See also: IRC log
TomB: propose we accept minutes
of 15th May telecon
... been through minutes and look OK to me
Resolved: accept minutes of 15th May meeting
<TomB> http://www.w3.org/2002/09/wbs/39408/f2f3poll1/
<scribe> ACTION: TomB to start questionnaire on date for f2f -- options are weeks of Nov 4 and Nov 11 [DONE] [recorded in http://www.w3.org/2007/05/22-swd-minutes.html#action01]
<TomB> ACTION: Daniel to send messages to thesaurus list, Semantic Web Interest, ... requesting feedback on the new SKOS use cases document [recorded in http://www.w3.org/2007/05/15-swd-minutes.html#action01] [DONE]
<TomB> ACTION: Jon and Alistair: Move SKOS issues over from Sandbox to Tracker on an ongoing basis [recorded in http://www.w3.org/2007/05/08-swd-minutes.html#action10] [CONTINUES]
<TomB> ACTION: Sean to look at the test environment supporting the SKOS semantics [recorded in http://www.w3.org/2007/05/15-swd-minutes.html#action05] [CONTINUES]
<TomB> ACTION: Alistair will look at raising the examples from the issues to test cases [recorded in http://www.w3.org/2007/05/15-swd-minutes.html#action06] [CONTINUES]
<TomB> ACTION: Guus revise his ISSUE-26 proposal to account for other options [recorded in http://www.w3.org/2007/04/24-swd-minutes.html#action07] [CONTINUES]
TomB: was talking with Guus just
before the meeting and expects to see output from Guus in next few days wrt ISSUE-26 proposal
<TomB> http://www.w3.org/2006/07/SWD/track/issues/31
TomB: Alistair suggests we open basic lexical label sematics at same time as issue 26
<scribe> ACTION: chairs to open issue 31 [recorded in http://www.w3.org/2007/05/22-swd-minutes.html#action07]
Alistair: doesn't this issue need
an owner?
... meant to have its status as open, so that it's on the
agenda
TomB: asks for volunteer to have owner of the issue?
Alistair: I'm willing if no one is desperately keen
TomB: are there any other steps that need to be taken?
Alistair: change the status in
the tracker
... I'll do that
<scribe> ACTION: open issue 31 in tracker for discussion next week [recorded in http://www.w3.org/2007/05/22-swd-minutes.html#action08]
TomB: issue 33--grouping constructs
<aliman> --done
TomB: we had discussion last
week, but now a 3rd proposal from Antoine
... goal of discussion today is to see if we have a rough
concensus so that we can move towards having a proposal
... to vote on in next week's call. Guus felt that we should
try to move this forward
... to some sort of resolution.
... calls on Antoine to put forward proposal 3.
Antoine: my proposal was based on
the diagnosis that the issue comes from
... the fact that two axioms are contradictory
... I suggested to release the axioms that define the domain
and range of the relationships
... and to introduce a new class that would be a superclass of
concept and collection
Bernard: familiar object-oriented pattern from gang of four. (e.g. common pattern in widgetry) called composite pattern ... sub-class which is also a component and a collection of components ...
<aliman> http://www.exciton.cs.rice.edu/JavaResources/DesignPatterns/composite.htm
Alistair: number of issues with
this proposal wrt semantic relationships
... in the current guide there's a rule called "collectable
properties" which was supposed to
... enable your to infer the direct semantic relationships
... other issues are that if you are doing query expansion,
when you have
... a collection that has a node in the expansion graph, so you
can artificially include
<TomB> [collectable properties http://www.w3.org/TR/swbp-skos-core-guide/#seccollectable]
Alistair: extra distance between
the concepts
... also you couldn't write an application that has no
knowledge of collections
... every application would have to have knowledge of
collections
... different to minimal fixed proposal, which means you don't
need any inference rule
... and doesn't need any knowledge of collections
Bernard: what's significance of not knowing about collections?
Alistair: really simple query
expansion doesn't need any collection stuff but guaranteed to
work
... with SKOS data
... For example, if you take a look at the NSDL metadata
repository, there is no
... collection stuff, but it is nevertheless a very useful
platform. If we went for
... Antoine's proposal then an application like that would have
to implement the collection
... stuff--it wouldn't work without it.
TomB: in W3C practice the
proposal that has the least impact on an existing spec is the
one that gets favoured
... In last week's discussion it seemed to me that we were
leaning towards the
... minimal fixed proposal. Does anyone want to add to
that?
Alistair: I agree with what Guus
said. I like the minimal fixed proposal because it does
... the smallest fix and takes an incremental approach. We can
take the approach
... and then explore issues with it. If there's a problem with
it, then we can take a look
... at other proposals (such as the relaxing domain/range or
the Bundle)
<TomB> +1 Alistair and Guus favoring Minimal Fix
Antoine: my worry is that in most
of the situations where we have seen collections
... the collections are included in the hierarchy.
Alistair: there's a difference
between a conceptual hierarchy and a systematic display
... the latter is a visual arrangement of terms, not really
part of what standards people
... would call a conceptual hierarchy.
... node labels are
introduced as a display convenience.
... In the minimal fixed proposal, let's say you're generating
a systematic display
... such as kinds of milk. In a minimal fix, you
declare a concept of milk, then goat's
... milk, cow's milk and
express narrower terms between them. Then itroducce a
collection
... say "milk by source animal". An application that gets that
data can ignore the
... collection data and just display the relationships between
kinds of milk.
... whereas an application that understands collections can ask
if (say) cow's milk
... a member of any collections?
... and then recurse upwards, to see if the "milk by source
animal" collection is also in a collection?
Daniel: I'm confused by what's asserted by collection membership and the siblings of a taxonomy
Alistair: collections are not
just siblings. You could have "milk by source animal", and
"milk by fat content"
... if you're actually thinking about classes, then "milk by
source animal" is not really a class.
... in the thesaurus community, "milk by source animal" is not
a concept, it's a node label
Daniel: but the ontology community wouldn't use this approach.
Alistair: the dependency of how
we model this and how we migrate from thesaurus to
ontologies
... the primary concern for SKOS is to make sure that it is
sufficient to represent thesauri and classification
schemes
... so that existing thesauri and classification schemes can
continue to have the features
... they're used to.
Daniel: we also have to do the same from an ontology-oriented perspective
Alistair: I am aware of Antoine's
point about the difficulty of implementing the algorithm
... I explain the algorithm to follow the narrower links, then
recursive collection-membership relationships
... there's another standardisation group (in the UK) that's
trying to write down a data model (in UML) for thesauri
... this group is having a big discussion about you handle
collections (aka Arrays)
Daniel: anticipate resolution from this group?
Alistair: they're influenced by
how SKOS does it.
... they're not hardcore data modellers.
... we'd have to present it in UML for them to understand.
Daniel: that might be useful for us to present it anyhow.
Alistair: because the UK group is doing it as a closed group it's not that public
Daniel: so it presents some interesting questions about how to map from SKOS to their data model
Alistair: I made a suggestion. Here's the link
<aliman> http://isegserv.itd.rl.ac.uk/blogs/alistair/archives/41 -> Initial suggestion of data model for thesauri for BS8723
Alistair: the idea that I had was
that we go with the minimal fixed proposal for now
... which forces me to expand on the algorithm
<TomB> +1 to go with Minimal Fix
Alistair: and if it's too complicated then we can change our minds. It's not the final word.
TomB: does anyone disagree with minimal fix?
Daniel: concern that we also need to address from ontology world
Alistair: need use cases to understand this better
TomB: sounds like we're moving to
minimal fix with reservation that we need to understand
... how ontologies map to thesauri.
... so could we put minimal fix on agenda for next week.
Alistair: yes, then we could resolve or not next week
<scribe> ACTION: Alistair to propose minimal fix for resolution of issue 33 [recorded in http://www.w3.org/2007/05/22-swd-minutes.html#action09]
TomB: to be talked about next
week
... propose SKOS issue resolution process to next week
<TomB> http://www.w3.org/2006/07/SWD/wiki/IssuesProcess
TomB: also want to make sure that
the document on the issues process actually corrresponds to the
way we are doing things
... find it slightly confusing about how to see issues that are
closed
<Antoine> ... beware of not losing it!
<scribe> ACTION: Ralph propose resolution to Recipe issue 1.3 [recorded in http://www.w3.org/2007/04/24-swd-minutes.html#action09] [CONTINUES]
<scribe> ACTION: Elisa to provide outline of work to be done by Apr 17 [recorded in http://www.w3.org/2007/04/24-swd-minutes.html#action11] [CONTINUES]
Meeting adjourned