Re: separate data/inidex

Nuno Souto wrote:

> In article <aa83v9029oi_at_drn.newsguy.com>, you said (and I quote):
> > >
> > >This myth has been around for a long time - and unfortunately was actually
> > >started by an Oracle consultant (who will remain nameless).
> >
> > And it wasn't me!
>
> Proof?
> :-D
>
> Narh, I think I know who it was. Mind you, back then it probably made a
> lot of sense. Problem with all these "gospel" statements is that they
> tend to be self-perpetuating. Which in a technical arena like IT is a
> recipe for disaster! Like Jonathan recognizes in his book, there is no
> such thing as a technical recommendation that remains valid ad infinitum.
>
> Reminds me of many "benchmarks", heavily promoted here as the be-all-and-
> end-all of performance evaluations. These were originally developed to
> make a Version 5 Oracle database run fast in a mainframe, period. Never
> meant to be then sold as "tests" for V6 and V7 databases. For obvious
> reasons. Yet they were sold as such, and promoted as the ultimate by all
> and sundry including Oracle Consultancy!
>
> Oh well, one day people will learn not to look for silver bullets...
> --
> Cheers
> Nuno Souto
> nsouto_at_optushome.com.au.nospam

It isn't entirely a question of looking for silver bullets. It is rather that you must start somewhere. And from where you start you tune or modify until you get what you want.

Of course no paradigm lasts forever. But you don't create successful applications by starting out with each new database by retesting every single basic assumption. If you did ... nothing would ever be created and deployed. Lets see ... I wonder if in this version the tables are still heaps? I know. We'll run some tests to see if that is still the case. ;-(

Daniel Morgan Received on Thu Apr 25 2002 - 17:22:48 CDT