Re: URL domain (too much SODA)

J.

>What happens to the theoretical basis for relational databases if one
>goes away from that, is unclear. It might well be improved if done
>correctly, but just what constitutes that correctness?

Agreed. The examples I've seen of departure from the Codd and Date model are not theoretical explanations. They are databases that seem to have been designed with an alternative model in mind. The operative word here is "seem". In the absence of specific documentation concerning the design intent, one often has to make inferences about what the designer might have been thinking.

I some cases, design intent doesn't matter. The only thing worth considering is
design consequences. But often, intent can help provide a clue to consequences.

--
Regards,
    David Cressey
    www.dcressey.com