Re: Multiple specification of constraints

"Dawn M. Wolthuis" <dwolt_at_tincat-group.com> wrote in message news:c25cbc$1c0$1_at_news.netins.net...
> There are definitely local constraints for specific applications that
would
> not be specified as global constraints.

No, there aren't. If app X deals with a completely different set of data than app Y, then you can easily apply each app's "individual constraints" as global ones, qualified by the subset predicates. And if they deal with the same data, then you can't have X writing data that violates Y's constraints, can you? That would be nonsense.

Do you have an example? If you're talking about simple data transformation, then calculating f(a) requires that a meets its constraints, and if you're storing results in b, that f(a) meet the appropriate constraints for b.

I'm not sure where the need for this division is.

> It would be foolish to specify all
> global constraints in one language and apply them with one data integrity
> engine, while coding local ones using different tools entirely, I would
> think.

I agree.

> Local constraints are just global constraints that have some
> additional conditional logic in them.

Then they're global. For example:

A => X & B=>Y

A and B can be propositions about the same data, and X and Y indicate their implications. A can be for app A, B for app B. The global constraint is their conjunction (the above can be written differently, of course).

> I better do some "real work" for a while, so I'm leaving many juicy
> responses unanswered for now.

Yeah, me too... hard to tear myself away, though...

• Eric