Am Donnerstag, den 18.10.2007, 12:10 +0100 schrieb Jeremy Carroll: > > ...how should I prevent - in a distributed, uncontrollable environment > > like the Web - the owner of u2 from adding a view based on u to u2 > > tomorrow? > > You don't have to. > > If u2 is grounded then either the new view relates u2 to the same graph > in which case the triple is true, or to a different graph in which case > it is false. > > In the latetr case (falsity), the named graphs approach to provenance > and trust then encourages you to rethink which sources you trust I guess we have different requirements. We do trust the sources, and want to allow building such a network of graphs. We can have hierarchical dependencies of graphs, as you propose. But we can (and want to) do more. Your proposal would be a kind of import mechanism, which is a bit more sophisticated than including a whole graph. Networked Graphs can even be used as a distributed rule mechanism and are much more expressive. Anyway, I do not think I should generally distrust every source, which uses information I have created myself and which I highly trust. This would basically be the result of what you propose. ;-) Best regards, Simon
Received on Thursday, 18 October 2007 13:38:46 UTC