Open BarryNorton opened 11 years ago
I would rather suggest the second one.
It raises an interesting question though.
To motivate Membership as an event, consider Robbie Williams. He was a founding member of Take That, then he left, then he came back (a separate event).
Now is his second membership a founding membership, or not? (Perhaps not if we call the event that: FoundingMembership)
If not, do we constrain (if we were to fully axiomatise) the earliest Membership event between two given Artists to be the only one allowed in the FoundingMembership subclass?
Well an artist could be part of one FoundingMembership event and several Membership events per group. I wouldn't suggest (more) constraints as a group could also be founded by more than two members, or?
On 3/2/2013 2:54 PM, Barry Norton wrote:
It raises an interesting question though.
To motivate Membership as an event, consider Robbie Williams. He was a founding member of Take That, then he left, then he came back (a separate event).
Now is his second membership a founding membership, or not? (Perhaps not if we call the event that: FoundingMembership)
If not, do we constrain (if we were to fully axiomatise) the earliest Membership event between two given Artists to be the only one allowed in the FoundingMembership subclass?
— Reply to this email directly or view it on GitHub https://github.com/motools/musicontology/issues/10#issuecomment-14328502.
The point of the would-be constraint is not that Robbie should preclude Gary's also having founded the band, but that Robbie's re-joining raises the possibility that one models his first membership without classifying that as a founding membership and his second one as such. Which is odd.
In fact, the more I think about it, the more I believe if you were a founder you're a founder for life, whether you have a current membership or not. This argues against the Membership subclass, but also argues against the simple subproperty unless mo:member were clarified to mean 'has at some point been a member', rather than being something like: http://dbpedia.org/ontology/currentMember
(That this feature is underspecified is consistent with its being marked unstable)
Yep. You are right in both points. So we should redefine mo:member to something like 'has at some point been a member' and maybe introduce more sub properties of mo:member such as mo:current_member and mo:founding_member. Finally, you can utilise the short cut descriptions (via mo:member and sub properties) and the more detailed ones (via mo:Membership) side by side.
Create a representation of founding members, perhaps either (Barry's initial suggestion) as a subproperty of mo:member or (Yves' refinement) as a subclass of mo:Membership.