Re: reconciliation of disparate models - Jeff

Jeff:

    I believe that OWL 2 defines "property chains" that
    would allow a direct Work-to-Manifestation relationship
    to be established at the ontology level.  For example,
    FaBiO uses this feature for its fabio:hasManifestation
    property. I don't know enough about OWL 2 to say if
    they've used it properly or not, but in principle such
    an addition shouldn't be difficult. (Assuming IFLA is
    willing to add it.)

    http://sempublishing.svn.sourceforge.net/svnroot/sempublishing/FaBiO/fabio.owl

    I would also note that I share SOME of the concerns about
    constraints, but not all. The constraints in OWL are on
    the model, not on RDF representations. We need to get
    used to living in an open world where information is
    often incomplete. Removing OWL constraints may help us
    feel better about ourselves, but not without an offsetting
    loss of instruction.

    ...
    I particularly like their premise that WEMI should be
    subclassed and those classes should be modeled on intuitive
    categories capable of handling the 80/20 rule. I think
    their model deserve more debate, but what I've seen so
    far seems pretty sensible.

    I do think it would be better if they pulled out their
    coupling to FRBRCore to a separate RDF document and
    imported it instead, but that's just me being picky.
    Since FRBR split the concept of "book" into 4 classes, we
    are now free (and actually obligated) to reconceptualize
    the term. IMO, fabio:Book makes perfect sense as a subclass
    of Expression. A "Book" can be "exemplified" in one of two
    (or possibly 3) ways: AnalogManifestation (e.g. "print")
    or DigitalManifestation (I think "ElectronicManifestation"
    would be better here) or ManifestationCollection
    (possibly meaning "both"). The way I read this is that a
    fabio:Book can exist as an abstract "Expression" without
    any manifestation, but then it gets "published" into one
    of these Manifestation subclass forms.

-- 
Tom Baker <tbaker@tbaker.de>

Received on Sunday, 13 March 2011 23:39:56 UTC