meeting record: 2004-04-01 SemWeb BPD WG telecon

Here are minutes of our 1 April meeting, with two corrections from
the irc record.  In irc I recorded that

   ACTION DanBri get rdf-thes contacts from Guus

was completed, however Brian correctly noted that this action is
not complete until Guus provides the data.  I have thus changed
this to 'continued'.

Also, Mike Uschold noted that I mis-recorded his comments, which
were intended to be:

  There was NO agreement on metaclasses or on any other specific
  issues. The agreement was on how to present our advice, [if we ever
  reach agreement]. Specifically: being non-judgemental in our advice,
  to emphasize consequences of modeling choices, and to avoid preaching.




                    SemWeb Best Practices and Deployment WG
                                  1 Apr 2004

   See also: [2]IRC log


          Guus Schreiber, Ralph Swick, Pat Hayes, Dan Brickley, Libby Miller,
          Jeremy Carroll, Marco Nanni, Natasha Noy, Brian McBride,
          Jacco van Ossenbruggen, Aditya Kalyanpur, Deb McGuinness,
          Tom Baker, David Norheim, Jos De Roo, Alistair Miles,
          Jen Golbeck, Mike Uschold

          Jim Hendler, Bernard Vatant, Aldo Gangemi (Aldo watched
          on irc, however, and commented on what was recorded therein.)




     * Action Item Review
     * Summary of New Action Items
     * Summary of resolutions
     * Agenda
         1. Admin
         2. Task Force Template
         3. Application and Demo TF
         4. World TF
         5. WordNet Task Force
         6. OPEN TF


Action Item Review:

   ACTION DanBri ask SemWeb CG about constraints and advice re WD, Note, etc.
     -- completed in

   ACTION Ralph talk with Guus and create a straw poll form for meeting times
     -- continued; to be done in next 7 days

   ACTION Guus to propose a format for TF description
     -- completed in [15]; on today's agenda

   ACTION JimH write description of WorldView TF
     -- completed in [17]; on today's agenda

   ACTION ChrisW write description of OPEN TF
     -- continued

   ACTION Aldo describe Wordnet TF
     -- completed in [18]; on today's agenda

   ACTION DanBri get rdf-thes contacts from Guus
     -- continued, though DanBri has started a page

   ACTION Aldo e-mail WordNet update to mailing list
     -- completed; see

   ACTION guus read bernard's e-mail with draft TMAP msg
     -- continued

   ACTION guus to finalize TMAP e-mail, notify cg, and send out msg
    -- continued, with the following discussion:

     Ralph; the CG asked that the message include explicit language
     that this was not yet a formal request for liaison

     Jeremy: is it possible to have a representative from the ISO TC on our
     Task Force without it being a formal liaison?

     Ralph: my opinion is that yes, it is possible for the Task Force to
     include ISO TC members without it being a formal liaison

   ACTION DanC to send mail requesting feedback on RDF/XHTML & GRDDL
   to both lists
    -- completed; see

   ACTION DanBri investigate publication of Turtle as SWIG note
    -- continued

   ACTION ChrisW present his time ontology in the next telecon
    -- continued

Summary of New Action Items:

   ACTION Guus ask Aldo if someone from Princeton is willing to participate
    -- recorded in

   ACTION Guus talk with Aldo about WordNet and produce a short-term plan
   for option 1
    -- recorded in

   ACTION Natasha look at classes-as-values DC:Subject issue
    -- recorded in

   ACTION Deb propose a simple pattern for OEP
    -- recorded in

   ACTION Jeremy write TF desc for XML Schema
    -- recorded in

Summary of resolutions:

   RESOLVED Minutes of the 18 March telecon [13] accepted with correction
   that Dan Brickley's action re: consulting SemWeb CG about Notes vs WD
   was completed.
    -- recorded in

   RESOLVED Date & time for next telecon 15 April 1400 UTC.
    -- recorded in

   RESOLVED Template for Task Force descriptions [15] accepted.
    -- recorded in




   <Scribe> PROPOSED date & time for next telecon 15 April 1400 UTC

   Pat: this time clashes with Data Access WG

   Jeremy: Japan is under-represented on the call now

   <Scribe> Ralph will get working on his action to get straw poll done

   <Scribe> RESOLVED date & time for next telecon 15 April 1400 UTC

Task Force Template

    draft task-force template [Guus 2004-03-31]

   DanBri: This template looks good.  I have already started to wrap
   this in HTML for the WordNet TF.

   <Scribe> RESOLVED: template accepted

Application and Demo TF

   On a possible role for the applications and demos task force
   [Brian, 2004-03-25]

   Brian: my objective was to start some discussion
   hoped this TF might act like previous Test Cases task forces

   Jos: test cases have my support

   Guus: reaction from Knowledge Web perspective?

   Marco: too early to say; need a more clear definition before
   starting work; there are many topics from which to select

   Libby: I support the idea of showing code fragments

   Guus: would FOAF fit here?

   Libby: yes, possibly

   Pat: the analogy with code is misleading; code does what it does,
   whereas ontologies do more

   Brian: important to have other task forces should sign up to provide
   illustrations while I floated the idea, I'm not yet signing up to
   coordinate the TF

   David: also support this, but also will have to limit my participation

   Jeremy: I am happy to sign up for XML Schema TF

   Guus: I expect to sign up for Porting TF

   Libby: I am willing will to help coordinate ADTF for a while

   Brian: willing to help Libby flesh out a proposal

World TF

   "World" TF description (informal) [JimH, 2004-03-25]

   Guus: two Notes; the first Note is intended to be more business cases
   and the second Note is intended to help those who are confused in
   where to start

   Pat: I expect to be involved in this task force. I worry that the first
   Note might have too much "drum beating" and not enough "practice";
   that it might turn into an advertisement

   Guus: expect the document to be short; a maximum of 5 pages

   Mike: read Jim's summary, looks OK to me

   Deb: I also volunteered to Jim that I would help on this

   DavidN: I would especially like to help with the first document

   Pat: might be good to have more representation from others in the WG
   who have different views

   Jeremy: document 2 needs participation from the DL community

   Mike: I think I can represent the business perspective

   Deb: I consult with some startup companies and can represent the
   startup perspective

   David: we know something about the business perspective also

   Jos: happy to review, but won't be able to participate in this task

   Jeremy: will also review

   Marco: France Telecom would like [to provide] feedback.  We are
   willing to try out a draft document within FT

WordNet Task Force

   WordNet Task Force - work outline [Also 2004-03-23]

   Guus: see Aldo's recent msg
   Aldo presents an important option 2 choices of representation (1) take
   the principal representation as-is and create an RDF Schema for it,
   without adding features (2) do a translation and introduce some RDF

   <AldoG> Above all, (1) does not try to "interpret" the original
   Wordnet schema in terms of ontologies.  (2) tries a mapping from
   the original schema to ontology data types (classes, individuals,
   properties, subClass, etc.)

   DanBri: my translation only took the nouns; this allows markup that
   looks nice

   Guus: my personal opinion is that in principal if there is a
   representation we should translate it as literally as possible and
   leave interpretation to the user

   Pat: this issue is broader than just WordNet

   <aldoG> I agree with Pat: it applies also to thesauri

   Pat: agree with a translation that takes everything as an individual
   and makes the fewest assumptions

   <aldoG> the problem with (1) is that we actually renounce to derive
   useful ontologies from wordnets and thesauri, which is the current
   trend of research in both academic and industrial/organization

   Libby: we followed the simple approach in calendaring

   Ralph: agree with starting simple.
   Can we get participation from Princeton?

   <aldoG> (2) is not necessarily "complex". For instance, Dan Brickley
   has ported (a part of) WordNet using (2)

   ACTION: Guus ask Aldo if someone from Princeton is willing to

   <danbri> From [20] (just
   drafted), "Document strategies for representing Thesaurus-like content
   using RDF/OWL --- Produce guidelines for transforming an existing
   thesaurus (or classification system, or similar concept-based
   taxonomy) into an RDF/OWL representation. Guidelines should describe
   strategies for converting into an RDF representation of thesaurus-like
   structures, as well as strategies for re-describing in RDF/OWL the
   content originally conveyed in the Thesaurus."

   Pat: others have done major work on aligning Cyc with WordNet too

   <aldoG> Christiane Fellbaum is interested in following the work, but
   she cannot ensure a lot of attention

   Guus: I don't agree with Aldo's "the problem with (1)..." -- these can
   be added later

   <aldoG> CYC2WN, as well as Sensus, SUMO2WN and DOLCE2WN (OntoWordNet),
   follow approach (2).  In fact in [21]
   I propose to deliver (1), which is almost ready, and then move to (2)

   Pat: re Aldo's (2), there are tradeoffs

   Jeremy: it is plausible to do this in two phases

   Guus: re Christiane Fellbaum; if she could nod agreement that would
   be nice and even make the TF work part of their distribution

   <aldoG> Yes, I agreed with Jeremy (I have even proposed 3 phases,
   since (2) can be done in a 'simple', or in an "enriched" mode)

   Brian: a key test of commitment from the WordNet folk is maintenance
   as WordNet changes

   Guus: this was part of my thoughts; make it part of their distribution

   <aldoG> Versioning is a major issue for every ontology production
   lifecycle ... we should also think on procedures to easily port new
   versions of wordnets and thesauri. Indeed, this issue has been very
   relevant in the fishery ontology project with UN-FAO ...

   <danbri> [the class-based approach has merit where you want to mix
   wordnet w/ other vocab, eg <Pig x:name="Porky"/>...]

   Jeremy: I don't hear any dissent on the first (simple) proposal

   ACTION: Guus talk with Aldo about WordNet and produce a short-term
   plan for option 1

   <aldoG> I will include the proposal to make the OWL port part of the
   distribution in the open letter to WN developers, also mentioning the

   DanBri: We made a namespace for WordNet version 1.6; it would
   be simple to make another namespace for subsequent versions.
   My original notion was to conflate all the versions in one
   namespace, but it seems that won't fly. I would like to see
   guidelines for converting thesaurii to RDF structures

   <aldoG> Only one namespace could be unacceptable for lexicographic

   DanBri: WordNet could be examples in such a note

   <aldoG> Yes, Dan, (2) is a way of arriving to a common methodology to
   port terminological resources to SW

   Guus: most of the translations that have been done are not complete.
   I propose that any translation we do should be complete; e.g. only
   noun trees and only a few relationships there are about 20
   relationships.  We should do all the relationships under option 1

   Jeremy: that should be a goal, but don't overconstrain those doing the
   work in the TF

   Guus: ok, a goal -- not a requirement

   <aldoG> In (1) it is very easy to do all the relationships. In (2),
   we'll need to distinguish among rels holding classes, individuals,

   <danbri> [aldoG, did you see the SKOS vocab aliman worked on? do you
   think that could be extended for use in describing w/net?]

   <aldoG> [yes, Dan, I heard about SKOS, I'll go through it asap]


   MikeU: There has been lots of discussion in the past few weeks.
   Agreements have not been reached on metaclasses, or on any other
   specific issues such as when to use classes as instances.
   Mixing part-of and subclass relationship also had some discussion.
   Consensus is emerging on how to present our advice; to avoid
   judgemental calls but to identify representation/modelling choices
   and explain consequences of the choices, without saying "bad" or
   "good".  In cases where we can agree that an outcome is probably
   a desirable one or not a desirable one we can say so.

   MikeU: Next steps: agree to pick a single thing to try the
   non-judgemental approach, pick a candidate for a best practice
   move forward on some simple low-hanging fruit

   Guus: would be good if this TF can produce a short note from a use
   case perspective

   MikeU: Deb wrote a paper [Living with CLASSIC] a few years ago

   Deb: section called "tricks of the trade";
   when do you model something as a concept or as an individual, when to
   use classes as instances.  Classic pointed out positive and negative
   implications of some choices.  The TF will propose a format for
   interchange; will need several iterations. I will look over
   Living With Classic paper to see what can be lifted.  We are not
   likely to attempt metaclasses to start

   Natasha: classes for DC:Subject have been discussed -- this is
   actually a technical problem propose to describe tradeoffs rather
   than say 'good' vs 'bad'

   Deb: this doesn't sound like a short one to me

   Guus: anyone who has tried to combine Dublin Core and OWL has been
   faced with this problem

   <aldoG> maybe we could also try to give some idea of alternative
   general choices concerning categories of content, like my LAB has done
   in the EU WonderWeb project [22],
   specially (public) deliverable 18

   Guus: I support this (Dublin Core) topic as a note

   MikeU: we agree it is useful, concern about how long it might take

   Natasha: do 2 topics, one simple

   <jacco> deb: sounds like the GOF design pattern template ...

   Deb: a template for explaining tradeoffs

   <aldoG> obviously, alternatives of content should be given even within
   the Living with Classic case, or wrt Dublin Core, anyway very simple

   Deb: a format that would help point out consensus quickly will help

   Mike: useful for all TFs to look beneath the surface of email and
   notice that there often is a kernel of agreement

   ACTION: Natasha look at classes-as-values DC:Subject issue

   <aldoG> Having a GOF for trade-offs would be great

   Natasha: note that classes-as-instances is a different issue

   <aldoG> Natasha, do you mean classes as concrete data types?
   therefore, the 'annotation' approach ...

   Deb: The Task Force would like to change from OPEN to OEP;
   Ontology Engineering Patterns.  This name change will facilitate search

   <Scribe> ACTION: Deb propose a simple pattern for OEP

   Guus: I only put on the agenda those TFs that were getting active
   discussion from the Cannes notes, another TF I'd like to see get
   attention is PORT - Porting Thesaurii to the Web

   Alistair: I am working on thesaurii conversions right now

   Guus: I am working on this as well, expect to involve someone else
   from my lab

   <golbeck> i would like to be involved in port as well

   DanBri: I posted a draft today

   <Guus> DanBri Alistair this paper might be of interest:

   <aliman> public-esw-thes has Stella Dextre-Clarke and Leonard Will
   from a team working on new british standards for thesauri, also
   doug tudhope from glamorgan + some others

   DanBri: would the WG prefer to use an existing public-esw list or
   start a new one?

   Guus: if a separate list, would like to have regular summaries sent to
   the WG list

   DanBri: agree

   Guus: What about the XML Schema TF?

   Jeremy: not urgent, though New York is an opportunity to have informal

   <jjc> ACTION: jjc, write TF desc for XML Sch

Summary of Action Items

   [NEW] ACTION: Deb propose a simple pattern for OEP
   [NEW] ACTION: Guus ask Aldo if someone from Princeton is willing to
   [NEW] ACTION: Guus talk with Aldo about WordNet and produce a
     plan for option 1
   [NEW] ACTION: jjc, write TF desc for XML Sch
   [NEW] ACTION: Natasha look at classes-as-values DC:Subject issue

    Minutes formatted by David Booth's [24]scribe.perl 1.67 ([25]CVS
    $Date: 2004/03/26 19:59:12 $



Received on Tuesday, 6 April 2004 12:30:16 UTC