See also: IRC log, previous 2008-01-08
<scribe> agenda: http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0071.html
PROPOSED to accept minutes of the Jan 8 telecon: http://www.w3.org/2008/01/08-swd-minutes.html
RESOLUTION: accepted minutes
next telecon: 22 January 2008
<scribe> ACTION: Quentin to review Editor's draft of SKOS Reference [recorded in http://www.w3.org/2007/11/20-swd-minutes.html#action02] [DONE]
<scribe> ACTION: Vit to review Editor's draft of SKOS Reference [recorded in http://www.w3.org/2007/11/20-swd-minutes.html#action03] [DONE]
-> http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0004.html Quentin's review of SKOS reference
-> http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0044.html Vit's review of SKOS reference
aliman: looked at those reviews
... neither quentin's nor vit's comments are about technical issues
... mostly about wording
Guus: propose to write a new version and ask
the reviewers
... send the new draft and a mail explaining how the comments by the
reviewer's were addressed
aliman: the schedule is quite aggressive
guus: you can put TODO's in the document
tom: there was some discussion on the naming of
some properties and a class which contain the word "relation"
... it might be clearer if we re-order the words
... "labelRelated" -->"relatedLabel"
<Antoine> +1
Guus: please take this into consideration for the draft
aliman: I chose the previous name to make a distinction with all of the "*Label" relationships
Guus: we should add a note to explain it
<vit> http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0004.html
<Quentin> http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0004.html
aliman: quentin pointed an inconsistency
between the text and the resolution at the f2f
... we can do a quick fix
Guus: let's add pointers to the issues in the document
<scribe> ACTION: Alistair send an email to the list by the end of next week that the reviewers can agree with and then propose publishing as WD by Jan 22 [recorded in http://www.w3.org/2008/01/08-swd-minutes.html#action02] [CONTINUES]
guus: move on into SKOS primer
-> http://www.w3.org/2006/07/SWD/wiki/SKOS/DraftPrimer Current draft of SKOS Primer
Quentin: I sent the review today
Quentin: mainly two comments: 1) we do not make
any reference in the primer to the semantics
... 2) a use case instead of separate examples would be quite useful
... other comments are related to the issues to be discussed later
marghe: I plan to send my review by next week
guus: the actions on marghe and Quentin to review the SKOS primer were not captured last week
<edsu> Quentin++
<scribe> ACTION: marghe to review the SKOS primer [recorded in http://www.w3.org/2008/01/15-swd-minutes.html#action04]
<scribe> ACTION: Alistair and Guus write draft section in primer on relationship between SKOS concepts and OWL classes for OWL DL users [recorded in http://www.w3.org/2007/11/06-swd-minutes.html#action05] [CONTINUES]
guus: the deadline for the previous action is 22 Jan
guus: move to ISSUE 36
-> http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0019.html Antoine on problems with closing ISSUE-36
Antoine: last week we made a resolution about
ISSUE 36
... actually when looking at the initial wording of the issue
... it is about linking relationships with the schema
... so the resolution is not complete, part of the problem still exists
GuusS: our resolution last week was an
amendment of a previous one
... we need to track these resolutions
... please look at the initial wording
<scribe> ACTION: Antoine to track the resolutions to ISSUE 36 [recorded in http://www.w3.org/2008/01/15-swd-minutes.html#action07]
<Zakim> aliman, you wanted to comment on label naming and to mention I have a placeholder for ... in reference
aliman: in SKOS reference we have a small note:
we haven't made any commitment on this issue
... it is implied that there will be a section showing a pattern for
querying
<Zakim> Tom, you wanted to ask about Tom's comments on Primer (in the agenda)
<Antoine> http://www.w3.org/2006/07/SWD/wiki/SkosDesign/ConceptSchemes/MinimalProposal?action=recall&rev=1
aliman: the reference has a section on SKOS and named graphs
GuusS: to Antoine: write what you think the resolution to issue 36 should be
Tom: there was some discussion on the syntax of
the examples
... graphs represented as pictures might be more readable
... it depends on the intended audience
<Ralph> I heard Tom express concern that N3 could be _less_ readable ?
Tom: which components of SKOS are basic and which are advanced?
Tom: which document should be cited for N3?
<Ralph> Notation3 (N3): A readable RDF syntax
<Ralph> Turtle - Terse RDF Triple Language
<edsu> Ralph: nice!
GuusS: it makes sense to use the same notation
in both documents
... unless there is a very good reason
... many people will read first the primer and then the reference
... i suggest to the editors to look at the pointers by Ralph
Antoine: we need to sync with aliman and seanb
<scribe> ACTION: Guus to schedule to discussion on the notation (syntax) used in SKOS examples in Reference & Primer in two weeks time [recorded in http://www.w3.org/2008/01/15-swd-minutes.html#action06]
GuusS: move to issue 44, only a few minutes for this
-> http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0052.html This and other threads in the mailing list
Antoine: there is some discussion on if
broader/narrows should be transitive
... some people are not convinced by our decision during the f2f
GuusS: I suggest that in the reference we state that broad/narrower are not transitive, discuss the rationale, and point to a specialization in which we define a transitive broad/narrower
aliman: there is a need for both of them
(transitive and non-transitive) in different use cases
... this is common pattern
... a design pattern to solve this is two have a non-transitive property and
a transitive subproperty
GuusS: I agree, I know this pattern
... but technically it cannot be a subproperty, it leads to inconsistent
semantics
aliman: one of the rules of thumb in OWL
reference is "do not mess with the vocabulary"
... I wonder if we should have rules of thumb for SKOS
GuusS: this is a different matter, SKOS is not a language like OWL
seanb: alistair's point is that if we allow users to make assumptions about the vocabulary, we can put interoperability at risk
GuusS: I suggest to leave this for now
<scribe> ACTION: Alistair to propose an approach to clarify which aspects of the extension module should be in scope for the candidate recommendation package. [recorded in http://www.w3.org/2007/12/11-swd-minutes.html#action06] [CONTINUES]
<scribe> ACTION: Alistair and Guus to prepare material for next week on Concept Schemes vs OWL Ontologies [recorded in http://www.w3.org/2007/10/30-swd-minutes.html#action04] [CONTINUES]
<scribe> ACTION: Guus to write up the issue [of Label Resource] and add to the issue list. [recorded in http://www.w3.org/2007/10/09-swd-minutes.html#action01] [CONTINUES]
<scribe> ACTION: Ralph to add pointer to Alistair's mail on grouping constructs as a note to resolution of ISSUE-39. [recorded in http://www.w3.org/2007/12/18-swd-minutes.html#action05] [DONE]
<Ralph> resolution of ISSUE-39
<Ralph> "RESOLUTION: Accept Antoine's proposal http://lists.w3.org/Archives/Public/public-swd-wg/2007Dec/0083.html as a resolution to ISSUE-39."
Ralph: do you agree with closing the issue?
<scribe> ACTION: Ralph to check whether the common interpretation of [recorded in http://www.w3.org/2007/12/18-swd-minutes.html#action10] [CONTINUES]
GuusS: the second action is a duplicate, sorry
ben: we hope we can have something for the reviewers in a couple of days
<scribe> ACTION: Ben and Michael to address comments by Tom [regarding maintenance of wiki document http://www.w3.org/2006/07/SWD/wiki/RDFa] [recorded in http://www.w3.org/2007/10/08-swd-minutes.html#action05] [CONTINUES]
<scribe> ACTION: Ben to prepare draft implementation report for RDFa (with assistance from Michael) [recorded in http://www.w3.org/2007/10/08-swd-minutes.html#action03] [CONTINUES]
<scribe> ACTION: Ben to distribute RDFa syntax draft to reviewers by Monday [recorded in http://www.w3.org/2008/01/08-swd-minutes.html#action10] [CONTINUES]
<scribe> ACTION: Diego to review RDFa syntax document [recorded in http://www.w3.org/2007/12/18-swd-minutes.html#action12] [CONTINUES]
<scribe> ACTION: Ed to review RDFa syntax document [recorded in http://www.w3.org/2007/12/18-swd-minutes.html#action13] [CONTINUES]
GuusS: if the document is available by the end
of this week, we have to postpone the decision by one week
... the decision might be scheduled for Feb 5th
<Zakim> Tom, you wanted to ask if there is an action on the RDFa editors to request a decision
<scribe> ACTION: Ben to prepare the email to request the decision for publishing on Feb 5th [recorded in http://www.w3.org/2008/01/15-swd-minutes.html#action18]
<scribe> ACTION: Ralph to review recipes document [recorded in http://www.w3.org/2007/12/11-swd-minutes.html#action18] [DONE]
-> http://lists.w3.org/Archives/Public/public-swd-wg/2008Jan/0062.html Ralph's review of the Recipes
<scribe> ACTION: Ralph see if W3C Systems Team can help with question on Apache conditional redirects [recorded in http://www.w3.org/2008/01/08-swd-minutes.html#action17] [DONE]
<scribe> ACTION: Ralph propose resolution to ISSUE-16 "Default behavior" [recorded in http://www.w3.org/2007/01/22-swd-minutes.html#action14] [CONTINUES]
<scribe> ACTION: Ralph/Diego to work on Wordnet implementation [of Recipes implementations] [recorded in http://www.w3.org/2007/10/09-swd-minutes.html#action03] [CONTINUES]
jon: the reviews are great, we are working on
integrating their comments
... we are still shooting for a pre-publication next week
... w.r.t. the comments from W3C Systems Team, not sure what to do
GuusS: if you can integrate ralph's and ed's
comments, we are in a position to publish a new draft
... decision in Jan 29
<scribe> ACTION: Jon and Diego to propose a decision on publishing the next Recipes draft by next week [recorded in http://www.w3.org/2008/01/15-swd-minutes.html#action23]
GuusS: (to the editors) make sure that the WG has the proposal
Elisa: planning to hold a call later this week and to work on our action
<scribe> ACTION: Vit and Elisa to include in the document all the target sections plus an allocation of sections to people and potentially a standard structure for sections [recorded in http://www.w3.org/2007/10/08-swd-minutes.html#action07] [CONTINUES]
[adjourned]
<Ralph> scribenick: ralph
Antoine: the idea is that transitive broader be
a superproperty of broader
... if we do that then statements using broader cannot be retrieved using the
super property
... the standard modelling pattern is good but we have a standardization
problem
Daniel: so there really aren't two types of
'broader'
... to me, there's only one kind of 'broader'
Sean: seems to me from reading the discussion
that people want to be able to query against 'broader' and get transitive
closure on query
... so there's really only one 'broader' but there's a way to query over a
more general notion
... the general notion would not be used in assertions
Daniel: I don't see a difference between query and assertion
Sean: there may be inferences I can draw from assertions
Daniel: in the OWL community you make limited assertions and do a lot of inferencing
Alistair: if transitive form is superproperty
we could have a convention that we only ever assert the subproperty
... but the superproperty is available for query
<Quentin> http://www.w3.org/2001/sw/Europe/reports/thes/1.0/guide/20040504/#3.9
Guus: [worries about community usage]
Alistair: choice of which is subproperty and
which is superproperty
... direct one could be 'broader' -- the one that people use -- which would
affect existing SKOS data
Sean: is there really an analog for the
transitive closure of 'broader' in current thesauri ?
... if I have a paper thesaurus, I don't really have a transitive closure
without having to do a lot of work
... the transitive closure is not actually represented anywhere
Alistair: agree, but the point is the
practicalities
... in certain applications it is convenient to compute the transitive
closure and then query it
Sean: that's fine, which suggests the pattern of using direct 'broader' in assertions and a transitive 'broader' that I use in queries
Guus: whatever we do, the transitive property
should be a superproperty of the direct one else the semantics are wrong
... the direct property: "a is a direct broader term of b", without saying
anything about transitivity
... remember that transitivity does not inherit
Daniel: I'm worried about this confusing the community
Guus: Sean points out that the community does not currently have this logical notion; they do it at query time
Sean: yes, I'm suggesting non-transitive
'broader' used in assertions and a different, transitive, relation that is
only used in query
... I hope this satisfies those who want 'broader' to be transitive in some
way
Alistair: I agree
Guus: but be clear that this is not [currently] being used in a pure logical way; it's a procedural thing
Ed: there's an example at the end of the primer, but it doesn't follow this pattern and would need to be changed
Quentin: the transitive version of broader and
narrower should be present; we're speaking of creating knowledge organization
systems so these should be taken logically
... in some application the developer might want to use SKOS as a simple
representation and might want some very simple logical inference
... without requiring the full capabilities of OWL
... their concept of hierarchy is a simplification of subsumption
... this might just mean that we need to look at a SKOS extension
... but I know there is opposition to extensions as they require additional
namespaces
Guus: if you're going to write an assertion, e.g. in a namespace document, you use 'broader' and if you want to write a query you can use 'broaderTransitive'
Alistair: if we do have 'broaderTransitive' or
'broaderClosed' in SKOS then two of the semantic conditions in the data model
become very easy to state
... e.g. 'skos:related disjointfrom skos:broader'
... and to assert some irreflexive relations
... I would like to see a broaderTransitive/broaderClosed superproperty
described normatively
... rather than omitting it or leaving it to a community extension
Antoine: agree
Sean: agree, and it would reduce the repetition of this discussion
Guus: any chance of getting this written up for discussion next week?
Alistair: are you suggesting we introduce two new terms in the SKOS vocabulary and include them in the editor's draft?
Guus: yes, in particular the editor's draft we're going to review next week
Ralph: I'd recommend sending this to the WG in a separate email
Sean: related to ISSUE 44
<aliman> Al's notes ...
<inserted> scribenick: aliman
quentin: transitive version of broader should be present -- speaking of creating KOS, forcing applications to take them logically. In some systems & applications, use SKOS as simple representation, and simple inference with it (not full OWL). SKOS vocab to do thesauri, taxonomies, hierarchies, concept of hierarchy very simplification of subsumption, as broader is loose meaning. SKOS...
transitive as super ...
antoine: problem, all statements asserted using transitive broader cannot be retrieved by
daniel: aren't two types of broader?
sean: if use pattern (transitive super) don't use that for assertions, use for querying? people want to query against broader, and get transitive when query; assertions about direct;
daniel: assertions vs. query?
sean: assertions -- directly asserted; may be inferences I can draw.
daniel: proposing two different types of broader, confusing. agree with you, make minimal assertions, do the rest by inference, legitimate.
guus: BT standard term in thesaurus community; what people state as BT is always direct broader; so by definition, our semantics of broader, if it is equal to BT, then it needs to be not transitive, otherwise people get confused.
quentin: as a sub-property, examples described as in skos core guide?
guus: sub-property has to be direct; if do that, what we call broader, will not be same semantics as thesauri, because only assert direct one. that's only way semantics.
aliman: other way around from guus, would affect existing SKOS data; if do as guus says,
sean: analog for transitive closure of broader in thesauri? If have a paper thesaurus, don't really have transitive closure, not represented...
aliman: required
sean: to have skos:broader as direct, and introduce some new super-property as transitive closure
quentin: I would agree as well.
guus: I you want to have a transitive and a direct, then transitive is always super-property. transitivity doesn't inherit
daniel: existing relations so, worried about confusing the community. How do you know which to use?
guus: sean is saying, in community, don't have logical notion. Do it at query/computation time.
sean: broader used in assertions, not transitive, then property used in query which is transitive. relatively clear statement, answers concerns of people requiring broader to be transitive.
guus: clear not being used in a logical sense; if want to get closure, have to do procedural thing; haven't seen logical use of thesauri yet.
sean: needs some careful explanation in reference and primer.
ed: in tail end of primer, example of doing it not the right way, will have to be changed.
antoine: will not be huge effort
... extension described in docs earlier, need to look at again.
guus: broader & broaderTransitive should be in spec, if write docs, use broader, if want to query, use broaderTransitive. If SKOS spec specifies broaderTransitive.
aliman: makes some conditions easier to state
guus: comes down to wording
aliman: would like to see super-property in the spec
antoine: i agree
sean: i agree
guus: i agree too ... can we have this in some short form in editor's. Two new URIs in SKOS vocabulary. Suggest broaderTransitive rather than broaderClosed
ralph: recommend separate email on this -- here's what we've done to editor's draft and why
guus: who is issue owner? temporary resolution of issue 44. I'll write it tonight.