pmika at cs.vu.nl
Thu Feb 23 12:18:59 UTC 2006
I agree with all you suggest except:
> I don't think foaf:published (with any usage) can be used for
> disambiguation, it is not inverse-functional as two persons may have
> exclusively published as co-authors, and thus share the same list of
foaf:published (or foaf:made for that matter) is clearly not inverse
functional as the same thing can be made by multiple persons. However, in
many cases they are "almost inverse-functional" and therefore highly
relevant for disambiguating instances (smushing).
The mentioned case with the publications is a good example: if you find two
foaf:Persons who have foaf:published the same publication, you cannot
conclude that they are the same person. However, if their name also matches
then you can be _very_ certain: it is highly unlikely that two authors of a
publication have the same name.
So alone it is not enough to disambiguate a person but taken together with
other information it can very well be.
Unfortunately, I don't think content-negotiation is or ever will be very
widespread, so I'm in favor of having a property for which we know what kind
of foaf:Document we get as a value.
More information about the foaf-dev