--- In Firebird-Java@yahoogroups.com, "Steffen Heil" <lists@...> wrote:

> If you compare NULL=NULL, however, you want to know, if two things -

that

> you do not know at all - are equal. This is undecideable. Therefor

NULL=NULL

> returns false as well as NULL<>NULL returns false.
>

Briefly, i would have preferred to have difficulties to deal with the
unknown than with the emptiness.

You are totally correct but i wonder how many people needs to deal
with unknown values and how many people needs empty values for which
they abuse null in real world.

We have NULL that is a huge instrument that represents the unknown and
we have nothing that helps us in dealing with the very well known and
much more practical emptyness of something.

Think about a number...the only values i can store are:
- the value
- null -> unknown
There's no way to tell "I know, it's empty. No, it's not unknown
because of that, it's empty and so it's different with something that
is not empty"
And maybe i can't sacrifice a real value to say "Hey, this value means
empty"...
But i can easily say it is unknown. :/

I think it's just a matter of what is more needed in practical
applications...the unknown concept and the logics that it brings is of
little use in great part of the real world.
Users want answers, yes or no, not "i don't know"...in the majority of
cases.

So i say that was a bad choice to give so much importance to the unknown.