David Cressey wrote:
> When those who know some body of theory deliberately make a design choice
> that the theory doesn't recommend, for reasons that are outside the scope
> of the theory, it can work out well. When those who don't know some body
> of theory design without regard to it, the results are frequently
> disastrous.
And in the case of SQL, from what I know most of the decisions were
made to test performance of an RM-like DBMS - not to test the utility
of the RM as a whole. It was a limited prototype which soon eclipsed
other efforts, for reasons probably related to its superiority over
network and hierarchical DBMSs.
Received on Wed Jun 01 2005 - 08:53:09 CDT
