[rdfweb-dev] Smushing the Semantic Web / Invalid data in IFP's

Ian Davis iand at internetalchemy.org
Mon Mar 22 13:09:51 UTC 2004

Perry Lorier said:
> Since I assume these all come from the LJ data, and LJ has one nick per
> person, this seriously shows that smushing on IFP's with the presence of
> bad data is practically useless.
This is showing a flaw in the LJ interface. Are they forcing people to
enter Jabber IDs?

In that idealistic fantasy world that I sometimes inhabit I'd say this is
LJs problem - they should be validating their data.

In the real world however, we're going to have to deal with this. At some
point your smushing is going to uncover a contradiction. You need to have
a strategy for handling this, e.g. rejection of the sources that caused
the contradiction.

It may be that using only inverse functional properties that we can't
force enough of these contradictions to be useful ("hey this guy's got a
LOT of email addresses.."). We probably need a smattering of functional
properties as well ("hey this one's got two dna fingerprints, now THAT
aint right...")


More information about the foaf-dev mailing list