Re: Does Codd's view of a relational database differ from that of Date & Darwin? [M.Gittens]

Marshall Spight wrote:
> Jan Hidders wrote:

>>
>>Not precisely, because "x > 5" would then be "false" but under the
>>"unknown value" interpretation the answer should be "maybe, maybe not".

Yes.

>>It's interesting if only because it is the formal definition of what the
>>"missing value" interpretation actually means.

My mistake, I meant "unknown value" not "missing value". The interpretation I'm talking about is the one where the truth value of a formula with unknowns (variables) is one of three:

TRUE: if for all possible values of the unknowns the formula holds FALSE: if for all possible values of the unknowns the formula is false UNKOWN: in all other cases

> [...] what are some other ways to deal with universal quantification
> besides unification?

For an example see 'description logics' which deal with a decidable subset of first-order logic that does include some universal quantification. There's a whole area of research dedicated to practical theorem proving.

> I'm guessing a bit here, but I suspect the "unattainable" part
> refers to uncomputability?

Yes. Usually a big bummer when you want to implement something.

> Overall, from a bang-for-the-buck perspective, the complexity
> and cost of *unknown* do NOT appear to be worth the modest
> utility. (Although I suppose it might be better to have SQL's
> null than nothing (ha ha,) if I couldn't have union types or a
> cardinality-0 solution.)

No argument there. Still, I think it's interesting to investigate (1) what exactly is the "unknown value" interpretation, (2) why is it so impractical and (3) to what extent can we in practice approximate this ideal.

• Jan Hidders