Re: Normalisation

Paul wrote:
> Jan Hidders wrote:
>

>>>Well, the "atomicity" definition doesn't hold up under scrutiny, and
>>>the "you shouldn't have more than one column about the same thing"
>>>definition is too informal.
>>
>>Yeah, that's what Date claims. Many disagree, including me.

No. But I think it is understood well-enough to be a useful concept.

> Are values in domains not atomic by definition, irrespective of any
> internal structure they may have? (Structure invisible to the relational
> operators, that is).

Yes, but that depends a bit on what you take as "the set of relational operators". If that includes nesting and unnesting then you make the structure of nested relations visible, and you should therefore consider them non-atomic.

• Jan Hidders