful to add a one-sentence summary of the review comments, to indicate open points that are still worked on
Jeremy Carroll: in general we can have comment inline to indicate work in progress
Peter Patel-Schneider: It will be hard to point to the distributed reviews from the Primer
Sandro Hawke: we should not point to the reviews, but reach a consensus on what needs to be done, and state that in the document
Additional other business
Peter Patel-Schneider: Agenda amendment: RIF statement on OWL compatibility http://www.w3.org/2005/rules/wiki/SWC#OWL_Compatibility
Peter Patel-Schneider: Possibly interesting to members of this group, please read if you are interested in the topic.