See also: IRC log, previous 2008-12-02
RESOLUTION: to accept minutes of the last telecon: http://www.w3.org/2008/12/02-swd-minutes.html
tomb: next telecon 16 dec
ralph: regrets for 16 dec
ACTION: Ben review RDFa Use Cases and propose transition to Group Note [recorded in http://www.w3.org/2008/09/30-swd-minutes.html#action02] [CONTINUES]
ACTION: Ralph to review the revised Recipes draft [recorded in http://www.w3.org/2008/12/02-swd-minutes.html#action15] [CONTINUES]
ACTION: Ralph/Diego to work on Wordnet implementation [of Recipes implementations] [recorded in http://www.w3.org/2008/01/22-swd-minutes.html#action20] [CONTINUES]
ACTION: Ralph post his comments on the editor's draft of the metadata note [recorded in http://www.w3.org/2008/11/25-swd-minutes.html#action03] [CONTINUES]
ACTION: Guus to look at OWL documents for review [recorded in http://www.w3.org/2008/10/21-swd-minutes.html#action10] [CONTINUES]
ACTION: Sean to add rdf:type and rdf:Property assertions to the skos schema [recorded in http://www.w3.org/2008/11/18-swd-minutes.html#action07] [DONE]
seanb: in latest version of schema, looking for link
<seanb> Schema is at: -> http://www.w3.org/2006/07/SWD/SKOS/reference/20081001/skos.rdf
ACTION: Guus discuss response to issue 157 with Sean [recorded in http://www.w3.org/2008/11/25-swd-minutes.html#action12] [DONE]
seanb: guus posted the draft for 157
-> ISSUE-157 Draft response [was Re: OWL WG LC comment for SKOS reference document] [Guus 2008-12-04]
ACTION: Antoine to write something in Primer wrt. ISSUE 160 [recorded in http://www.w3.org/2008/12/02-swd-minutes.html#action14] [CONTINUES]
ACTION: Antoine propose 1 or 2 SPARQL examples showing named graph usage [recorded in http://www.w3.org/2008/11/25-swd-minutes.html#action14] [CONTINUES]
Alistair: Guus wrote a draft
proposing to rearrange the examples
... however Peter wasn't satisfied
... Guus noted some things that OWL DL wouldn't support
... Antoine noted issues with the rearrangement as this would introduce
stylistic differences between the Reference and the Primer
... Guus proposed that the OWL Full examples be collected under a caveat
... we need to find a way to resolve issue 157
-> issue 157; Last Call Comment: SKOS and OWL 2 analysis
Antoine: was Peter's comment about SKOS itself?
Sean: Peter's worries are not about the data
model itself but about the presentation of the examples
... Peter was concerned that if we used OWL Full patterns in the examples
this would encourage [more] people to use OWL Full
seanb: can't stop people doing anything, but peter keen we don't use rdf:value in example
antoine: people should use patterns about to be
hidden, because most simple and common ones
... uncomfortable with encouragement
Alistair: if this is a problem we need an alternative
antoine: clear warning about OWL DL
... reference has normative value, so people won't come in to primer first
... so maybe sub-optimal
seanb: we have these three examples, a literal, a blank node with rdf:value, and a URI -- you want to see all three with equal value?
antoine: literal one is most common,
simplest
... wouldn't use rdf:value, but means don't need URIs for notes, so could be
beneficial
seanb: came late, but alistair had alternative proposal to consider them as annotation properties, so then problem using them with literals and objects goes away
antoine: at first glance ok with that, already
something we have for labeling properties, so re-using this solution is
proper way to do it
... might still be problems with use of annotation properties
seanb: rdf:value is still an issue, but would solve some of the problems
antoine: anything about rdf:value in OWL 2?
seanb: not mentioned, has no semantics
antoine: what is problem with it?
tomb: where is use of rdf:value documented? mentioned in primer, in reference too?
aliman: just in an example in reference
antoine: in rdf semantics, listed with containers, collections etc. just gives ideas on use, no formal constraint
tomb: more like a usage convention
antoine: small paragraph of 7 lines, meaning can vary
<Antoine> http://www.w3.org/TR/rdf-mt/#ReifAndCont
Alistair: to make Peter happy we should look at
how rdf:value is being treated in OWL2
... suppose we made all of the properties be owl:AnnotationProperty ?
... but leave the Primer as it is
Antoine, Sean: I'd be happy
Sean: Guus might not be satisfied
seanb: i would support that
... discussing with simon jupp, why aren't documentation props annotation
props? tells a consistent story
... veering towards substantial change, I'm not sure it is.
antoine: given that skos:note is for annotations, easy to defend.
seanb: object properties (semantic relations), then labeling and documentation properties (annotation properties)
antoine: difficult to defend doc props not as annotation props
Alistair: someone might want to add an extension to SKOS that placed cardinality restrictions on some properties
Antoine: complicated because there are many annotation properties in the SKOS world
aliman: i can live with them as annotation props
seanb: ralph, do you see this as a substantive change? more editorial?
ralph: not sure ... from two viewpoints: Would any developer of a skos tool have to change their code to conform to altered version of reference? Or, likely that any existing LC comments would disagree with this proposed change? I.e. would anyone feel this is unacceptable change?
seanb: difficult to say
ralph: prefer to err on side of caution, if not sure, consider it substantive
seanb: none of comments mention documentation properties; but maybe not mentioned because totally happy;
ralph: suspect relatively few looking at OWL DL vs OWL Full differences in detail
seanb: I would be happy with this change
tomb: what are implications if do consider this substantive?
ralph: we would need to do another last call
... formally, we don't have huge variety of choices; more than an editorial
change, peter feels it's important.
... but could say, we don't consider this change invalidates any other
reviews, and don't consider that skos implementers have to make changes.
seanb: if happy to change to labeling props, then I'm surprised if we're unhappy to similar change to doc props.
tomb: could we assert this is not substantive change, not hide problem but ... to go to another last call over this seems to far on the side of caution
ralph: i agree another last call is overkill. i wouldn't want to wordsmith too finely to avoid substantive change question, just say we have made the following changes.
seanb: we're still giving people option to comment, so this isn't final.
ralph: purpose of this part of review process is to make sure, if somebody did look very carefully at doc props and decide exactly right, they wouldn't comment. need to make sure any reviewer who could care is on notice that changes have occurred.
tomb: need to take an action?
aliman: need to redraft response to peter on 157
ACTION: seanb to redraft response to peter on ISSUE-157, where skos doc props are annotation props, and rdf:value example is dropped from skos reference [recorded in http://www.w3.org/2008/12/09-swd-minutes.html#action10]
tomb: any objections to this?
[none]
<Ralph> Sean, please make sure these sorts of things are explicit in the Changes section
aliman: need to approve sean's drafts on namespace issues
seanb: approved and sent already
tomb: so look forward to next week, we should
vote next week on a resolution to request candidate rec in first week of jan,
then on next day ralph will ask for extension of charter to end of april
... Need to propose to go to candidate rec on the list, so can vote at next
week's call.
ACTION: alistair send email with editors' draft proposed for CR before next telcon [recorded in http://www.w3.org/2008/12/09-swd-minutes.html#action11]
tomb: can say time has run out if no response by next week
Alistair: on 157, can Sean send his revised response based on today's discussion directly to Peter?
Tom: yes, that would save time
PROPOSED: to respond to ISSUE-157 commenter, proposing skos documentation props are made OWL annotation properties and rdf:value example is removed from skos reference
tomb: any discussion?
RESOLUTION: to respond to ISSUE-157 commenter, proposing skos documentation props are made OWL annotation properties and rdf:value example is removed from skos reference
seanb: i'll check with guus, what if he doesn't like it?
tomb: then we have to go to list somehow
Alistair: we'll make a new editors' draft
incorporating all the changes we proposed to commentors
... will be contingent on Peter and Guus being willing to live with the
resolution to issue 157
tomb: ask if, implementations, ed and antoine
what's the status of the LoC subject headings?
... any issues to resolve there?
edsu: we're both working on separate things,
but kind of the same. what LC is doing, what their not doing. lcsh.info was
done as an experiment by me, to get feedback on whether correct and/or
useful.
... people here see it is useful, good feedback, most concerns come back to
having it live at real domain in LC rather than my domain, no concerns about
skos implementation details.
... antoine and I discussed things specific to LCSH, meaning to write email
on the topic, basically that LCSH has sets of concepts that are more
specialised than skos:Concept, e.g. topical concepts, geographic concepts,
form/genre concepts, things specific to LC. antoine and I discussing, i've
been inclined to specialise SKOS, whereas antoine inclined to use
skos:inScheme to identify separate...
... groupings.
tomb: also wanted to ask about other vocabs in clay's department. i was interested to see if marc relator terms are now declared as rdf properties? have been for past 2-3 years, but proposal to declare smae URIs as both rdf properties and skos concepts. antoine you discussed that with clay?
antoine: not that. specific concept scheme of relations?
tomb: i went into skos reference, it's not a contradiction to say that a skos concept is also an rdf property... but it makes my head spin :)
edsu: i talked to rebecca about it, better to leave definitions as they are. if want to declare elsewhere then fine, but leave existing ones as are and keep uris stable
tomb: proposal was to additoinally say they are
also skos concepts
... nothing formally says you can't do that, but i'm wondering what it
means
edsu: potentially confusing
tomb: can have sub-property relation between
uri as property, and a broader relation between same uris
... meeting in may, rdf schema doesn't distinguish preferred labels, so
feature of skos attractive, but labeling properties are not committed to skos
concepts, so could use preflabel with properties
antoine: but problem with semantic relations, restricted to skos concepts.
tomb: i sent a not to clay, wanted to follow
up
... if using marc relator terms as implementation, need to resolve that.
... anything else to discuss?
seanb: on topic of implementations, some work here (manchester) simon jupp building api infrastructure for skos and an editor as well, so can use as implementation experience
edsu: i've been emailing with simon, he's been using lcsh.info as testbed
seanb: simon makes use of schema to drive the application, so if we change schema he can cope with that
ralph: two action items not carried forward...
ACTION: [PENDING] Ralph to report on use of RDFa metadata in Recommendations. [recorded in http://www.w3.org/2008/11/18-swd-minutes.html#action02]
ACTION: Guus and Jeremy to give concrete implementation examples of the use of rdfs:label w/ SKOS [recorded in http://www.w3.org/2008/10/07-swd-minutes.html#action10] [DROPPED]
tomb: meeting adjourned
Tom:we dropped that action re: rdfs:label last week