Re: foundations of relational theory?

"Tony Gravagno" <g6q3x9lu53001_at_sneakemail.com.invalid> wrote in message news:jfgrpvkaluu7ad4g5ube55f397babruo16_at_4ax.com...
> This comes to the argument "guns don't kill people,
> people kill people". Therefore, don't blame the model for data
> corruption (any model), blame the people who corrupt the data by act
> or omission.

But that leads me to the idea that I want a model that will make it hard to make accidental mistakes. I agree that a malicious insider is something that no system can defend against.

It's important that integrity be enforced at the "bottom" level, and not any higher; it sounds like you've got this idea covered.

I'd still argue that a declarative integrity enforcement system is better than a procedural one. Also, having it be centralized (rather than procedural) opens the possibility of writing applications in other programming languages besides the one the database prefers. Allowing only BASIC cuts one off from quite a good deal of recent programming language advancements. I saw firsthand Pick losing contracts on that basis alone, in the mid-1980s.

Marshall Received on Tue Oct 28 2003 - 04:45:10 CET