Stu Weibel wrote:
>I am of the opinion that pragmatics are more valuable than principle
>here, so I come down on the side of 15 elements and a small number of
>sub elements.
ok
>I also would very much like further well-reasoned opinions on which way
>to jump with Relation:
>
>a.) a small number (4?) of explicit attribute types (Relation.IsPartOf,
>Relation.HasPart...etc.)
>
>b.) exactly 2 attributes for all time (Relation.Type, Relation.Target)
>with enumerated lists of types... perhaps a DC default list and then any
>scheme-qualified lists that others come up with. If so, what is the
>minimum, default list for DC-Simple?
I was reading this thinking I would far prefer option (2), since it doesn't
take a redevision of the DC itself every time we want to add a relation
type, which we will doubtless want to do. But then it seems to me that
there is nothing special about Relation in this respect, and that with dates
as well we could have Date.Type and Date.Content, with an initial list of
approved types. And with other subelements as well.
Are there implications for representation in RDF? What do the people
working on the data model think?
p
|