See also: IRC log
<RalphS> previous 2007-07-24
<seanb> apologies -- will be a couple of minutes late dialing in....
<RalphS> Scribe: Elisa
<RalphS> scribenick: elisa
Proposed: Minutes of two weeks ago accepted http://www.w3.org/2007/07/24-swd-minutes.html
Proposed next telecon will be in two weeks, 21 August
<RalphS> F2F poll results; 11 answers thus far
W.r.t. on F2F poll, Guus proposes that there may be another way to go on this
there is still another week for responses
there are issues with several of the dates -- conflicts with other meetings, holidays
Guus is still proposing to have the F2F on October 8-9 in Amsterdam, based on results of the poll at present
The meeting topic will be primarily SKOS, which means a full agenda already, even without other topics
Guus proposes that F2F should be 8-9 October in Amsterdam
Ralph second
Resolved to meet 8-9 October in Amsterdam
<RalphS> Simone, you didn't answer http://www.w3.org/2002/09/wbs/39408/skos-ftf-amsterdam/ yet -- can you travel to Amsterdam on 8-9 October?
Guus will plan logistics; need to set up web page
<RalphS> (same question for Sean)
<scribe> ACTION: Antoine & Guus will take on setting up logistics [recorded in http://www.w3.org/2007/08/07-swd-minutes.html#action01]
<Simone> @Ralph: I think has some logistical problem at office, but my boss has holidays and I'm waiting for he
<scribe> ACTION: Guus to email a proposal to the list about the issue of containment [recorded in http://www.w3.org/2007/07/24-swd-minutes.html#action01]
<scribe> DONE: http://lists.w3.org/Archives/Public/public-swd-wg/2007Jul/0164.html
<RalphS> ok, Simone; thanks.
Main point is that containment should not be handled at the level of SKOS constructs --
it should be handled outside of SKOS, although it is important
may need to be resolved at the F2F
Ralph: alot of the SKOS discussions have been what we generally need for the language (SKOS) but depends on other context
context including RDF, etc.
Ralph had kept generally quiet on some of these issues, deferring to others on what should be in the
thesaurus related vocabulary
Guus: in particular issues such as containment, provenance
Question regarding language, containment issues (Antoine?)
Was wondering if these issues dealt only with statements, does not know if the element of the statement is also contained in the concept scheme based on the definition of containment
Guus; there is a more general issues, possibly in 2003 -- discussion on social implications of RDF statements
what can other people say about your resources ... paragraph was deleted from the original RDF
specifications
Guus is concerned that we are going down the same road here
Antoine's point would be to keep to the simple solution, SKOS would not try to deal with containment of statements,
but we should keep the notion of containment of concepts
Keeping inScheme -- which should address this
Guus: thus you could do about half -- concepts but not statements
Antoine: the obvious mechanisms don't handle this relationship between concepts and concept schemes
Guus: wondering why we need InScheme, given that ontology people have not been talking about this ...
<Zakim> aliman, you wanted to say, I guess the question to ask is: what breaks if we remove skos:inScheme?
Alistair: question is why did we decide we needed it in the first place, and what breaks if we remove it
It would be good to have a short list of the use cases that helped make that decision
I have two thesauri, and merge the two graphs, but then want to display the contents from only one
Alistair: we used InScheme to do this before we had named graphs --
<RalphS> +1 to using the use cases to decide if a feature is important
Guus: if you talk about a concept in a scheme, we look at the namespace URI, and don't trust that
everything in the namespace belongs to the scheme
Sean: if they don't make the assumptoin about the namespace and graph overlapping ... answered by the
fact that tools / developers make the assumption that statements are about elements that happen to occur in the same graph
<RalphS> why aren't people using rdfs:isDefinedBy ?
Ralph: why aren't people using isDefinedBy
Guus:
we could actually
say in the document that this is the preferred practice
Jon:
the discussion is related to issue 36, and 36 and 35 are both confused ... issue 36 is that the
broader /narrower relationship between concepts is actually an issue with the concept scheme
question is does the concept exist without the container
how do you express properties of the container under the current circumstances
is broader / narrower actually an issues wrt the concept scheme
Guus: yes - you need to use reification to use isDefinedBy
Jon: this goes to another conversation -- about whether broader / narrower is about relations within or across concept schemes
is the concept scheme a required property of a concept -- membership in a concept scheme
are concepts aggregated within a concept scheme
<Zakim> aliman, you wanted to say the use cases for skos:inScheme were quite simple originally, nothing to do with authenticity or provenance
these are questions we need to answer before addressing the issues around inScheme
Alistair: it was never a requirements for inScheme -- original issues were much more naive, did not
address provenance, etc.
focus was more on how to pull schemes apart once merged
there may be some reasonable work-arounds based on the original use cases
Jon: what kind of containment are we talking about -- containment or aggregation
Alistair: this isn't a question we've asked in any depth to date
<RalphS> Jon: if you delete a concept scheme, do the concepts "in" it go away?
from a traditional thesaurus point of view, the thesaurus is composed of the concepts, thus if you
destroy the thesaurus you destroy the concepts, but from a semantic web point of view, you might
want to do this differently
Jon: what we probably want is aggregation rather than containment
Ralph: does aggregation have any specific interpretation in this context
Jon: thinking about OO development
<Zakim> RalphS, you wanted to wonder if relationships between concepts are non monotonic
Ralph: does this map onto Sean's earlier comment
Jon: this is a problem for someone exposing data on the web --
Ralph: is there a community of people who would want to destroy data using this mechanism ...
I'm a little worried that people might be viewing the semantics of inScheme in a way that might be nonmonotonic
if you have two concepts a and b and a concept scheme that might define relationships between them, and another concept scheme that defines different relations between them
are they the same concepts
if the relationships that scheme 1 defines when you refer to scheme 2 should still be true ...
Guus: we have the notion of owl:deprecated
<Zakim> seanb, you wanted to ask about whether we should be considering some abstract model
Ralph: ... two concept schemes would not want to deprecate each other
Sean: wondering if it would be easier to do this without the RDF ... thinks this muddies the water a bit,
worrying about named graphs, triples, etc.
instead we should think about what aggregation might mean, then worry about translation to underlying triples that might represent it; thinks about this similarly to how he thinks about the owl abstract syntax
Ralph: wondering if this is something we could describe in terms of the owl abstract syntax --
<Zakim> aliman, you wanted to advocate opaque URIs and to respond to sean to ask: fair enough, but if we model above RDF, then how do we model? What language do we use?
Sean: we need to be clear about how the model fits together without worrying about the triples
Alistair: is sympathetic, but we need to model it somehow, and if we aren't using RDF, how would we model it
Sean: the activity we've had with respect to owl 1.1, we've used abstract syntax to define the model ...
<RalphS> Elisa: there may be something in UML that helps, but answering the questions in English is a first step
<RalphS> Elisa: maybe it's not a bad idea to draw a picture
<RalphS> ... try to tease out what we consider "inside" SKOS and what we consider at this moment to not be inside SKOS
<RalphS> ... having spent a lot of time on provenance, I have mixed feelings about what kinds of constructs should be used
<RalphS> ... Deb McGuinness has developed PML (Proof Markup Language) to describe some of these things
<RalphS> ... I use PML when something is not precise enough
<RalphS> ... 'draw a picture' -- i.e. create a metamodel
Sean: not suggesting that we have something formal, but drawing the picture would be very useful (i.e., creating a metamodel)
Alistair: we should start with some simple use cases and then go from there
edsu: focus on the use first then the solution makes sense
Alistair: taking an operational approach first was the only way to do this originally
Jon: are the use cases something we could effectively add
Alistair: I'm thinking of little stories, more specific to the use cases we currently have, that would
describe the behavior
Jon: we're lacking the simple user stories
Antoine: I agree, and some of the things we have in the use case document could provide the context for that
we could describe the strategies that need to be used to solve them
Alistair: Jon's use case is really rich with little stories and could be used for that
Antoine: his organizations use case could also be used as a basis for this
Alistair: should we set up a wiki page to describe user stories as a starting poitn
Guus: I think the notion of having a collection place for the different options here would be good
Antoine: if the time schedule is within weeks, he can set up a wiki as a starting point
Guus: we have two concepts, a scheme that describes a relationship between them, a second concept scheme ...
Ralph: how do we actually think applications are behaving
Guus: we will see how far we can get with this
Continue Guus' two actions, Alistair's action
<aliman> ACTION: Alistair to state the difference between the two flavours of the SimpleExtension proposal for issue 26 [recorded in http://www.w3.org/2007/07/17-swd-minutes.html#action05] [CONTINUES]
Ralph: spent some time cleaning up documents, etc.
at telecon last week Shane ? took an action to start fixing up the editor's draft for the syntax document
Guus: new working draft soon?
Ralph: yes, though can't provide an exact schedule in the moment -- they are behind where they wanted to be, but converging on the technical issues
there are details on how to spell certain things, for example, but close to consensus on most things
Guus: so it would be good to have a new working draft soon. how will this work, as two working groups have to approve this
Ralph: doesn't see difficulties
actions continued
Guus: would it be good to schedule some time for Recipes at the F2F?
Jon: yes, that would be useful, to get consensus on whether certain issues should be addressed, also useful to have a hard deadline for a draft we can discuss
<berrueta> +1 to put the recipes into the agenda at the F2F
Guus: we will reserve some time for the recipes ..
<RalphS> Elisa: I have made some progress and would like time on the next telecon to talk about this
<RalphS> Jon: regrets for 21 & 28 August telecons
<RalphS> ACTION: [DONE] Diego to repeat test without q values re: ISSUE 58 [recorded in http://www.w3.org/2007/07/24-swd-minutes.html#action02]
<RalphS> ACTION: [PENDING] Guus to move ISSUE-26 forward [recorded in http://www.w3.org/2007/07/03-swd-minutes.html#action04].
<RalphS> ACTION: [PENDING] Guus to post user experience reports for ISSUE-26 [recorded in http://www.w3.org/2007/07/03-swd-minutes.html#action05]
<RalphS> ACTION: [PENDING] Alistair to state the difference between the two flavours of the SimpleExtension proposal for issue 26 [recorded in http://www.w3.org/2007/07/17-swd-minutes.html#action05]
<RalphS> ACTION: [PENDING] Ralph propose resolution to Recipe issue 1.3 [recorded in http://www.w3.org/2007/04/24-swd-minutes.html#action09]
<RalphS> from earlier comments, putting them in the record:
<RalphS> Antoine: RDF containment would enable to describe the provenance of statements...
<RalphS> ... but not the provenance of concepts in concept scheme ...
This is scribe.perl Revision: 1.128 of Date: 2007/02/23 21:38:13 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/Jon/edsu/ Succeeded: s/amplifications are helping/applications are behaving/ Found embedded ScribeOptions: -implicitContinuations *** RESTARTING DUE TO EMBEDDED OPTIONS *** Found Scribe: Elisa Found ScribeNick: elisa Default Present: Ralph, edsu, Guus, Antoine_Isaac, Elisa_Kendall, JonP, Sean, Simone, Alistair Present: Ralph edsu Guus Antoine_Isaac Elisa_Kendall JonP Sean Simone Alistair Regrets: Tom Daniel Diego Agenda: http://lists.w3.org/Archives/Public/public-swd-wg/2007Aug/0009.html Got date from IRC log name: 7 Aug 2007 Guessing minutes URL: http://www.w3.org/2007/08/07-swd-minutes.html People with action items: alistair antoine guus propose ralph resolution[End of scribe.perl diagnostic output]