Re: Date and McGoveran comments on view updating 'problem'
From: paul c <toledobythesea_at_oohay.ac>
Date: Thu, 11 Dec 2008 19:48:36 GMT
Message-ID: <o8e0l.3$xD3.0_at_edtnps83>
>
> Correct, there is no restriction onto S and SP whatsoever. There are
> two natural conditions on delta relation D, however. It should have
> the same header as join of S and SP, and be disjoint with the S ^ SP
> view.
Date: Thu, 11 Dec 2008 19:48:36 GMT
Message-ID: <o8e0l.3$xD3.0_at_edtnps83>
vadimtro_at_gmail.com wrote:
> On Dec 11, 5:37 am, paul c <toledobythe..._at_oohay.ac> wrote:
>> In the meantime, I'd like to point out that based on >> >> http://arxiv.org/pdf/0807.3795, >> >> SP v S = S can't be true given the usual attributes Date ascribes to SP >> and S, for example Supplier_Name or somesuch is in S but not in SP. I >> think it could be true for SP{S#, P#} V S{S#} = S{S#} with a further >> proviso that S# is a foreign key.
>
> Correct, there is no restriction onto S and SP whatsoever. There are
> two natural conditions on delta relation D, however. It should have
> the same header as join of S and SP, and be disjoint with the S ^ SP
> view.
Thanks. BTW, I know many people agree that those "conditions" are "natural". By my narrow view, which I like to think of as 'pretty pure' theoretical, I'd rather think of them as mere psychological conditions. (I say "pretty pure" rather than "purely" because I don't yet understand some of the arguments about infinite terms in a wff that uses "Exists", so I usually ignore them. But when thinking about the A-algebra, I always happily ignore the question of finite machine memory.) Received on Thu Dec 11 2008 - 20:48:36 CET