Re: reconciliation of disparate models - Karen, Jeff

Karen wrote:

    I would be fine with the constraints in FRBR and RDA if someone could show
    me that they do NOT interfere with interoperability with other data. Some of
    this may be due to my lack of understanding of how linking works.

    The issue that I brought up to the FRBR discussion list was that of linking
    fully FRBR-ized library data with non-frbr-ized bibliographic data, such as
    the data that is coming out of efforts to make citations linkable. I believe
    that in order to enable this we have to add some super-classes to FRBR. But
    if the FRBR community will not agree to that (which so far they do not),
    then anyone faithfully adhering to the library dictates about metadata may
    be SOL. In other words, I can buy into FRBR if some flexibility is allowed,
    but not if it is intended to be inflexible (which seems to be the case).

    Corey, take a good look at how FABiO has defined its FRBR sub-classes (*).
    They seem to have an entirely different interpretation of Expression and
    Manifestation, to wit (whit?) - a book is an Expression in FABiO, and an
    ISBN is an Expression property. How hard will it be to work with a
    combination of FABiO data and library data (in the latter it is very clear
    that a book is a Manifestation and in RDA ISBN is actually defined as
    "Identifier for the Manifestation")?

    * Also, in terms of sub-classing FRBR entities, don't the sub-classes
    inherit the same constraints as their parent classes? I don't see how
    sub-classes help us *expand*.

Jeff wrote:

    I like Jon's answers.

    WorldCat makes a distinction between W and EM and I (in
    a crazy "what's a URI?" sort of way) but we should all
    admit it's a disappointing experience.

    ...
    I don't think "FRBR" (in the "final report" sense)
    or "RDA" (in the RDF vocabulary sense) is a problem
    because, as Jon has suggested, they can both accommodate
    open-world assumptions. I think the problem is "FRBRer"
    because it attempts to be closed-world using OWL as it's
    expression. OWL is inherently open-world (as we should
    expect the future to be).

    I don't think FRBRer is "wrong" in its re-conceptualization
    of library metadata, it's just anti-cataloger.

    To summarize, I think FaBiO generally has it right.


-- 
Tom Baker <tbaker@tbaker.de>

Received on Sunday, 13 March 2011 23:20:30 UTC