Alf Eaton wrote:
> "The system can't differentiate between two people with the same name",
> that should have been.
Ok, it might sound wild, but there is a solution for this: assign a
different URI per paper and draw equivalences between them.
It's highly unlikely that two people with the same name co-authored the
same paper, therefore hashing the subject of the paper and the author's
unicode name would yield a unique identifier, which is what we were
asking in the first place.
Now, this identifier is indeed unique, but probably too much, since at
most one paper is associated with that author URI, if we don't draw
equivalences.
The cost is indeed high, compared, for example, at ignoring the issue,
but I think that truly unique URIs + equivalences are the best way
around the homonym problem.
In piggy-bank, we designed the system so that each 'tag' is your own and
then we draw an automatic equivalence if two tags are homonyms *and* if
there is no 'differentFrom' relationship between the two tags.
I would apply the same concept to people.
--
Stefano Mazzocchi
Research Scientist Digital Libraries Research Group
Massachusetts Institute of Technology location: E25-131C
77 Massachusetts Ave telephone: +1 (617) 253-1096
Cambridge, MA 02139-4307 email: stefanom at mit . edu
-------------------------------------------------------------------
Received on Fri Jul 15 2005 - 15:45:39 EDT