Re: A possible way of going forward with OWL-R unification (ISSUE-131)

Just one more, higher level, point.

I'm *extremely* leary of working into a standard somewhat non- 
traditional semantics. The RDF and OWL experience should make clear  
that doing simple, somewhat odd by seemingly harmless things can  
hugely come back to chomp you in the buttocks. Once in a standard, it  
becomes much much harder to change or otherwise work around.

Obviously, if bizarre stuff that's hard to work with gets into  
widespread implementations and applications, that's a force for them  
as well. But at least there the choice *is* forced and usually guided  
by attempts to use those implementations. Thus, it's much clearer the  
costs and benefits of change.

So I would argue for conservativity at the moment. If during CR, for  
example, it becomes clear that we should add something more to OWL-R,  
that would be a good reason for going back to last call.


Received on Thursday, 17 July 2008 11:14:14 UTC