Re: Sixth normal form

"paul c" <toledobythesea_at_oohay.ac> wrote in message news:xV5si.22613$rX4.14717_at_pd7urf2no...
> Brian Selzer wrote:
> ...
>> It's somewhat unrelated, ...
>
> I was with you up 'til the first comma in the last paragraph. It sure
> doesn't have to do with redundancy. First there were rigid keys, then
> contingent keys, now we have immune keys. Newcomers please note - don't
> believe this mumbo-jumbo, it is all undefined mysticism, as is the term
> "key update".
>

Are you saying that Codd's use of "key update" in his definitive1970 paper was also mumbo-jumbo?

Isn't it strange that people who can't come up with a solid argument resort to misrepresentation, nitpicking over terminology, and outright derision. Often they end up with their foot in their mouth.

> Also, I would like to know if anybody has ever invented a relational
> algebra or calculus that supports "update". AFAIK, update is either a
> system colloquial, environmental or a language, say SQL, term. Yeah, I
> know people say "update" a tuple et cetera and I think that's okay if it
> is agreed and understood just what exactly the system or dbms they have in
> mind is going to do about the "update". In the recent threads, I don't
> and I'm pretty sure we don't either.
>

It should be obvious that relational algebra and calculus are meant for queries, not updates. Unlike the operators of relational algebra, neither insert, update, nor delete returns a result. Nor does relational assignment. More importantly, the algebra and the calculus involve a single database value (set of relation values), a snapshot of reality if you prefer. An update operator (as D&D call them) involves two snapshots of reality, and truth in one does not imply truth in the other.

Before you make a complete fool of yourself, I suggest you read Codd's 1990 book. You should be able to find it online now. I did. I can't remember where, though.

> p
Received on Thu Aug 02 2007 - 03:50:19 CEST