Re: [RIFWG] [Requirements?] A vision for the RIF

Yes, you need some kind of a non-first-order trick to have a useful notion
of constraints. At least, this is all I've seen.


	--michael  


> I'm not particularly up on the literature on integrity constraints, but 
> what I've seen seems to fall in line behind Reiter and Michael (roughly 
> that integrity constraints are not first order/are "epistemic"). (And 
> this isn't just in the LP or database literature.) That doesn't mean 
> that there isn't something I've not seen, of course. So I'm wondering 
> if anyone has some pointers to alternative conceptions (preferably post 
> "On Integrity Constraints"). (That is, I'm looking for evidence that 
> there are pre-existing such conceptions. The "I can imagine" constraint 
> seems a bit weak :))
> 
> I'm definitely missing what's at stake :)
> 
> Cheers,
> Bijan.
> 
>