Re: Multiple specification of constraints

"Mike Sherrill" <MSherrillnonono_at_compuserve.com> wrote in message news:bp4s30td62buv8vo101rthuhm6gja6vl8v_at_4ax.com...
> On Wed, 25 Feb 2004 12:39:18 -0600, "Dawn M. Wolthuis"
> <dwolt_at_tincat-group.com> wrote:
>
> [snip]
> >My point is that with the way that software is written today, quite often
> >data constraints are coded in multiple places, even multiple languages.
>
> That's certainly true. But the problem is not necessarily in having
> multiple expressions of a constraint. The problem is in having
> multiple uncontrolled expressions of a constraint.
>
> >I'm suggesting something rather radical --
> >decouple the constraint specification and validation logic from the
> >database.
>
> Why? It's not crystal clear from what you posted.

Apparently, Dawn thinks DOS 1.0 has advantages over the relational model. Ignore her. She is an idiot. Received on Wed Mar 10 2004 - 04:04:15 CET