>> Yep, very right. But there are some other 'errors' or constraints that >> the SKOS schema itself does not prevent/enforce, e.g. 'No two concepts >> in the same concept scheme may have the same value for skos:prefLabel in >> a given language.' We could include just one more constraint or >> 'warning' rule. >It's entirely natural and healthy for the machine-readable bits not >to capture everything. Otherwise, we'd have no reason to go on inventing >ever richer formalisms :) (OWL, RIF Rules, ...) Vicious argument, especially for formal model designers ;-) I actually also agree, but in that case some guidelines could be provided in the docs, especially when the things not dealt with by formal systems are as important as this prefLabel constraint is for SKOS (and would the 'uniqueness of terms' be for the Term-as-class solution we were discussing). Antoine
Received on Thursday, 1 June 2006 17:44:58 UTC