Skip to topic | Skip to bottom
Home

RDFTM
RDFTM.ConCall20060221r1.5 - 21 Feb 2006 - 12:08 - StevePeppertopic end

Start of topic | Skip to actions

Agenda and Minutes of 2006-02-21 meeting

TIME: 11:30 UTC+01.00 (Western European time)

DOCUMENTS:

AGENDA:

1. Review action points from 2006-01-27:

  • LMG: Provide email comments on VP's Draft of the formalization Before reviewing proposals, those who wish to propose a formalization should provide examples on the Wiki relating to the following mappings described: topic names, binary associations, scope. Deadline March 3rd.
  • SP: edit complete document
  • VP: write prose description of alternative solution to Identity issue. This must be proposed by March 3rd.
  • SP: collect vocabulary for n-ary associations (including talking to Chris Welty) and also look at how it can be applied to unary associations
  • VP/SP: start general discussion on unary associations
  • VP: send email of reification example that causes problems. We need this by March 3rd.
  • SP: check with Ralph on status of Survey as WG Note

2. Discuss issues raised in the draft:

  • (Status) Note or Recommendation? On the advice of Ralph Swick we go for a Note in the first instance (because the SWBPD WG will be winding up on May 1st). A Recommendation might be possible after that, via an Incubator Activity. Document should state that ultimate goal is Recommendation.
  • (1.2) Do we need to state that it is NOT a goal "to provide the mapping between the RDF and Topic Maps models"? It needs to be made clear that we are not creating mappings at the model level. FV will provide additional text to clarify this.
  • (3.4) Is the description of rdfs:label in OWL correct? YES; at least, it's close enough for now.
  • (3.4) Is there agreement on having built-in guidance for rdfs:label? YES
  • (3.4.1) Is it correct to introduce a new property rdftm:variant-scope? NO, the approach is flawed because it doesn't cater for multiple variants. We need a separate blank node for each variant. STP to rewrite.
  • (3.6.1) Is it OK not to support multiple signatures? YES
  • (3.6.1.1) Should it be possible to overrule the assumed equivalence between rdf:type and tm:type-instance? We can't prevent it, but don't mention it specifically.
  • (3.6.1.2) Should it be possible to overrule the assumed equivalence between rdfs:subClassOf and tm:supertype-subtype? ditto
  • (3.7) How to handle datatypes other than string that are mapped to names? it would be an error; this should be pointed out.
  • (3.7) Is it true that RDF only allows xsd: datatypes, whereas TMs allow any type? If so, how do we handle the situation? It is not true. RDF does not have this restriction.
  • (3.8) How to support reification of n-ary associations? Follow suggestion in document rdftm:RoleProperty class.
  • (3.9.1) How do we know when a name or internal occurrence is scoped by a language? Define rdftm:Language for the class of all RFC3066 languages
  • (3.9.1) What is the mapping between a language tag and a language topic? RDF uses RFC3066 so we need to support that namespace. Applications can be allowed to provide their own mappings to other namespaces.
  • (3.10) In the absence of guidance for a property, do we assume occurrence? If so, should we simply drop the rdftm:OccurrenceProperty class? Keep the class; applications will be able to allow users to set options to support this assumption.
  • (7.2) Do we want to support reification of roles and, if so, how? [TBD]
  • (7.2) Do we want to support reification of topic maps and, if so, how? [TBD]
  • (Re. PPS's comments) Are TM reification and RDF reification really equivalent, or is the latter about annotating the statement rather than the relationship that it represents? Action Point LMG: Look into RDF definition of reification and ascertain whether it is the same or not. Start discussion on mailing list, if necessary.

3. Discuss issues raised by reviewers (this agenda point was not covered):

  • Bernard Vatant (Feb 10)
    • Thinks identity handling is deterministic. Change 3.3 and 7.1.
    • Suggests using owl:equivalentClass and owl:equivalentProperty in addition to owl:sameAs in order to avoid OWL-Full
  • Peter Patel-Schneider (Feb 11, Feb 13)
    • Doesn't like the term 'proxy': use 'symbol' instead?)
    • Disapproves of use of 'resource': use 'RDF node' instead? What then with 'property'? Better to tidy up usage of resource so that it doesn't cause obvious "offence"?
    • Dislikes talk of 'ambiguity' re. RDF identifiers
    • Thinks we should either support all of OWL or none of it
    • Points out that properties are resources (fair enough; fix wording)
    • Says that RDF reification is "*extremely* problematic" and doesn't mean what we think; this also affects scope
    • Was confused by role types
    • Points out that it is unnecessary to create a second (inverse) statement in OWL
    • Sceptical about use of scope to capture language tags ("appears to be non-monotonic")
    • Uses the term "IRI reference"
  • Elisa Kendall (Feb 16)
    • Recommends going for informative rather than normative mappings
    • Takes issue with focus on guided translations, fearing limited utility because mappings will be project-dependent
    • Recommendation to use annotation properties in #4 in order to avoid OWL Full (?)
    • Comment regarding typed statements in TMs (#5)
    • Comment on container construct in OWL

ATTENDEES:

  • Lars Marius, Fabio, Steve

ACTION POINTS:

  • VP: Write prose description of alternative solution to Identity issue
  • VP: Send email about reification example that causes problems
  • LMG/VP: Provide examples of application of Q, ODM and FOL (?) formalisms to mappings of names, binary associations, and scope as given in draft Guidelines of Feb 10th.
  • FV: Provide additional text for section 1 that explains the goals more clearly (ref. apparent confusion in comments from ODM people)
  • SP: Update draft based on today's discussion
  • SP: Draft replies to BV, PPS and EK
  • LMG: Investigate exact nature of reification in RDF: Is it equivalent to TM reification or not? Discuss on mailing list, if necessary.

OTHER DECISIONS:

  • Put off the issue of how to cover unguided translations until we have the guided translation nailed down.

TIMELINES:

  • All Action Points to be completed by March 3rd.

NEXT MEETING:

  • 2006-03-07 15:00

-- StevePepper - 21 Feb 2006
to top


You are here: RDFTM > MinutesOfConferenceCalls > ConCall20060221

to top

Copyright © 1999-2017 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Fabio's Wiki? Send feedback