Re: NULLs: theoretical problems?

Jon Heggland wrote:
> Keith H Duggar wrote:
>
> > So it's required to "write DEF" for all nullable attributes and a
> > formula is true only if the nullable attributes are defined. How
> > is this different from dropping the requirement to "write DEF" and
> > simply defining that a formula is false if any of it's variables
> > is undefined? And how is that different from allowing NULL and
> > defining that a formula is false if any variable is NULL?
>
> Note that SQL does not actually do this, and Jan Hidder's example
> depends on it. TRUE OR NULL is TRUE, not NULL (or FALSE).

SQL is rarely if ever in my mind when asking a logic question.

> Would you have it otherwise?

I'm not trying to have anything. I'm trying to understand what the "write DEF" prescription buys us over say the "Interactive Mathematical Proof System" of Farmer, Guttman, and Thayer that has exactly the property I described that any formula is false if any variable is NULL. Again, I am failing to grasp what the "write DEF" prescription buys us. I would like to understand.

KHD Received on Fri Aug 31 2007 - 03:59:51 CEST