Re: Designing database tables for performance?

On Feb 27, 8:17 am, DA Morgan <damor..._at_psoug.org> wrote:
> >
> > You talkj about it as if this was indeed not good enough..Writing crap
> > is easy but writing good books is difficult and noble. Somebody has
> > got to do the thinking before somebody builds the system. Else the
> > product becomes ORACLE (or SQL Server or DB2)
>
> If this was still the 1970s and 80s I would agree. But the ideas were
> fleshed out. The ideas were implemented. The market has spoken.

The market is the arbiter of commercial success, and nothing else. In particular, the market has nothing useful to teach us about science or mathematics. I am a supporter of markets in general, and the EMH as well, at least in its weak form. But one has to understand their limitations: they are in essence an elaborate form of popularity contest.

And: markets change their minds. It is easy to adapt your "the market has spoken" argument to a circa-1985 argument that says we should not pursue C++ (or anything else) because "the market has spoken" and chosen C: it is fleshed out, implemented, and it is what the market picked.

> The only thing new coming out of discussions of theory seems to be
> whining about how things were not implemented in a purist fashion.

Really? How many academic papers from the last ten years of computer science research have you read? Nothing in there except whining? Or by "coming out" do you mean releasing a product? Is commercial activity the only legitimate form of human activity?

> Conveniently ignoring the fact that every attempt to do so
> has been a commercial failure.

Stem cell research has been a commercial failure; I guess we should abandon that, eh? Fusion power? No one's made a billion dollars from the Calculus of Inductive Constructions; I guess it's worthless then.

Marshall Received on Tue Feb 27 2007 - 12:55:02 CST