3 ms·

I think his point could have been made clearer, but he's really just saying that there will often be a mutable natural key that is captured in the modelled data

I think his point could have been made clearer, but he's really just saying that there will often be a mutable natural key that is captured in the modelled data, and that it makes sense to enforced constraints over that key if you know that the constraints exist.

chris_wot 13y ago
Ok, maybe. But I can only think of examples where if the schematic was normalized properly then it would be a non-issue.