At 11:52 PM 11/3/2000 -0400, Claudio Valderrama C. wrote:

> Hmmm, you are trying to implement poor man's version of DEFERRED
>constraints, but manually controlled: deactivate a constraint, do all the
>operations and then activate the constraint. Will the constraint check the
>data when it's activated again?

Yes, but this is an inefficient as well as inconvenient way to
pretend to have deferred constraints. Two people I respect as
database designers (Jim & Monty Widenius) both think that uninforced
constraints have value. So why not?

Regards,

Ann