At 04:41 AM 6/1/2004, johnsparrowuk wrote:

>I seem to remember reading an artical about this years ago.
>Apparently there was some 'discussion' inside the Interbase company
>when Borland bought it, about the usefulness of SP (and I assume
>triggers too). I got the impression the people at Interbase saw no
>advantage in it, but were forced by Borland.

Actually, it was Jim who didn't see a reason for stored procedures,
though he did recognize the benefits of triggers. He associated
stored procedures with Sybase whose TransactSQL really needed the
performance boost of stored procedures. The BLR language allows
a program to pass an entire subroutine to the engine - conditions,
branching, looping, variables - so you got the processing on the
server side without putting the code there.

Sometime around 1990, we realized that people were creating views
with triggers as a way to produce the encapsulation that stored
procedures offer. The transfer to Ashton-Tate (and Jim's associated
departure) accelerated a strong push for Stored Procedures.

As for the stored procedure vs declarative language issue, as
a general rule, I prefer declarative language to procedural language.
(Yes, I sort of favored QUEL, too.) But the flexibility of triggers
and stored procedures is quite a challenge to emulate in any
declarative language, so I think we'll have procedures and triggers
for a while.

Regards,


Ann