It sounds like reification with a generic fuzzy/probabilistic magnitude is not in scope for this property. Reading [1], it seems that circa is used with a point in time and/or one or both endpoints of a time segment. [1] https://en.wikipedia.org/wiki/Circa As an aside, this is a great book about modeling reification and n-ary properties: http://patterns.dataincubator.org/book/ Such a reification pattern would be generally useful for working with OpenCog AtomSpace "Weighted, Labeled Hyper" graphs (and PLN). On Mar 4, 2015 12:52 PM, "Vicki Tardif Holland" <vtardif@google.com> wrote: > > On Wed, Mar 4, 2015 at 12:37 PM, Wallis,Richard <Richard.Wallis@oclc.org> > wrote: > >> >> Personally I believe that creating subTypes Date is still a good >> pragmatic solution to the requirement(s) >> > > I agree. I'm fond of Role for many things, but a subtype is a better tool > for this. > > - Vicki > > Vicki Tardif Holland | Ontologist | vtardif@google.com > >
Received on Thursday, 5 March 2015 06:24:48 UTC