Re: Modelling Considered Harmful

From: Kenneth Downs <knode.wants.this_at_see.sigblock>
Date: Tue, 26 Apr 2005 18:53:56 -0400
Message-Id: <lft2k2-8fa.ln1@pluto.downsfam.net>

Paul wrote:

> Kenneth Downs wrote:

>> Correct.  The Relational Model (capital R, capital M) can be used as a
>> guide
>> for building relational databases.  An employee database cannot be used
>> as
>> a guide for building people.  The first is a model, the second is record
>> keeping.

>
> A model in the sense of a "mathematical model" is just a simplified
> conceptual version of reality though (or fantasy, although you could
> argue this is just a part of reality that exists in our minds). Useful
> for making predictions. e.g. Newton's and Einstein's models of gravity
> aren't a guide to creating gravity, they just extract the important
> parts that descibe it in a simplified way. I don't really understand the
> difference you're trying to make between models and record-keeping. A
> row in a "person" table of a database is just some simplified
> information that describes that person.
>
> The relation model is about how to represent data in general, and any
> specific database is a model of some specific area of interest. So two
> levels of models.
>
> Paul.

It is not a dramatic distinction, but I believe it is real. By the definitions I offered in my OP, the RM is a Model but a database is a system of records. They are not the same thing, and whenever you call a horse a car you can run into trouble. Nothing deeper than that.

-- 
Kenneth Downs
Secure Data Software, Inc.
(Ken)nneth@(Sec)ure(Dat)a(.com)

Received on Tue Apr 26 2005 - 17:53:56 CDT