Re: RM's Canonical database

From: Cimode <cimode_at_hotmail.com>
Date: 3 Jul 2006 12:41:59 -0700
Message-ID: <1151955719.599553.94200@p79g2000cwp.googlegroups.com>

Bob Badour wrote:

> No. The predicate of the view constrains what the application interacts
> with.

Wrong. A very limited perception and misunderstanding of RM true nature.

Any predicate is the projected product of one or more relations and constitues the formal assertion of *how* the result relvar (you call view) should respect predicate theory. (Each occurrence verifying TRUE). predicate of a view does not constrain anything at presentation time (in this case applicative layer) because it is a consequence, it is the operation itself and the characteritics of relvars operated that determine what the application can or can not see.

> Thus the application only interacts with the subset of the data
> that complies with its rules.

A meaningless sentence. application do not set rules with any subset of data. Received on Mon Jul 03 2006 - 14:41:59 CDT