Re: Proposal: 6NF

From: Bob Badour <bbadour_at_pei.sympatico.ca>
Date: Fri, 06 Oct 2006 14:22:28 GMT
Message-ID: <EEtVg.1274$cz.18103@ursa-nb00s0.nbnet.nb.ca>

David Cressey wrote:

> "David Portas" <REMOVE_BEFORE_REPLYING_dportas_at_acm.org> wrote in message
> news:1160133403.293728.245040_at_m73g2000cwd.googlegroups.com...
>
>
>

>>I have not dismissed nulls arbitrarily. There are good reasons to
>>reject SQL's definition of a null. I'm still waiting for your
>>definition. Apparently you don't have one for numerics so dare I ask if
>>you have one for boolean? Do you have a null for a graphics image type
>>or a date type or a cartesian co-ordinate? How does your null behave
>>under the operators for those types? I think you have your work cut out
>>here...

>
>
> I don't know what Brian's going to say, but I'll give you my definition.
>
> A null is a marker indicating the absence of a value, in a place where a
> value might normally have been expected.
>
> A null is not a value. A null has no type. And to borrow Roy's phrase,
> "make of it what you will."

Roy's phrase has relevance and utility for reports where a human reads the report and where lacunae and whitespace have utility related to primate psychology. A dbms is a formal logic system where it has no place at all. Received on Fri Oct 06 2006 - 09:22:28 CDT