Re: Modelling Considered Harmful

mountain man wrote:
> His model, and the RM of the data belongs to the 1970's, prior to the
> release of Oracle, and in the 1980's DB2 and SQL Server, among
> other RDBMS software.
>
> Codd's and Date's model of the data is thus antiquated, except for
the
> purpose of historical reference, seeing as though it has been
realised
> for over 30 years within the machinery of software (RDBMS vendors).

Apparently the above "belongs to the 1970's" and "antiquated" [sic] are meant to indicate a deficiency, but keep in mind that math, logic, physics and many other fields of study are much older as well.

And as far as being realised - well no, it hasn't.

> Harmful? Yes. Database theorists and academics are having a field
> day in generating useless pedagogic literature that bears absolutely
no
> reference to the technical reality and/or USE of modern RDBMS
> software.

I don't understand how this would be harmful even if it were true. Theorists and academics aren't all charged with providing useful tools to industry; basic research requires creativity, risk, and often foolishness. Much useful information has come from such.

In short: industry capitalizes on the findings of academia, and while the practices in any field can be a field of study in its own right, how I drive my car has no bearing whatsoever on, for example, fuel science. Are you certain that the findings in academia have no bearing on practice? The usual course of events is that vendors completely ignore useful results from academia in favor of whatever their users scream most loudly about, and whatever other vendors are doing... so even when there is useful research around, it's ignored.

• erk