Re: (OT)Dynamic inheritance (was: Object support in the relational model??)

>> OO Prescription 3: Computational Completeness

> Thanks for reminding me of that. But resource to authority never
> went well with me, you see, I'm Baptist.

:-) In general I agree, just a time saver sometimes.

> Unfortunately I don't have TTM in hand now, and AFAIR this was the
> one part in it that didn't convince me. I can't see why one couldn't
> implement D in Scheme or some other Lisp dialect, for example -- or
> even Java, C# or some other C derivative -- and have users defining
> functions in it. I haven't checked, but I think this is the approach
> taken by Alphora in Dataphor.

The gist was that they wouldn't prohibit using D as a sublanguage, but D should not be hobbled in such a way as to *require* another language. All that 'impendence mismatch' I suppose. Anyway, Java, C etc don't support Relations, so why on earth would you want to use them for anything?

It seems to me that it was a big mistake to not make SQL computationally complete from the start. I know it's getting there, but, I still get annoyed when there is a task that I can't do in SQL (which, admittedly, is not so often nowadays). My head just has no space left in it to be truly fluent in more than one language. But then, I never got the hang of French either.

Regards
Paul Vernon
Business Intelligence, IBM Global Services Received on Fri Jun 14 2002 - 19:02:54 CEST