Re: RM formalism supporting partial information
Date: Sat, 17 Nov 2007 04:30:59 -0800 (PST)
Message-ID: <73841fed-9e55-43d7-b258-a3216629b94d_at_e10g2000prf.googlegroups.com>
On Nov 17, 7:54 pm, Jan Hidders <hidd..._at_gmail.com> wrote:
> As you already noted <=3 essentially represents your approach. You may
I agree that ~1, ~2 and ~3 will probably all be the same.
> also have noticed that <=1 is Zaniolo's approach. The <=2 represents
> the does-not-apply interpretation of null values. I strongly
> conjecture (i.e. haven't formally verified) that <=1 and <=2 are the
> same and <=3 is different, but ~1, ~2 and ~3 are all the same.
However I was thinking that if anything <=2 was the odd man out. In fact, consider the following simple example...
r1 Name
fred
r2 Name
fred bill
Clearly for each tuple t1 in r1, there exists tuple t2 in r2 such that t1 <= t2. Therefore r1 <=1 r2.
NFP[{Name}](r1) = { { (Name,fred) } } is an element of int(r1)
NFP[{Name}](r2) = { { (Name,fred) }, { (Name,bill) } } is an element of int(r2).
int(r1) is not a subset of int(r2), so it is not true that r1 <=2 r2
However r1 <=3 r2 is true. Received on Sat Nov 17 2007 - 13:30:59 CET