Re: What is that "more" that makes E-R model truly independent ?

"paul c" <toledobythesea_at_oohay.ac> wrote in message news:B4Lpi.6049$rX4.4385_at_pd7urf2no...
> David Cressey wrote:
> > "beginner16" <kaja_love160_at_yahoo.com> wrote in message
> > news:1185291874.325737.75510_at_w3g2000hsg.googlegroups.com...

> This reminds me of a quote from another book, "Logic" by Wilfrid
> Hodges, which Hugh Darwen has mentioned from time to time:
>
> "The task we now approach is that of *formalizing* logic. To formalize
> is to strip away the concepts which give meaning and application to the
> subject, so that nothing remains but bare symbols. Translation will
> disappear first, situations will fade away next, and finally even truth
> will make its exit."
>
> I think this means that a dbms is a logic machine. When the machine
> gives us its answer to a question, consisting of "nothing but symbols",
> can we then look to a conceptual model to tell us what meaning to take
> from the answer? If so, does it happen much in practice?

Good quote, P.

If you'll grant me one proviso, I'm going to claim that it happens nearly always in practice. The proviso I'm asking you to grant is that if a user has merely an intuitive grasp of "what the data really means", that means that the user has a conceptual model of the data in his/her brain, even if it might be at an unconscious or subverbal level.

What happens a great deal in practice, in my experience, is that different users (and other stakeholders) have mutually contradictory conceptual models of the data, and what it means in the real world. They may or may not be aware of these contradictions. And a database cum application that may be working "perfectly" from the point of view of abstract mathematics may in fact be delivering a great deal of negative value to the enterprise because of the misunderstandings between people who affect the data or are affected by it.

Most of the regulars in this newsgroup prefer not to deal with such issues. That's fine, in theory. In practice, it is not. Received on Wed Jul 25 2007 - 19:43:44 CEST