SWBPD WG Face-to-Face

4 Nov 2005


See also: IRC log


AListair, Danbri, BrianMcBride
ChrisW, dbooth, FabGandon_, Alistair




<Raphael> hi RalphS:-)

<aharth> hi, morning!

<RalphS> yes

<dwood> yes

<ChrisW> we can hear you

<RalphS> Guus: 11 people now in the room

<RalphS> would someone write here the names of all who are present, please?

<dwood> Guus, My apologies for not being there!

<ChrisW-scribe> Agendum: Waitinf for latecomers to wander in

<RalphS> I wish I could be there too

<ChrisW-scribe> small groups greeting and mumbling with each other

<RalphS> Scribe: ChrisW

<RalphS> ScribeNick: ChrisW-scribe

<RalphS> Chris, could you write here the names of all who are present, please?

Guus thanks Andreas & DERI

<danbri> +1

<RalphS> Alistair Miles

<RalphS> Dan Brickley

<RalphS> Brian McBride

scribe: Andreas, Libby,
... BenNguyen


scribe: Jacov, Raphael

<Raphael> Giorgos


<Raphael> Vassilis

<Jacco> s/jacov/jacco

<Raphael> George

two giorgos


evan, elisa



Jeff Pan and Aldo Gangemi arrived

zakime, meetingroom also has Aldo, JeffP

jjc: should we establish criteria for moving notes forward now

<danbri> [oh, huh... i didn't show up as registered; am sure I did, but I guess must be mistaken]

jjc: a lot of documents

guus: can we have a concrete proposal?

ralph: what are you suggesting? To change the practice or repeat it?

jjc: concerned about running out of time and publishing unfinished notes as a result
... make sure to apply process properly, regardless of time

<aharth> dwood, i'll ask the technician again...

<RalphS> yes, concern about rushing things through as we get close to the end of January is certainly justified

guus: clear goal of meeting is to get to the notes within the contrants

<danbri> [ for remote participants --- http://www.flickr.com/photos/danbri/59645352/ ]

guus: have a proposal by end of meeting for coord. group regarding future of WG

<RalphS> not cantankerous, Jeremy -- you raised a good point

Agendum: PORT TF


SKOS Core Guide


Discussing: http://lists.w3.org/Archives/Public/public-swbp-wg/2005Oct/0195.html

Alistair basically summarizing above note

ITEM 1: 'Extending' SKOS Core

Discussing: http://isegserv.itd.rl.ac.uk/cvs-public/~checkout~/skos/drafts/appextensions.html

<RalphS> this seems to be a good primer on how to use RDF for vocabulary definition, applied to a specific case

<RalphS> and as such is quite useful

<RalphS> DWood: is there a the difference between "extending the SKOS Core" and "extending SKOS"?

alistair: skos == skos core

danbri: concerned that its much easier said than done
... does not seem to be tool support for it - this would set back adoption

<RalphS> "its" == "RDFS inferencing" (Alistair's term)

jjc: we (W3C) made the decision a while ago to make RDFS inferences cheap enough that web tools will support them
... there is some support in Jena for this
... the hard bit is figuring out which schema to dereference

brian: what scale of inference

alistair: things like Mesh, dewey-decimal, etc. woudl publish their own extensions

guus: agree w/ jeremy that we shouldn't back off from inference
... AListair was too optomistic about the number of triples (underestimated)
... real scalability issues but that shouldn't stop us

<RalphS> I agree with Guus -- let's use RDFS and encourage the tools to support subPropertyOf and subClassOf efficiently

jacco: don't see that there's a problem if you warn users to use subproperty reasoning when you need it

danbri: at least three parties, publishers, extender, and data sets, and users
... yes, that's four

brain: not convinced we shoudl recommend something that is currently impractical


<RalphS> [I didn't understand which bits Brian thinks may be impractical]

jjc: wants to spank Brian

guus: have to be aware of scalability problems

danbri: non-rdf experts will be downloading rdf tools, its a different thing to publish things for which there is no tool support

<dwood> I don't think we have to limit ourselves to available tools, as long as our recommended solutions are implementable. I am *always* worried about scalability.

danbri: this community (thesauri?) are not SW experts. if we encourage them to extend skos there is a risk they will reinvent owl
... badly

<RalphS> [I appreciate DanBri's partial photo of the meeting room http://www.flickr.com/photos/35468151816@N01/59645352/ -- Thanks, DanBri]

jjc: a wider issue: there are statements in SKOS that have a wider semantic ... meaning
... that is, formal semantic constraints
... e.g. constraint on language tag
... basically, that SKOS is already moving into the territory danbri was worried about?

alistair: there are other constraints, like no two concepts shoudl have the same preferred label
... but some concept schemes where you may want to break that

danbri: how far skos is extended to become a modeling framework is a concern

alistair: we promote the extension of owl core until it begins to tread on owl

^owl core^skos core^

<Zakim> bwm, you wanted to ask whether there is existing practice of this form of extension

alistair: there is some experience: gemet thesaurus
... not enough pracitcal use but a lot of interest

<dwood> Is it reasonable to explicitly state that SKOS should not be extended in any manner which is already defined in OWL?

jjc: it just describes a simpl e use of RDF the way it was intended

<RalphS> +1 to Jeremy's statement

Brian: (previously) questioned whether the experience was enough to make a "best practice"

<RalphS> this is showing poeople how to use RDFS

Guus agrees with Jjc

alistair: hybrid ontologies, ranges. Class of things called enviromnental organiztions, they have project type, class of concepts, simple taxonomy. Please share thoughts abou this.

<RalphS> [I'll ask via mail later (after looking further myself) whether there is prose to explain the use of skos:definition vs. rdfs:description]

<RalphS> ITEM 2: Importing SKOS Core into OWL (DL) Ontologies

Alistair: one schema for SKOS

some of the bits are not compatible w/ OWL-DL

proposal: have an rdfs version, and owl-dl version, and a merged version in owl full

<dwood> This raises an interesting point in general: Is there any reason that legitimate uses of RDF or RDFS should be prohibited from conflicting with OWL? I think that it be best practice NOT to, but shouldn't necessarily be prohibited.

some of the owl-dl problems:

some subprops of rdfs:label

most properties are just rdf:Property

scribe: lots of annotation properties so can't be "extended" (subprop'd) in OWL DL

<RalphS> [I understand the speakerphone challenge now, seeing the size of the room in http://swordfish.rdfweb.org/photos/2005/11/04/2005-11-04-Pages/Image20.html ]

fabian: concerned that there is some redundancy so that in the merged (owl full) version you have two things

<RalphS> [I hoped we could use lang properties to permit inclusion of multiple rdfs:label properties in a single schema information resource]

alistair: yes, also a problem for maintenance
... but ok because it makes it simpler

guus: is there a need for OWL-DL in the thesaurus world, do we need to worry about it?

jjc: see some semantic constraints that are outside owl-dl
... so agree w/ guus that may not need to worry about it

danbri: owl-dl tool referencing skos data would be polluted if there isn't an OWL-DL version of skos

brian: seems like a common problem - how to describe different versions in different languages

<RalphS> Brian raises a good point -- this is suggesting a "best practice" for dealing with RDFS + OWL-DL would be useful in the community

brian: do we need a best practice

ITEM 3: Problem with SKOS Collections

<RalphS> I vote for favoring RDFS:)

discussing: http://isegserv.itd.rl.ac.uk/cvs-public/~checkout~/skos/drafts/collections-5.html

discussion to clarify the thesaurus terminology

basic issue: the "narrower term" is a collection which is not supposed to be a skos:concept

scribe: but range of narrower term property is skos:concept

aldo: this is not a collection, this is just an "unnamed" concept

<dwood> I am concerned, but not surprised, that SKOS (as a major new use of RDF) has so quickly run into issues surrounding OWL-DL. The concept of an "OWL-DL version of SKOS" sounds nasty and brutish. +1 to work regarding guidance.

guus: the problem is that "chairs by function" is a meta-class

<RalphS> "concerned, but not surprised" -- yep:(

jjc: no examples of this - would be good to have some

alistair: some times these "collections" are ordered and the order means something

mikeU: could use annotation properties indicating metaclass membership for OWL DL version

<dbooth> But a "collection" does not say what the ordering means.

alistair: problem is that a concept can appear in two places so an annotation property won't tell you which place is the one being faceted

<RalphS> yeah, I'd like to see if there's a different relationship (than owl:disjoint) between skos:Concept and skos:Collection

jjc: agree w/ aldo that collection and concept are the same

<RalphS> hmm. perhaps not the same but there's a connection

<dbooth> jjc: Things like union create unnamed classes. Perhaps this is what's happening.

guus: the problem is that skos:concept is used for anything that can be used in annotation

discussing solution seciton of http://isegserv.itd.rl.ac.uk/cvs-public/~checkout~/skos/drafts/collections-5.html

<RalphS> [I'm understanding now, participating in this 'remote' mode, why it's so tempting to type thoughts in irc rather than interrupt the f2f conversation.]

jjc: is the problem just that skos:concept is misnamed
... should it be named skos:namedConcept ?

alistair: there are coordinated (combined?) concepts also

guus: there are two kinds of concepts: indexing and those for organizing the indexing concepts
... separate naming issues from organizing issues

jjc: or a compelling example of why that won't work

guus: clearly presented - propose we hav ea breakout session to resolve the issues
... for planning - it has been suggested that SKOS be one of the primary outcomes of SWBP
... taken very seriously in library community
... perhaps should be taken into a rec track

<RalphS> [Isn't this question of "what's the future path" part of tomorrow's agenda?]

guus: would beg the question as to what we need to finish in swbp, and what to leave for new wg?

alistair: defer judgement on whether it shoudl be rec
... in this wg would like to resolve these three issues + possibly one or two recent ones

<RalphS> [I heard Alistair say "defer judgement on rec to [??]", didn't hear to whom he thought it should be defered]

[to people who know about "that kind of thing"]

guus: would rather finish the basic uses of skos in swbp and save the rest for next wg

<RalphS> hmm. would "the people who know" be the users of SKOS or someone else?

danbri: only meaty problem is the final one (collectsion/concepts) so I think it could be moved to new wg

[guus is going around the room]

dbooth: not speaking for HP - but seems like an application of SW technology and not really in W3C domain
... but may not happen if it isn't

ralph: would like alistair to give a better sense of where the community is
... much more "horizontal" an effort than some things W3C has considered
... don't think WG has time to finish all three issues raised today
... makes an excellent note on how to use RDF

<dbooth> On one hand, SKOS is an application of RDF, and thus might be done outside W3C. On the other hand, it's clear that the outside community is not ready yet to be able to do this work correctly. Thus it may make sense for W3C to do it correctly, and it will give a lot of visibility of Sem Web technology to the larger community.

<RalphS> specifically, item 1 makes an excellent example of how to use RDFS and I'd like to see that item be the principal milestone for end of january

can't hear you

<RalphS> +1 to seeing SKOS work continue in W3C

dwood: would like to see skos continue beyond swbp and rec track

<RalphS> [oops; network disconnect between MIT & Galway?]

jjc: who would participate in a wg?

different library community people

shoudl be involved

not typical W3C people

so, consider whether they are likely to join WG

<RalphS> [I'd like to know how those folk have been participating in the SKOS design up to this point]

spoken to Michael Wilson who is UK w3c person

lots of interest

[they've been participating on the mailing list]

<Zakim> Ralph, you wanted to suggest that extensions be the primary SWBPD goal

danbri: 23 posters on mailing list + a lot of observers

<RalphS> DanBri: there were 23 distinct individuals posting on the [SKOS] mailing list last month


<RalphS> [I guess that's the public-esw-thes@w3.org mailing list]


<dbooth2> Scribe: dbooth


Temporary chair for this session: ChrisW

Aldo: TF was about bringing WordNet (large english lexicon). Work: 1. Create data model for WordNet. 2. Create ontology, relying on existing project but not within W3C TF.
... Data model has been created by Guus and finalized by me. Covers the WordNet schema from the prolog source
... One document distributed by Mark. Some general issues on the work of the TF, some raised by Jeremy.

<RalphS> Jeremy's comments on WN

<RalphS> Aldo's response to Jeremy

Aldo: Also want to apply WordNet model to Lexical Markuup Framework.


Viewing http://www.cs.vu.nl/~mark/wn/wn-conversion.html on screen

Guus: Resolved RDF/OWL in a different way than SKOS. Didn't have RDF separate file issue. Everything that is an object has an owl propoerty statement.
... Can read the same file and ignore the extensions.
... Issue was: What elements of WordNet should get URIs? DanBri made the point that URIs should have a human-readable format, and we've done that. AFAIK this current version could lead to consensus. More importantly, we have thhe transformation program, which is important to Princeton people.
... Main difference from current: relations are partial.
... Use 2.0 because 2.1 does not have the prolog yet.
... What's in the prolog source will be in the RDF source.

Aldo: Wordnet source uses something that is not completely finished.

Guus: This comes under the category: What should be the maintenance strategy? I suggest start from Jeremy's email suggestion.
... First point: define a mapping from WordNet to OWL.

Brian: What are the TF's plans?

Aldo: Most of the port is done. Should be published/used. Issues on how to use it. Translation 100MB.

<dwood> What is Princeton's position on maintenance?

Aldo: One option: small files. Depends on tech you want to use. possibility: submitted a paper to OWL REC conference in Genoa. Proposed a panel for discussing WordNet, lexical resources and SemWeb and they're willing to do it, so it's low hanging fruit.
... Princeton people are busy: new things to do with WordNet. Best we can do is to provide them with somethign we can use. (First step of plan)
... Second: Involve linguistic assoc.
... Third: harmonize the various wordnet efforts, such as neufchatel and others, to avoid doing the same thing multiple times.

Jeremy: Timescale?

Brian: That third one is more an objective than a plan.

Aldo: Can propose to them to join the effort.
... Re timescale: can provide stuff to Princeton in the next few weeks.
... On second one, the conference is in May, so it would have to be after that.
... On the third one, I'm already in communication with LMF ISO committee.

DanBri: It's fantastic to reach out to these communities, but the traditional W3C way to do that is through publishing working drafts. Gets muchh wider readership, gives greater transparency.

Aldo: We have a draft of a WD.

Guus: my take on timing: If we appoint internal reviewers, within 6 wks we can get a WD out.
... Including WG decision.

ChrisW: WD as item 0 in Aldo's plan?

Guus: Yes, suitable for princeton and language community.
... If we take the normal process of gettign their input, it would push us past Feb.

ChrisW: But the WD seems like a vehicle to reach out to others.

<RalphS> [boo on asking Princeton to comment formally on Editor's Draft]

Guus: Yes. Princeton has already looked at it.

<RalphS> ["boo" == -1]

Brian: WD would be based on the doc linked from the agenda?

Guus: Yes.

Brian: When I read it, the doc seemed more about the internal process for doing the translation. I would have thought we would want to publish a description of WordNet in RDF: here's the structure and how to use it, rather than a tech note on the rational for the decisions we made. Good doc, but not the doc we want to publish.

<Zakim> Ralph, you wanted to ask the WN TF how far from a first WD they think the current Editor's Draft is

Guus: This is not for the user's of wordnet, but the process we discussed. For getting consensus on this version, that's what this doc addresses. I agreee you'd need a different doc for users.

<dwood> +1 on a public WD first

Ralph: +1 to danbri's comment on publishing a WD to include the community. Concerned about princeton being asked to comment on the ed draft. There's a balance between getting input from interested people and having public visibility. I encourage getting this published as WD for public comment.
... How much more time needed (editorial) before asking for WG review?

Guus: Current WD could be reviewed now. It's in pretty good shape for the purpose that we've discussed. Though I agree with Brian that we'd need a different doc for users later.

Jeremy: Brian asked who the doc is for. But TF charter gives various objectives, and took 18 months to reach. I have no confidence that it can now be done in 6 weeks.

Guus: I now have a student working on this.

Jeremy: Prereq for second doc are political issues.

Guus: Not for writing it. Getting consensus yes.

Ralph: We can put whatever level of caveats we want in the status of the doc, but I agree that consensus takes a long time, and the first step is to put a draft out, as rough as necessary, and let the community comment. Then we can start getting feedback to address Brian's concerns. I'd like to have a draft out in public and all comments recorded in the list archives.

Guus: Current draft is adequate for pub in my opinion.

Jeremy: My main concern is what do we say about future expectations for it.

Guus: Future work is separate discussion.

<RalphS> We can note in the SOTD that the WG charter expires in January and we make no promises about what will happen to this WD post-January

ChrisW: Guus is saying something significant in the TF has changed, and more work is happening now (in the past month) so you cannot judge the rate of progress by looking at the whole past year.

danbri: It would be great if the work were finished, but nothing wrong with publishing as far as we got.

<RalphS> +1 to DanBri's "publish what we've got to save the world having to reinvent this"

<dwood> yep, get it out to the world

danbri: Also tech discussion relates to planning and roadmap. I've always liked mapping synsets on to classes. Has a popular idiom.
... Issue: Same URI can be used for the class and lexical representation? And what does URI dereference to?

<Zakim> jjc, you wanted to make two review comments on current draft

<RalphS> Guus: I think the WordNet work could continue within the Semantic Web Interest Group

<THE_CHAIR> ping

<RalphS> [the last line I see from scribe is ":45] <dbooth> ... Issue: Same URI can be used for the class and lexical representation? And what does URI dereference to?"

<Ben> connection problems here with the wireless

<Ben> david will send email of what's been missed

<THE_CHAIR> dbooth will send back scribe

<RalphS> <insert DBooth's mail here>

<RalphS> </insert>

<RalphS> :)

<Ben> danbri: thinking about the relationship between skos and WN

<Ben> ... a lot of worj to be done in skos needs owl people

Danbri: Re relationship between SKOs and Wordnet. Lots of SKOS work needs OWL people's brains.

<Ben> ... people approve of the skos approach, they also want to talk about terms

Danbri: Faceted stuff scares me, because it is reinventing OWL.
... Strawman work item might be taking terms in wordnet and using them with SKOS. SKOS-only WG would be unhealthy -- only thesaurus people.

Alistair: __ have terminology markup, and they know a lot about multilinguality, and they've started looking at SKOS.

Guus: Happy to have these options on the table.

Aldo: Community of Wordnet community are dynamically holding. We're providing something to support the mainstream.

<THE_CHAIR> ^evolving^evolving^

Aldo: Might want to suggest practices to transform.

<RalphS> [would the TF like to solicit reviewers now?]

<RalphS> [ah, Guus already on it]

ChrisW: No decisions made.

Guus: I propose we assign WG reviewers to ed draft. Volunteers?

Volunteers: Alistair, Brian, Ben, Jacco, Jeremy
... plus DanBri, though also a TF member.

<scribe> ACTION: Ralph to copy the WordNet to CVS space and post URL [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action01]

Review deadline: 25-Nov-2005

DanBri: mailing list for SKOS worked well for us.

ADJOURNED FOR LUNCH -- continuing at 2:00pm local

<RalphS> [54 minute lunch break]

<RalphS> Mark's WordNet conversion document is now copied to http://www.w3.org/2001/sw/BestPractices/WNET/wn-conversion.html

<RalphS> ?!

<RalphS> oh well

WG thanks Stefan for hosting!

<Ben> jacco: brief overview of what MM has been doing in the last weeks

<Ben> ... things don't look ready yet

<Ben> 3 issues

<Ben> 1 first deliverable

<Ben> 2 possible structure on 2nd deliverable

<libby> stefan has left...

<Ben> 3 future work (vocabularies image annotation, and if recharter extend to video annotation)

<Ben> 1- image annotation on the semantic web

<Ben> contrbutors: giorgos, raphael and JAcco, many other contributors

<Ben> goal: provide a survey and overview of the vocabularies and tools for people who want to do image annotation on the semantic web

<Ben> discuss them in the context of use cases

<Ben> main challenges: you run a risk of mentionning some projects and missing others

<Ben> lost of work and lost of progress in the area: problem of keeping up to date.

<Ben> ... problem ensuring documents outside the w3c namespace remain stable. This has been moved out of the main draft

<Ben> if the main draft is ready, this can be published as a wg note

<Ben> 4 use cases with example solutions

<RalphS> [I propose using the W3C Wiki for the 'living' resources material]

<Ben> to do list:

<Ben> 1- lot of clean up

<Ben> 2- draft suffered from time pressure

<Ben> 3- better integration of sections

<Ben> 4- important use cases missing: news related images, scientific images, solution of media productions services use case (still working on this)

<Ben> Issues, the use case examples are seperate documents. Should this be integrated into the main document.

<Ben> Problem: these solutions might get outdated. especially the vocab.

<Ben> However the main principles remain valid, so long as people know they should use updated version of the vocab

<Ben> jeremy has raised the problem of patent problems

<Ben> jjc: ACTION: verify or not if there is a patent issue and if there is clarify distinction

<Ben> jjc: general rule not to discuss patent issues on the public mailing list

<Ben> Jacco: scoping. there is a broad range of images to be annotated. there is a problem about scoping the document. For the moment it is manageable in size.

<Ben> ... most drafts assume that SW technologies are a good idea and want to know how to use it in a best way

<Ben> ... this document is different: public for this document are not yet convinced by using SW technology in order to annotate image documents (but they already use meta data)

<Ben> ... the question is: what communities to target ?

<Ben> ... EWIMT workshop in london soon, also SWAMM workshop in may at WWW conference

<Ben> ... can mike and libby comment on this ? Mike has submitted a long email that I have not had time to read yet.

<Ben> Chris: this is a big WG, I have trouble remembering the references of deliverables

<libby> I still owes a written review

<Guus> q\q?

<Zakim> ChrisW, you wanted to request referring to deliverables by name not number

<Ben> jacco : 1- overview of the vocab and tools out there for SW annotations for images

<Ben> 2- interoperability between SW and non SW approaches

<Ben> Chris: call 1 overview and 2 interoperability

<Ben> Mike: general remarks: a lot of good content in the document.

<RalphS> MM] Review of Image Annotation Document [Mike Uschold 2005-11-03]

<Ben> ... discussion on images not relevant for the target audience

<Ben> jacco: we dont want to give people false expectations.

<Ben> .. the SW in itself wont solve all the issues with image annotation

<Ben> Mike: lots of use cases, general positive view on the note

<Ben> .. there is much to do still. Areas of concern: the objectives need to be more clear. The stated objectives are weak.

<Ben> .. include: SW is a benefit to this area

<Raphael> Discussing Mike's review: http://lists.w3.org/Archives/Public/public-swbp-wg/2005Nov/0020.html

<Ben> .. more emphasis on motivating the use of SW technologies.

<Ben> .. the description of the use cases were inconsistent in style and format

<Ben> ... need a list of all the use cases

<Ben> ... looked at a few use cases. Not acceptable to dump the reader an rdf ontology. The user should have a text description or graphical representation.

<Ben> Raphael: this will be done (?)

<Ben> Mike: the document does not have a coherence, a lot of cut and paste.

<Ben> ... a lot of specific comments I will not go through.

<Ben> Jacco: thanks, comments were helpful

<Ben> One of the main goals to have a structure to show were the document was going to. Is direction ok ?

<Ben> Mike: yes

<Ben> Mike: I volunteered to be internal reviewer

<Ben> Guus: decision to publish as WD under the condition comments get addressed ?

<Ben> ... would you recommend publishing as WD ?

<Ben> Mike: I think so but there is a lot of work to be done

<Ben> Jacco: I'd like a second review

<Ben> jjc: its a long review

<Ben> guus: lots of detail. Is there anything essential to be changed before you are happy with the document as a whole ? [to Mike]

<Ben> mike: more coherent stories + summary of the use cases into the document.

<Ben> Guus: is it juste editorial process ?

<RalphS> <ben apparently agrees to scribe off-line; insert his comments here>

<Guus> q"

<FabGandon> guus: nothing prevents the current doc to move to working draft

<Ben> jjc: I'd prefer a style where the document is interesting and not trying to sell the document

<Ben> Mike: I agree

<Ben> jjc: I thought this was an interesting document

<Ben> mike: most of the other notes assume that the semantic web is useful. It felt like this document has target readers in the MM domain

<Ben> .. in order to feel they should bother reading the document, it should help to motivate them to show the potential the SW has to show in this area

<Ben> jjc: the document needs to discuss that

<Zakim> Ralph, you wanted to object to publishing in current state

<Ben> Ralph: i agree with mike

<Ben> it shouldn't be a drawback from publishing

<Ben> ... would not obejct to pulish in this form

<Ben> ... would leave to othe editors to do the editorial work before it goes public

<Ben> Ralph: jane needs to go through the administrative process of joining the WG before the document may be published

<Ben> Guus: I have mailled her about this

<Ben> ... it is an action she needs to fullfill

<Ben> ... jacco to take an action about reminding her

<RalphS> I also have sent mail to Jane to point her to the Invited Expert application

<FabGandon_> scribe: FabGandon_

ibby: to review after revision

lLibby: document touching too much different problems

<RalphS> [btw, the Invited Expert administrative process is not just administrivia -- it's getting explicit agreement to certain IPR terms from the individual, and that's really what's important here]

<dwood> yes?

<dwood> OK

<dwood> Does anyone have a switch and some cable?

<RalphS> [meeting room scribes going off-line due to network issues at DERI]

<danbri> maybe folk who are connected could duplicate some of real scribing, for ralph + david

<danbri> eg. "david - i reviewed it too, tho i didn't have some of the links. i liked it. and desc of use cases...

<danbri> ...2 things came to mind

<danbri> ...wasn't clear to me where the work is going

<danbri> ...could that be made clearer?

<danbri> ... was left w/ feeling "ok interesting - but what next?"

<danbri> j: see the 'what next' docs

<danbri> david: title seems broader than the actual paper

<danbri> ...suggest narrowing the title

<danbri> j: i'll have to think about that

<danbri> d: what i read was a good description of use the cases

<RalphS> j = jacco

<danbri> j: goal is not to only do use cases. but also to do tools + vocabs. use the use cases to organise this discussion

<RalphS> d = dbooth

<danbri> ...prob is that the meet re tools + vocabs is in the other vocabs

<danbri> dbooth: since i didn't see the other docs, i can't comment

<danbri> jacco: may be we can scope title better

<danbri> raphael: [missed point re a proposal from ralph]

<danbri> jacco: i like suggestion from ralph to do it in a wiki context

<danbri> raphael: ok

<danbri> jacco: 1 more request for libby -- -re targetting communities for public feedback

<danbri> libby: certainly

<Ben-scribe> Guus: online scribing is off. We try to fix asap. Even wire has a problem.

<danbri> guus: i'm going to propose to this wg to publish an edited version, at editor's discretion

<RalphS> I object to publishing this until Jane Hunter's status is corrected

<danbri> ben-scribe --- is my psueo-scribing ok here?

<danbri> (missed point vfrom brian)

<danbri> jjc: [missed]

<bwm__> brian: can I take that for each use case, one of the contributers to the document has specific knowledge of that use case?

<bwm__> jacco: yes

<danbri> guus: publish 2 conditions. (1) IP issue resolved to ralph's satisfaction (2) jane hunter ack / status is resolved

<RalphS> yes, once Jane Hunter has Invited Expert status I am OK with publishing

<danbri> ... under those conditions i propose to publish this as a wd of swbpd wg

<danbri> ...with editorial discretion

<danbri> ralph: what does editorial discretion mean?

<danbri> guus: take on board comments from wg

<danbri> ralph: i'd feel more comfortable with specifics, eg. 'take on mikes xyz comments' not a general open door to changes

<danbri> guus: mike, jeremy, libby, david's comments

<RalphS> fine to that list of specific comments

<ChrisW> !-2^dbooth^dbooth^

<danbri> danbri: relays jjc's concern to ralph

<danbri> ralph: ack'd

<danbri> guus: accepted by consensus. thanks editors!

<danbri> (no objections etc)

<danbri> action on taskforce coordinations to prepare a version for publication

<danbri> jacco: existing standards won't go away

<danbri> ...existing standards interop w/ rdf etc [missed detail]

<danbri> raphael: there's currently 3 or 4 versions of mpeg 7 [ in rdf/owl?? -danbri]

<danbri> ...don't model things in same way

<danbri> ...semantic vs syntactic interop

<danbri> guus: are you saying the standard is ambiguous?

<danbri> jacco: problem of lack of formal semantics...

<danbri> chris: re wordnet, skos ... similar role --- formalising things that haven't been yet

<danbri> [tm also, i think]

<danbri> mike: one way to get around the 'is it semantic or syntactic' q is to just avoid that way of couching the question

<danbri> [valentina arrived just after lunch]

<danbri> [missed]

<RalphS> oh, when the connection was lost zakim erased his memory

<danbri> [...]

<danbri> mike: this is great, ... but scope seems enormous

<danbri> ..."boy that's a lot of work!"

<danbri> jacco: i think impossible before feb

<danbri> mike: perhaps a cursory treatment of each?

<danbri> mike: can't poss do a good set of recommendations on that schedule

<danbri> ...10-12 pages, several parags for each

<RalphS> oops

<RalphS> sorry, Chris

<ChrisW> ok

<danbri> danbri: re doing MM stuff in SWIG ... my main worry is patent policy stuff, given that relfecting other orgs work into w3c specs has IPR issues

<danbri> guus: [summarises for ralph]

<danbri> ralph: i agree w/ danbri's concerns

<danbri> ...maybe other options

<danbri> ...eg. to work on specific vocabs

<danbri> ...that seems to me that it might be a candidate for a 'w3c incubator group

<danbri> [ a new mechanism within w3c]

<danbri> ...an incubator [aka "XG"]

<danbri> jacco: pls post a pointer

<RalphS> Incubator Group is an option for working on vocabularies

<danbri> jjc: on dan's point re intellectual policy

<danbri> ...w3c patent policy only applies to rec-track

<danbri> ...so if there is concern re IP, then that could be an arg for patent policy

<danbri> ...even in a wg context, targetting a wg note, patent policy is largely silent

<danbri> ...so, in as much as there are patent in this

<danbri> ... a rec-track might be needed

<danbri> [...]

<danbri> jacco: if the wg is to be rechartered, when might we know this?

<danbri> ralph: you'll know that, when there's a proposal to the ac, to recharter groups

<danbri> ...its not likely to happen much before end of january

<danbri> jjc: historically, many groups have over-run

<danbri> ralph: [missed]

<danbri> phil t: what are we talking about exactly re new charters?

<RalphS> yes, true that groups have over-run their charter but that's not recommended

<danbri> guus: could be a simple exstension, or it could be new wgs etc

<danbri> phil: seems significant work and potentially part of a new charter

<danbri> guus: in our 1st meeting in Cannes, was rightly recognised that link to MM was important; yet difficult and hard to progress.

<danbri> ...we've seen precisely that happen.

<danbri> ...i think, over course of the 2 days, we're getting a sense of what each tf needs

<danbri> ...summarise in afternoon

<danbri> phil t: i agree w/ jjc's point re ipr

<danbri> ...originally at time group started, different aspirations around semweb

<Zakim> Ralph, you wanted to say that we should work within our charter but the work we do suggests where to go in the future

<danbri> ...does it need to change? maturity? areas of interest?

<danbri> ralph: [as above]

<danbri> ...i strongly recommend that this wg work within its charter scope

<danbri> ...both timelines and otherwise

<danbri> ...thats not to say that we shouldn't continue working up until very last moment

<danbri> ...the MM TF helps us all understand this area

<danbri> ...all these questions (what happens where, etc) ... are informed by what goes on in the TF now

<danbri> ...but please don't assume you can continue on past end of january unaffected

<danbri> ...q of what happens after Feb ... we'll keep talking about that between now and then

<danbri> guus: as we discuss each TF, ... will make a list w/ status of each TF, get a sense of scenarios...

<danbri> [...]

<danbri> guus: to contiue this, need to make it part of a new charter

<RalphS> Incubator Activity Approved, Call for XG Charters [Member-only link]

<danbri> ...within 6 months we might have a better stab at 1st note but not a great deal more

<danbri> mike: there is an overlap betwene this [?] note and an [?] OEP one

<danbri> guus: multimedia perspective... refelcts a different standards community

<danbri> mike: agreed

<danbri> ...what would be done, ...eg take EXIF thing for pixel, ... and show owl:sameAs pixelsize or whatever

<danbri> ...the things you use to declare mappings, will be the mechnisms we talk about in OEP

<danbri> jacco: re incubator, i think ralph was suggestiong that mainly for vocab work

<danbri> guus: we won't make a decision now; this is input for final discussions

<danbri> g: what was decision re structure?

<danbri> j: not deciding now

<danbri> g: we're plannign to end this deliverable on time?

<danbri> j: will re-visit at end of this f2f

<danbri> g: and discuss then the structure?

<danbri> j: do that offline, well, on the mailing list

<danbri> g=georges(?sorry)

<danbri> guus: ..[missed]

<danbri> ...depends on outcome of futures discussion tommrorow

<ChrisW> guus: woudl be useful to put a stake inthe ground re: interop note

<danbri> raphael: ...is there already an XG?

<danbri> danbri: none that i know of

<danbri> [example: there is some offlist discussion amongst rdf/mapping/geo folk, http://esw.w3.org/topic/GeoRDF and nearby has some related text... but not yet an XG proposal i think]

<danbri> jacco: [...] VRA Core

<danbri> ...specific things for images. A bias towards cultural heritage.

<danbri> ...Mark V has been working hard on a schema rdfs/owl for VRA

<danbri> (mark van assem)

<danbri> ...if we could have a vra schema w/ similar acceptance to Dublin Core's

<danbri> jacco: so q to the wg... is this an actviity suited to this wg?

<ChrisW> can we get the URL?

<ChrisW> is this document on line?

<Zakim> Ralph, you wanted to say that XG is definitely appropriate for the VRA vocabulary work and to remind us about "best practice"

<danbri> ...how do we move it forward?

<danbri> [...]

<danbri> ralph: VRA vocab... that kind of work is exactly what an incubator group is appropriate to do

<Raphael> This one Chris ? http://www.w3.org/2001/sw/BestPractices/MM/vra-conversion.html

<danbri> ...where you have some idedas you want to work out with a smaller group of peers, before proposing w3c recommendation track work

<danbri> ...if you're already at the stage where you think there's community consensnsu developing, and think you're ready ... then a separate wg should be proposed

<danbri> ...neither to me seem to match original expectations for this wg

<danbri> ...more about nominiating some current practices as best practices

<danbri> ..premature to consider that work as a best practice

<danbri> ...maybe for an XG, possibily for a WG

<danbri> guus: i have a number of doubts

<danbri> ...maybe other suggestions

<danbri> ...i wrote the first vra schema, as a test... was approached by Linda C by vra

<danbri> ...unsure about how it is being maintained

<danbri> ...not sure this is a problem

<danbri> ...in terms of content, it is exellently done. a nicely used 'dumb down' principle (from dublin core)

<danbri> ...but needs to be a community behind it

<danbri> ...also: this is very skos related

<danbri> ...has many of the issues...

<RalphS> [MM/ALL] Proposed f2f agenda addendum [Jacco 2005-10-28]

<danbri> ...closer to skos

<danbri> jacco: tools + vocab doc uses the schema

<danbri> chris: using vra is similar to skos?!

<danbri> guus: many issues that are in the same .... from same community

<danbri> vra = visual resource association

<danbri> ... vra core categories, v 3.0

<danbri> ...is owl Full

<danbri> ...specialises DC

<danbri> aldo: it's not a thesaurus...

<danbri> jacco: more like dublin core

<danbri> al: dublin core and skos are natural partners

<danbri> guus: yes, natural partners

<danbri> chris: [not same, just related ok]

<danbri> al: imporant for dc and skos to grow together

<danbri> ...to have them related sociallly is important

<Ben-scribe> Danbri: we talked earlier about bringing some work into the IG. The more patents they have, the more scared I am. This is comes from DC, so seems OK. I don’t car if its incubator, TF, IG, or even the DC group, so long as it’s discussed in public. Maybe DC would be a good choice.

<danbri> al: dcmi usage board moving towards endorsing external refinements

<danbri> [rather than trying to do everything monolithically]

<danbri> mike: those [dc] extensions... any concern about rubberstamping incompatible extensions?

<danbri> guus: built into dc, the dumb down rules, ... fact of life that people need the extensions

<Ben-scribe> Danbri: lots of discussions in the DC world. This should allow cleaner extensions.

<danbri_> [oof, network vanished briefly]

<dwood> 15:15-16:30

<dwood> RDFTM TF


<danbri_> guus: welcome valentina

<danbri_> [coffee until quarter-two]

<Benjamin> jjc: meeting for the xml datatypes after end of this meeting

<raphael> Jacco is scribing off-line ...

<raphael> jjc makes an announcement for a small working group before the dinner tonight on XML Schema datatypes

<Jacco> Jeremy: XML Schema datatypes break out session 45 minutes after the end of this meeting

<Benjamin> valentina: the document contains a survey of existing proposals

<Jacco> Andreas: Dinner starts at 20:00

<Jacco> Guus: tomorrow we start at 9:00, on SE and Schema datatypes, coffee break, moved OEP discussion to after lunch

<Jacco> Now: A+D, tutorial, vocab. management task force. SKOS-break out directly after coofee break, after lunch OEP

<Jacco> Now: A+D, tutorial, vocab. management task force. SKOS-break out directly after coofee break, after lunch OEP

<Benjamin> ... we have studied those proposals to identify the mapping issues to address, to define guidelines for rdf topic maps interoperability

<Jacco> DWood: possible for me to participate in the SKOS-break out?

<Jacco> DWood: possible for me to participate in the SKOS-break out?

<Jacco> Dan: no, is in parallel with SPARQL

<Benjamin> val: would like comments on structure of document


<Jacco> Valentina presenting slides, sorry I missed that

<Jacco> valentina is posting url of her slides...

<dbooth2> Valentina: (in response to question about auth challenge on the page) strange.

<Benjamin> Slide URL is : http://tesi.fabio.web.cs.unibo.it/RDFTM/RDFTMMappingGuidelines

<Jacco> intro with glossery of terms so both communties agree on this

<Jacco> 2nd section of draft is about requirements

<Jacco> today I will show you the informal description section

<Jacco> proposed solution on the data level of RDF/TM interoperability

<Jacco> - how to author RDF or TM documents with max operability

<Jacco> discuss goals and non goals, see slides

<Jacco> non goal: provide unified model

<Jacco> approach: define vocabulary for the mapping

<Jacco> + guidance

<Jacco> we have no addressed issues on guided translations

<Jacco> it possible to obtain a better translations if you have guidance

<Jacco> slide explaining LTM syntax

<Jacco> 1st block represents TM assertion, a is a topic with name "name"

<Jacco> in TM we have 2 types of IDs

<Jacco> in RDF you id by URI

<Jacco> in TM the URI can be a subject id, when the topic defines a information resource descring a subject

<Jacco> or it IDs an abstract concept

<Jacco> correction to the above: URI is either a subject locator or a subject identifier

<Jacco> topic also have variant names

<Jacco> within a certain scope s

<Jacco> last block is a binary association, p is the name of the association

<dbooth2> Informally, a "Subject locator" is for identifying a document directly; "Subject identifier" identifyies a document that describes a concept, and thus indirectly idenfies that concept.

<Jacco> valentina discusses examples of typical mappings

<dbooth> ScribeNick: aliman_scribe

<dbooth> Scribe: Alistair

danbri: pat hayes had comments on this point ... formalisation should happen soon ...

<RalphS> Pat Hayes' comments

see pat hayes email ...

<dbooth> Pat Hayes email: http://lists.w3.org/Archives/Public/public-swbp-wg/2005Oct/0216.html

val: if a resource is explicitly defined as type information resource, then we knoe the URI must become a subject locator ...

<danbri> [the point i think: the lack of a claim that something is an InfoResource doesn't allow us to assume it is a mamber of the class of non-INfoResources ]

pat said 'how can you distinguish between a resource that is not an information resource, and a resource that is not asserted to be not an information resource ...

scribe: val: but this is guided translation, and you should be explicit about what is and is not an information resource ...

and without guidance, assume not.

jjc: not negation, because talking about human input, so ok to make a closed world assumption re human input to guidance for translation

val: yes. If you don't have this particular guidance, you fall into the unguided translation ...

we don't deal with identity in the unguided translation, so this coudl change in the unguided translation ...

this point for unguiged transaltion is undecided as yet.

<dwood> I see in the slides (http://tesi.fabio.web.cs.unibo.it/twiki/pub/RDFTM/RDFTMMappingGuidelines/RDFTMGuidelines.pdf on Slide 5) that mapping between RDF and TM Models has been dropped. It was originally one of Steve's goals. Could someone please ask Valentina (when appropriate) why the requirement was dropped and what the issues are? Is this mapping theoretically possible? Do they intend to address it in later work?

al notes that angle brackets are missing for the long URLs in notation3 example

val: we define that iso:topic-name is a NamingProperty ...

if we had no subject locator, we would use one subject identifier chosen at random as the URI of the resource ...

danbri: what do you do with the others?

val: if you have more than one subject locator, one is URI, others are owl:sameAs
... when you have more than one subject identifier in the topic, one becomes URI for the resource ..

others become subject identifier properties.

Al; this is wrong

<RalphS> subjectIdentifiers are indirect, so it would be in appropriate to say that two subjectIdentifers are owl:sameAs -- rather the semantics are that the (identified) subject is owl:sameAs

val: in topic maps you are not pointing to the abstract subject, you have a resource that represents the document describing the subject ...

<RalphS> e.g. [ rfdtm:subjectIdentifier :a ] owl:sameAs [ rdftm:subjectIdentifier :b ]

danbri: sameas is ok ... you have the property subjectIdentifier turn one into URI ... ?

<RalphS> ^ implied by some relationship between :a and :b

val: talk about it later ...

jjc: in terms of document this is wrong, because it contradicts MUSTs, becomes non-deterministic

<RalphS> I think it's incorrect (re: slide 11) to say that a subjectIdentifier URI becomes the URI of the topic resource

val: identity is the most complex issue

<RalphS> ... a subjectIdentifier appears always to be an information resource that is (intended to be) interpreted by a human

jjc: if you choose the first, you get round tripping, if you choose at random, you don't

<danbri> [..ooOO ("could you get different graphs, but the same inferences out via the sameAs claims?") ]

jacco: is the order of the subject identifiers relevant?

<RalphS> [I can't follow the voices enough to know whether I should interject that comment or not]

val: we would like to discuss this, lars marius has the idea presented, i see all the points made here ...

<RalphS> yeah, there are details of these mappings that need further discussion but the f2f time probably can't get down to that level of detail

still an issue, would like to discuss further on the list and work out a solution.

<RalphS> perhaps f2f time is best used in discussing approach

guus: what is status of the task force?

<danbri> [jeremy ... do you think there's a precise definition possible for "deterministic" mapping? --- thinking of yr graph comparision stuff, and trickyness of comparing graphs w/ bnodes in them]

<dwood> Ralph, "If the topic has only subject identifiers, one of them becomes the URI of the corresponding resource" implies that this is a convenience for an implementation, not a theoretical decision.

<Zakim> David_Wood, you wanted to ask Valentina about the mapping of RDF to TM Models

<raphael> Al is still scribing but offline ... (connexion lost again!)

<bwm_> valentina: idea was to finish in December

<bwm_> ... but this was not possible

<bwm_> guus: this is a draft

<bwm_> ... we could have an editors draft

<bwm_> ... and it will take time to get wg consensus on publishing a working draft

<bwm_> ... my guess is this is a 3-4 month process to get to wd status

<bwm_> ... and not sure this is realistic

<bwm_> jjc: would it be better to take current wd to a note

<bwm_> guus: its not in a final state

<bwm_> ... we would like editors to say they think they are done having processed all comments

<bwm_> ... maybe all we should do is publish this survey

<bwm_> ... maybe that would not be all we wanted, but it is something

<bwm_> valentina: we have already addressed reviews on wd

<bwm_> jjc: so it would be acheivable to produce a note

<dwood> Are you discussing not attempting to publish http://tesi.fabio.web.cs.unibo.it/RDFTM/RDFTMInteroperabilityGuidelinesSteveDraft ?

<bwm_> ... the current doc is unlikely to be finishable in the time frame

<dwood> or a different document?

<bwm_> ... better to aim at something achievable

<bwm_> guus: my suggestion is

<bwm_> ... take current wd and bring it to note by end of charter of wg

<bwm_> ... have internal dicsussion with task force about what could be done beyond the charter

<bwm_> ... then we could look at options

<bwm_> ... the earlier you do this, the more time we have to discuss options

<bwm_> ... that would be a consensus of the tf as a whole

<bwm_> ... if the wg gets rechartered we are committed to ...

<bwm_> valentina: I will schedule a tf telecon to discuss this and post results to the list

<bwm_> guus: I'm looking for possible ways to continue

<bwm_> ... and as a minimal thing complete the note

<RalphS> I believe the "current draft" is http://tesi.fabio.web.cs.unibo.it/RDFTM/RDFTMInteroperabilityGuidelines as per the Agenda

<bwm_> ... and don't underestimate the time it will take to do that

<bwm_> ... first it has to be reviewed by the WG

<RalphS> (and per Valentina's msg of 27 Oct http://lists.w3.org/Archives/Public/public-swbp-wg/2005Oct/0215.html )

<bwm_> jjc: its realistic to do that in the time frame but its not trivial

<bwm_> guus: any more comments

<bwm_> jjc: I will send email on the doc we had to review

<bwm_> davidW: what are turning into a note?

<Zakim> David_Wood, you wanted to ask if the "current wd" under discussion is the Guidelines for RDF/TM Interoperability at

<bwm_> guus: the survey

<bwm_> ... because that's all we have time for

<bwm_> davidW: thats fine

<bwm_> guus: any more comments

<bwm_> danbri: I wanted to ask about reification but that takes too long

[scribing posting offline notes] ...

davidw: because too hard?

<RalphS> (Guus clarifies; the Survey WD is the document targetted at Note)

<bwm_> note: bwm's scribing is unofficial

val: there we many discussions re mapping at the level of the models ...

personally i think it's not possible ...

we haven't discussed in TF

val: next steps too address n-ary relations issues, and reification ...

n-ary relations mandatory to collaborate on a vocabulary (???) ...

we need to identify the best formalism for defining mapping rules ...

<RalphS> [thanks, Brian, informality acknowledged and very much appreciated]

writing down of the rule formally

and resolve the identity issue

guus: what is timeline?

val: no official timeline, discussing with tf members, also depends on the WG future

guus: not completed by 1st feb
... so what is required to finish this work? ... can then look for organisational strucutre to do this ...

val: idea was to finish december, but not possible ...

think we can have a draft for 1st feb ...

guus: this is an editor's draft, will take some time to get consensus in WG to get to official WD status ...
... this is an editor's draft, will take some time to get consensus in WG to get to official WD status ...

maybe several months ... not realistic

jjc: time better spent bringin current WD to working group note?
... publishing a final version of this document is solid achievement realistic in time
... current sketch of doc will take a lot of time to finish ...

better to aim at acheivable goal.

guus: suggest take the current WD to Note. Have a discussion within TF to work out possible work to do beyond current charter ...

then if we have specific proposal, then we can look at options to satisfy ...

earlier you do this, earlier we can adapt WG planning ...

but this needs consensus of TF as a whole, to say that we are committed to continuing this work.

val: will schedule next telecon and decide these things, then post to list.

guus: expect proposal for how to continue, and as minimal thing to produce this Note ...

jjc: realistic to do within timeframe, but not trivial!

davidw: is uni bologna wiki draft the thing we're going to make into a note?

guus: no only the survey.

<danbri> http://tesi.fabio.web.cs.unibo.it/twiki/pub/RDFTM/RDFTMMappingGuidelines/RDFTMGuidelines.pdf <- accessible now

<benadida> and now, Ralph? :)

<libby_> ralph: http://www.flickr.com/photos/nicecupoftea/59738682/

<libby_> (not very exciting)

<libby_> http://swordfish.rdfweb.org/photos/2005/11/04/ <- photos from earlier

<bwm_scribe> ben: the tf has been meeting regularly making progress on rdf/a

<bwm_scribe> ... sent two docs that represent the latest rdf/a

<benadida> http://www.w3.org/2001/sw/BestPractices/HTML/2005-rdfa-primer

<bwm_scribe> ... the first one summarises the issues

<bwm_scribe> ... reflects our current thinking

<bwm_scribe> ... there is a big issue

<bwm_scribe> ... the curie issue


<bwm_scribe> ... the rest is fairly uncontroversial

<bwm_scribe> ... i can go through and do some examples

<bwm_scribe> guus: we will try to project now

<bwm_scribe> ... can you indicate the main changes

<bwm_scribe> ... after boston discussion

<bwm_scribe> ben: the main development since then

<bwm_scribe> ... is the simplificcation of the inheritance rules

<bwm_scribe> ... originally we tried to represent arbitrary rdf graphs which generated a lot of triples

<bwm_scribe> guus: we now have the doc on screen

<bwm_scribe> ben: with simplified inheritance of attributes

<bwm_scribe> ... but cannot represent arbitrary rdf graph

<bwm_scribe> ... but greatly simplifies things

<bwm_scribe> ... guus should I do a few important examples

<bwm_scribe> guus: a few examples and then questions

<bwm_scribe> ben: look at 2.1

<bwm_scribe> ... using prop attributes to designate rdf triple

<bwm_scribe> ... nothing that is surprising

<bwm_scribe> ... once something is an object, it cannot be retasked as a subject

<bwm_scribe> ... have to repeat the uri to do that

<bwm_scribe> ... 2.2 shows qualified links using rel attribute

<bwm_scribe> ... can be used on any xhtml element

<bwm_scribe> ... example is creative commons license

<bwm_scribe> ... the rel can be interpretted as a qname, but I'll get back to that

<Zakim> jeremy, you wanted to mention xml:lang issue and to note XMLLiteral error in examples from primer

<bwm_scribe> jjc: back in 2.1 should those not be xml literals rather than plain literals

<bwm_scribe> ben: yes - thats an error

<bwm_scribe> danbri: can you create plain literals

<benadida> http://www.w3.org/2001/sw/BestPractices/HTML/2005-rdfa-syntax

<bwm_scribe> ben: yes - we don't discuss that in the primer

<bwm_scribe> ... see full spec

<bwm_scribe> ... something about datatype plain - missed

<bwm_scribe> ... the syntax doc describes how to handle literals (uri above)

<danbri> todo for danbri: check range of foaf:name, if it allows xml literal

<bwm_scribe> ... we are working on how to make a literal by concatenating the child element content

<bwm_scribe> ... section 3 introduces about attribute

<bwm_scribe> ... allows to make statement about subjects other than current document

<RalphS> 3.1 Qualifying Other Documents

<bwm_scribe> ... e.g. about a photogrpah in the document

<bwm_scribe> ... in 3.2 we see the qualification of junks of documents

<bwm_scribe> ... give an xml id to an element

<bwm_scribe> ... can use link and meta attributes to make statements about chunks

<bwm_scribe> ... in 3.2 first example box, section id = person, link element off of that

<bwm_scribe> ... there has been talk of syntactic sugar for rdf:type but they are not resolved yet

<bwm_scribe> ... possibly using the class attribute

<bwm_scribe> danbri: <section id="person">

<bwm_scribe> ... is that about the document or the person

<bwm_scribe> ... the claims are about the person not the document

<bwm_scribe> ben: that's ok -

<jeremy> http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2005Sep/0020

<bwm_scribe> danbri: that will be a big conceptual leap for xhtml people

<jeremy> (answers dan questions)

<bwm_scribe> ben: the point is to be able to support difference between person and their home page

<RalphS> [I think this example is probably a bad idea -- not good practice]

<bwm_scribe> ... we can designate a chunk of the document to represent the person

<bwm_scribe> al: it would be much safer not to go there

<bwm_scribe> ... assume frag's within documents denote sections of documents

<bwm_scribe> ben: I'm confused about the risk here

<bwm_scribe> jjc: this question is answered in rdf concepts document

<bwm_scribe> ... the frag id is interpretted as an rdf frag id and can denote the person

<dwood> So how does one attach a property within a section to the base document?

<benadida> dwood: use the about="" attribute to override the subject

<bwm_scribe> ... it is interpretted in an rdf graph with respect to rdf mime type

<bwm_scribe> al: what happens when you want to annotate the document

<bwm_scribe> al: what happens when you want to make statements about the section of the document

<RalphS> yes, this confusion between fragment of a document vs. Person is real and will continue

<bwm_scribe> jjc: it gets difficult

<bwm_scribe> al: then you use one uri to denote two things

<bwm_scribe> danbri: i said this would be rathole

<bwm_scribe> al: that's why I say don't go there

<bwm_scribe> danbri: can you do blank nodes

<bwm_scribe> ben: yes

<bwm_scribe> ... lets take this point offline - I didn't think we had any controversy yet

<RalphS> Alistair suggests we could choose a different example and I suspect he's correct

<bwm_scribe> ACTION: ben to contact alistair on use of frag id's [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action02]

<bwm_scribe> - action should include david booth and danbri also

<bwm_scribe> ben: section 4.1 bnodes is a natural progression what we said

<RalphS> [I suggest that as there are at least 4 people who want to discuss this that "off-line" mean "by mail, copying the WG list"]

<bwm_scribe> ... if you don't have an id on a fragment then you get a bnode

<bwm_scribe> ... using link rel

<bwm_scribe> ... see example

<bwm_scribe> ... if you want to give the bnode a name you can

<bwm_scribe> ... see second example

<bwm_scribe> ... ignore the square brackets for now

<bwm_scribe> ... you can relate bnodes using the about attribute to designate what bnode you are referring to

<bwm_scribe> danbri: what is the scope of the bnode name

<bwm_scribe> ben: the whole document

<danbri> example: if we had some wrapping document eg CDF 2.0 that had sub-islands of rdf/a ...

<bwm_scribe> davidB: in the generated rdf, what is the current document? The entire document?

<bwm_scribe> ben: I'm confused - what do you mean by conversion

<danbri> would those sub-islands of rdf/a (each with their own <head>, arguably?

<danbri> ...have bnode IDs in the same space

<bwm_scribe> davidB: in the example you show the xthml and the resulting RDF

<bwm_scribe> ... the resulting RDf refers to the current document

<bwm_scribe> ben: its the current uri so the xhtml document

<bwm_scribe> ... section 4.2 has limited support for reification

<bwm_scribe> ... in some useful cases

<bwm_scribe> ... but I'd like to skip this

<bwm_scribe> guus: should you mention reification in the primer

<bwm_scribe> ben: good point

<bwm_scribe> ... we should make this more of a primer

<bwm_scribe> danbri: I'm going to propose you drop reification

<bwm_scribe> jjc: I tried to make that point

<bwm_scribe> ben: I urge you to check out the creative commons and foaf examples

<bwm_scribe> ... the big issue is the compact uri issue

<benadida> http://www.w3.org/2001/sw/BestPractices/HTML/2005-10-27-CURIE

<bwm_scribe> ... a large part of which is my fault

<bwm_scribe> ... there was an issue iwth requirements

<RalphS> I note that Creative Commons has a use case for which they think reification helps, so I challenge DanBri and Jeremy to (continue to) help in showing how to express the Creative Commons case (approximately 'who wrote this license')

<bwm_scribe> back

<bwm_scribe> scribing interrupted by network failure, continueing offline

<bwm_scribe> ben: at the end of section 2 in the primer there is a good example of a curie

<bwm_scribe> ... second issue is how can we have subjects and objects that can be uri's or curies

<bwm_scribe> ... so we introduced square bracket notation

<bwm_scribe> ... we considered a lot of different options

<bwm_scribe> ... e.g. different attributes names

<bwm_scribe> ... but this sames cleanest

<bwm_scribe> ... importantly it gives us a solution ot the b-node problem

<bwm_scribe> ... I will put together a doc summarizing our discussions

<bwm_scribe> guus: I have a question about curies

<bwm_scribe> ben: yes

<bwm_scribe> guus: this could create a timing problem

<bwm_scribe> jjc: I beleive this issue is bang on charter for this task force

<bwm_scribe> ... the length of uri's is a significant barrier to the deployment of semantic web technology

<bwm_scribe> ... if you look at microformats or wiki

<bwm_scribe> jjc: its knowledge meeting the web

<bwm_scribe> guus: he was supporting ben's point about the need for curies

<bwm_scribe> danbri: consistency is important

<bwm_scribe> ... can rel and href take the same kind of values

<bwm_scribe> ben: good question we are discussing that now

<bwm_scribe> ... [missed]

<RalphS> Ben: we're trying for backward compatibility for some HTML syntax; e.g. rel="next"

<bwm_scribe> ... if you take html from a prior version then rel etc should still work on xhtml 2

<bwm_scribe> guus: i'm going to give this discussion max 5 mins

<bwm_scribe> jjc: rel="next" is an unqualified name i.e. from xhtml 2 namespace

<bwm_scribe> danbri: you mentioned no follow in passing as if that has now been accepted into the namespace

<bwm_scribe> ben: we have not added no follow into the namespace - bad example

<bwm_scribe> jacco: current tools will not parse this correctly

<bwm_scribe> ben: yes - we are looking at that

<bwm_scribe> guus: lets have tech discussion on the list and focus now on planning

<bwm_scribe> ... what are the task force's plans

<Zakim> Ralph, you wanted to also ask for a sense of the WG about RDF/A as a separate module

<Zakim> RalphS, you wanted to see if there is a sense of the WG resolution regarding the utility of CURIEs

<bwm_scribe> ralph: there are two important things the tf needs the wg input on

<bwm_scribe> ... unlike xml namespaces, there are other ways to abbreviate uris

<bwm_scribe> ... it is largely an aesthetic argument being made by those who express concern about curies

<bwm_scribe> ... we have seen a lot of pushback on the curie idea

<bwm_scribe> ... before we adjourn f2f, please can we have the sense of the wg on whether we agree that sw technology requires not typing so much in docs

<bwm_scribe> ... second thing relate to question about parsers

<bwm_scribe> ... rdf/a is being done in context of xhtml 2

<bwm_scribe> ... which is sufficiently different to xhtml 1 that tools have to change

<bwm_scribe> ... there is a desire to use rdf/a with other kinds of xml docs

<bwm_scribe> ... how interested is the wg in making this a module that can be used with say svg or xhtml 1

<bwm_scribe> guus: i'm glad that the tf has made so much progress recently

<bwm_scribe> ... it gives wg as a whole a problem

<bwm_scribe> ... we need time to reflect and think about these issues

<bwm_scribe> ... not comfortable getting an opinion now

<bwm_scribe> ralph: that's why i said before end of meeting

<bwm_scribe> guus: like to hear ben's opinion on what can be done by 1/Feb

<RalphS> (i.e. end of meeting tomorrow)

<bwm_scribe> ben: is 1/Feb when we should wrap up by

<bwm_scribe> guus: from a minimialis tpoint of view - yes

<RalphS> we also have a deadline from the HTML WG charter

<bwm_scribe> ben: the primer is our main goal in the next couple of weeks

<jeremy> (further point might be #7 in http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2005Oct/0057 )

<dwood> Doesn't the RDF/A primer rely on getting the syntax doc completed?

<bwm_scribe> ... then do xslt to extract triples

<bwm_scribe> ... we hope to have these docs in business shape by 1/Feb

<bwm_scribe> ... we need to discuss their status in terms of wg notes

<bwm_scribe> guus: I suggest your propose a schedule to bring to wd including review by wg

<bwm_scribe> ... need editors draft before Christmas, preferably earlier

<bwm_scribe> ben: i think we can do that

<bwm_scribe> jjc: this tf is also a tf of the xhtml wg

<bwm_scribe> ben: if it becomes a note - there is the question of where it lives afterwards

<bwm_scribe> guus: that all has to be discussed

<bwm_scribe> ... I would appreciate an output before the end of the charter

<RalphS> We've been viewing our primary deployment path as via the HTML WG

<bwm_scribe> ... and the primer is a good candidate

<bwm_scribe> ... I'd like to see how to mark up your home page

<bwm_scribe> ... because that was the original goal

<RalphS> that's why I asked for a sense of the WG resolution on whether that should remain our primary deployment objective

<bwm_scribe> raph: i appreciate where guus is going

<bwm_scribe> ... this tf has viewed its mission as helping html wg find a solution to include in xhtml 2 specification

<bwm_scribe> ... may be that xhtml wg has a tighter deadline

<bwm_scribe> ... they are supposed to be in last call 9 months ago

<bwm_scribe> ... if our primary deployment task is through xhtml 2

<bwm_scribe> ... we must be part of that

<bwm_scribe> ... if the sense is that a separate module is the way to go we are less tied to their schedule

<bwm_scribe> ... but then this wg would have more to do

<bwm_scribe> ... and we may not be able to do that

<bwm_scribe> jjc: tight coupling means that eventually the payback is huge

<bwm_scribe> ... doesn't matter if its not done by the end of this wg

<bwm_scribe> guus: do you have plans for continuation of this wg

<bwm_scribe> ... if bp does not go ahead, does it continue in xhtml

<bwm_scribe> ralph: there is a lot of pressure to make progress before the end of January

<bwm_scribe> jjc: there is another issue

<bwm_scribe> ... that xml lang gets lost

<bwm_scribe> guus: lets note that

<jeremy> (further point might be #7 in http://lists.w3.org/Archives/Public/public-rdf-in-xhtml-tf/2005Oct/0057 )

<jeremy> I think that will be a big issue ....

<bwm_scribe> davidW: it seems odd to publish the primer without the spec

<RalphS> If the RDF-in-XHTML work remains in its current state much longer there are many people who will be severely criticized

<bwm_scribe> ... can we get the syntax doc done

<danbri> [fwiw I sent mail re rdf/a and reification; jeremy if you could take a look i'd be v grateful]

<bwm_scribe> ben: I think so - the syntax doc is nearer done

<bwm_scribe> davidW: lets get it up for review then

<bwm_scribe> ben: ok

<bwm_scribe> guus: there are issues to resolve

<bwm_scribe> ben: yes

<bwm_scribe> guus: I'd like to see a schedule

<bwm_scribe> ACTION: ben to produce schedule for getting docs done [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action03]

<bwm_scribe> jjc: we could ask tf to publish what we have now for wider review

<RalphS> ACTION: Ben produce schedule for getting RDF/A editor's drafts docs ready for WG review [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action04]

<bwm_scribe> guus: you might want to consider publishing what we have if resolving issues will take a while

<FabGandon> For records info on CURIE: http://www.w3.org/2001/sw/BestPractices/HTML/2005-10-21-curie

<ChrisW> NEW OEP EDITOR'S DRAFT: http://www.w3.org/2001/sw/BestPractices/OEP/SemInt/

<ChrisW> for discussion tomorrow

<ChrisW> (I don't have email access)

<bwm_scribe> guus: its nearly 6pm

<bwm_scribe> ... we started 30 mins late and finish 30 mins late

<bwm_scribe> ... thank you ben

<bwm_scribe> ... thanks david and ralph for joining us

<bwm_scribe> ... there is an updated schedule for tomorrow on the list

<RalphS> [happy to finally know how to pronounce "quay".ir

<bwm_scribe> Andreas: dinner is in Quay street, near spanish arch K C Blakes

<RalphS> ... as "key" :) ]

<bwm_scribe> ADJOURNED

Summary of Action Items

[NEW] ACTION: Ben produce schedule for getting RDF/A editor's drafts docs ready for WG review [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action04]
[NEW] ACTION: ben to contact alistair on use of frag id's [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action02]
[NEW] ACTION: Ralph to copy the WordNet to CVS space and post URL [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action01]
[DONE] ACTION: ben to produce schedule for getting docs [recorded in http://www.w3.org/2005/11/04-swbp-minutes.html#action03]
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2005/11/04 17:59:40 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.127  of Date: 2005/08/16 15:12:03  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/:/,/
FAILED: s/jacov/jacco/
Succeeded: s/Rafael/Raphael/
Succeeded: s/dwood:/dwood,/
Succeeded: s/note/note in the SOTD/
Succeeded: s/holding/evolving/
Succeeded: s/jjc :/jjc:/g
Succeeded: s/mike :/mike:/g
Succeeded: s/ :/:/g
Succeeded: s/ibby/Libby/
Succeeded: s/david/dbooth/
Succeeded: s/coching/couching/
Succeeded: s/Ralph:/DWood:/
Succeeded: s/Ralph: poss/DWood: poss/
Succeeded: s|Slides: http://tesi.fabio.web.cs.unibo.it/RDFTM/TDFTMMappingGuidelines||
Succeeded: s/(if I typed the URL correctly)//
Succeeded: s/same here ...//
Succeeded: s/correc /Slide /
Succeeded: s/Subject indicator/Subject identifier/
Succeeded: s/section-id = person/<section id="person">/
Succeeded: s/yes/ the whole document/
Succeeded: s/know/now/
Succeeded: s/by end/before the end/
Succeeded: s/action ben/ACTION: ben to/
Found Scribe: ChrisW
Found ScribeNick: ChrisW-scribe
Found Scribe: dbooth
Inferring ScribeNick: dbooth
Found Scribe: FabGandon_
Inferring ScribeNick: FabGandon_
Found ScribeNick: aliman_scribe
Found Scribe: Alistair
Scribes: ChrisW, dbooth, FabGandon_, Alistair
ScribeNicks: ChrisW-scribe, dbooth, FabGandon_, aliman_scribe
Default Present: Ralph, David_Wood, Guus, Brian, Jeremy, ChrisW, Alistair, DanBri, Andreas, Libby, Ben, Fabien, Jacco, Raphael, Giorgos_Stamou
Present: AListair Danbri BrianMcBride
Agenda: http://lists.w3.org/Archives/Public/public-swbp-wg/2005Nov/0018.html

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Got date from IRC log name: 4 Nov 2005
Guessing minutes URL: http://www.w3.org/2005/11/04-swbp-minutes.html
People with action items: ben ralph

[End of scribe.perl diagnostic output]