Re: Dreaming About Redesigning SQL

"Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in message news:bn5v3h$1gka$1_at_gazette.almaden.ibm.com...
> "Anthony W. Youngman" <thewolery_at_nospam.demon.co.uk> wrote in message
> news:yzzLHJAS9Yl$EwIB_at_thewolery.demon.co.uk...
> > I just feel that as a *mathematical* *theory*,
> > any attempt to enforce it is likely to be counter-productive, because
> > the real world is not a theory ...
>
> This is probably a bit philosophical, but I disagree.
>
> If a theory does not reflect the 'real world' then that theory is broken.

The key point here is that the theory does, in fact, reflect the real world much better than Wol's ad hoc crap.

> [snip]
> > >Now, I might be generous and say that the above is predicated on the
> > >assumption of data independence - the idea that the physical layout to
> store
> > >the data be separated from the abstract representation that user use to
> > >view, create, query and modify that data. So if don't believe in this
> > >principle (which I suspect you don't) then I guess we will struggle to
> > >convince you of the essentiality of the relational model.
> >
> > Yes - I most definitely DON'T believe in that.
>
> Cool. We make progress. You don't believe in physical data independence.
> Maybe you should therefore show you understand the benefits of data
> independence and show how they are not really so beneficial in your
opinion.
> Otherwise I'll assume you don't understand the benefits and stop the
> conversation here.
>
> > You notice I gave Lauri a
> > formula to calculate how long a query would take. I don't think he
> > succeeded in convincing me he could reciprocate in kind (I've still to
> > study his answer), and
>
> > proving that a query will complete in finite time
> > is of no use if "finite time" equals "longer than I can wait".
>
> Agreed.
>
> [snip]
> > Actually, I would contend that the MV model gives you exactly what you
> > are looking for here - execution time IS proportional to complexity. We
> > do expose the theoretical to the physical, and the "cost in lack of
> > generality" is negligible.

Given that Pick/AQL cannot even express simple questions using simple sufficient data structures, his assertion is fatuous at best. The cost is enormous when one must physically restructure one's data and add considerable meaningless complexity just to ask a simple question.

Given the complexity one must add simply re-casts MV as single values, Wol is apparently too stupid to understand what that says about MV. Received on Wed Oct 22 2003 - 15:25:05 CEST