Re: RM formalism supporting partial information

On Nov 16, 10:51 am, David BL <davi..._at_iinet.net.au> wrote:

> From your middle two examples we have
>
> r1 = { (x1={}, x2={}) }
> r2 = { (x1={a}, x2={}) }
>
> Then taking various interpretations of projections...
>
> I(proj({},r1)) = {}
> I(proj({},r2)) = {}
>
> I(proj({x1},r1)) = {}
> I(proj({x1},r2)) = { (x1=a) }
>
> I(proj({x2},r1)) = {}
> I(proj({x2},r2)) = {}
>
> I(proj({x1,x2},r1)) = {}
> I(proj({x1,x2},r2)) = {}

That was a bit lazy. I should be using the (tau) function to map a relation to its body. Received on Thu Nov 15 2007 - 23:07:36 CST