Sort of... withdrawn:-) Having discussed with Ian, too, he made me realized that the OWL RL rule set can be (and will be) applied on a larger set of graphs, and this affects the completeness criteria only. If that restriction is required for the completeness then, well, let it be:-) It will not affect the way rules will be used I guess. ivan Ivan Herman wrote: > Boris, > > for my understanding: does it mean that, for OWL RL, a compliant > implementation using the rule set would be expected to disallow the > usage of BNodes in the rules? At first glance this seems to be fairly > restrictive and not really understandable for neither the users nor for > implementers; indeed, the rule set seems to be absolutely oblivious to > the fact whether some of the resources are BNodes or not. Consequently, > I would expect rule based implementers to ignore this restriction. > > Can you explain the rationale for this? Is it _really_ necessary to > disallow BNodes? > > Ivan > > Boris Motik wrote: >> Hello, >> >> It turns out that, when we introduced anonymous individuals into OWL 2, we implicitly introduced them into all the profiles as well. >> I only noticed this yesterday while creating the complete grammars -- good thing people suggested that I do that! This is clearly an >> error, as anonymous individuals increase the complexity of reasoning in some cases. I have corrected this error, but, because it did >> involve a small change to the language, I thought that I should at least bring it to everyone's attention. >> >> Regards, >> >> Boris >> >> > -- Ivan Herman, W3C Semantic Web Activity Lead Home: http://www.w3.org/People/Ivan/ PGP Key: http://www.ivan-herman.net/pgpkey.html FOAF: http://www.ivan-herman.net/foaf.rdf
This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 29 November 2008 11:48:05 GMT