Re: The Null Problem is a Non-issue

On 2014-09-02, Derek Asirvadem <derek.asirvadem_at_gmail.com> wrote:
> The context is, I have a couple of papers written re Relational Databases,
> and I would like to find technically qualified people who can review them.

with the hidden context "and not declare them to be rubbish, even if they are!"

8>< other preliminaries snipped

Of course you don't want any answer of mine, but this is Usenet, I am giving you one anyway. Feel free to ignore it.

> 1. The Null Problem does not exist in the physical universe (note,
> I did not say "Nulls do not exist").

I am quite prepared to agree that a null problem does not exist in the physical universe at a human scale.

> 2. The Null Problem exists in the unreal universe of discourse, beloved
> of pseudo-theoreticians, who cannot relate to the real universe.

The null problem exists in theoretical discussions about relational databases because many people have failed to understand the Relational Model, or the nature of the problem they are trying to solve with nulls.

It also exists in actual databases due to a similar lack of understanding.

> 3. The 3VL and 42VL that is yet to be resolved is close to their heart,
> and they will never resolve it, because if they did, they would not have
> a Thing to argue endlessly about.

Anyone who thinks that any level of multivalued logic is the answer to a problem of database design is wrong, even E.F.Codd when he did this.

> 4. The 3VL and 69VL is irrelevant to the physical universe.

True.

> 5. The pseudo-theoreticians on the SQL Committee implemented the Null
> Problem in the SQL standard, thereby inflicting upon the physical universe
> a degree of pain that it did not have. Whereas programming was previously
> a finite task, these creatures of the deep have made it an infinite
> one, where every query must be scrutinised for an issue that does not
> exist. This is merely the insane enforcing their insanity on the sane.

It is a committee including members whose actual aim is not to create the best possible standard. What do you expect?

> 6. The commercial RDBMS vendors (the above four) have implemented
> a finite SQL, which does not have The Null Problem, thereby relieving
> their customers from being reduced to the capabilities of the insane.

What about Oracle? It is as good (or as bad) as the rest.

You are wrong anyway, it is possible to use any of them in ways that involve some form of a "null problem".

> 7. Hundreds of thousands (Millions?) of developers who use such products
> have never heard of The Null Problem or 3VL or 56VL. (They still have
> to deal with queries that return Null, as distinct from having to deal
> with The Null Problem.) For three decades.

And many, many of them have produced flawed systems because they have heard of nulls and do not understand them.

> 8. The Null Problem is a non-issue in the physical universe.
> It remains a glorious and sanctified issue in the unreal universe.
> For three decades.

You already said this.

Eric

-- 
ms fnd in a lbry