Re: Time ontology extension for non-Gregorian calendars

> These might be models for you.

Excellent, thanks!

> From a technical perspective, there is a concern not to compromise 'backward compatibility' in the sense that data (individuals, a-box) that follow the current ontology do not become inconsistent or get harmful entailments as a consequence of a new axioms (t-box).

Sure, I don't think it will be a problem

 > Note the approach I took for the 2017 version of OWL-Time was to
interpolate super-classes with more generalized definitions above the
existing classes that were designed for the Gregorian Calendar. That
is where those 'GeneralizedAAA' classes and 'GeneralBBB' datatypes
came from. I suspect a similar strategy should be applied to support
your use-cases.

Yes, I think it should be fine.

I'm thinking of doing this short spec at the same time as two other
specs which will be independent: one for Indic calendars and one for
Chinese calendars (that I don't think belong in the OWL-Time ontology
but I'm happy to change my mind). Building the three of them at the
same time will allow me to be sure that the OWL-Time extensions really
work.

It will probably be a few weeks/months before I come back with a
proposal, in the meantime I'll probably have some question on a few
details.

Best,
-- 
Elie

Received on Monday, 25 May 2020 06:51:52 UTC