I hesistate to offer an opinion, but this seems to make a point for good
authority
control. The email address could be an attribute of an entity that is
the
creator. My past opinion has been more that authority control was nice
to have,
but I questioned it at times considering the work it took.
Any opinions?
Gary Masters
Dan Brickley wrote:
> Here's a slogan for thinking about 1:1 and its relation to the Dublin
> Core datamodel issues:
>
> Don't ascribe properties to an object that are
> really properties of some other related object.
>
> In other words, don't pretend that the email address of the creator of
> an image is a property of the image. Don't pretend that the fax number
> of the office of the publisher of a book is a property of that book.
>
> Here's another:
>
> Don't make category mistakes.
>
> The philosopher Gilbert Ryle used the term 'Category Mistake' to talk
> about this problem (it's a kind of mistake philosophers are
particularly
> prone to):
>
> "It is one big mistake and a mistake of a special kind. It is,
namely,
> a category mistake. It represents the facts of [....] as if
they
> belonged to one logical type of category (or range of types or
> categories), when they actually belong to another." [1]
>
> Now I'm not normally prone to quoting philosophers on DC mailing
lists, so
> am a little wary here, but I do think this nicely characterises the
> problem. Talk of "metadata records" is a red herring in the 1:1 and
> data model debates. The issue at heart is about the confusions that
> follow from making category mistakes, from ascribing properties to the
> wrong (type of) thing. This hooks up the 1:1 issue to the
> work/manifestation/copy way of thinking and to architectural (RDF data
> model) and conceptual (Schema harmonisation) issues.
>
> It also explains the concern some people had when they saw DC which
used
> constructs such as DC.Creator.Address.Street.PostCode. To think of the
> postcode of the street of the address of the creator of a resource as
if
> it were an attribute of a Document is a mistake of a certain sort. It
> suggests that we are making a category mistake, an exercise in poor
> data modelling, and a violation of the 1:1 principle. If RDF gives us
> anything, it is a formalism whereby we can draw out these complex
> relationships in an explicit graph diagram and make explicit what once
> was implicit.
>
> Dan
>
> [1] Gilbert Ryle, 'The Concept of Mind', 1945.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|