A statement on dbdebunk.

One of the last posts on dbdebunk ("On conceptual modeling and database design") included the following :

<quote>
My material is sound because unlike all the stuff floating in the industry, the logical level is formal and the conceptual level was developed to allow 1:1 mapping to it.

· Conceptual model: informal business concepts
(enterprise-specific)

· Logical model: formal representation (as much semantics as a
system is capable of "understanding") (enterprise-specific)

· Physical model: implementation in hardware/software
(enterprise-specific)

· Data model: A "formal mapping construct" to map conceptual to
logical models (universal theory of data)

</quote>

Note in particular :

Conceptual model === INFORMAL
Logical model === FORMAL
Conceptual model "ALLOWS A 1:1 MAPPING" to the logical model

Now, it seems to me that "allows a 1:1 mapping" means that there is some kind of isomorphism between the things that are mapped in such way.

And the fact that some kind of isomorphism between two things can be found, implies that any quality or property that holds/exists/is proven for either of the things mapped, necessarily also holds/exists for the other thing mapped.

So if model X has the property of being formal in some sense, and model Y is in some way isomorphic to model X, then it necessarily follows that model Y is also formal in that same sense as model X is formal .

So the statements quoted here, seem contradictory to me, if not quackery.

Anyone care to correct me on this, or comment in any other way ? Received on Fri Aug 18 2006 - 08:35:57 CDT