Re: A Normalization Question

From: Marshall Spight <mspight_at_dnai.com>
Date: Fri, 09 Jul 2004 16:16:01 GMT
Message-ID: <5xzHc.23197$WX.21006@attbi_s51>

"VHarris001" <vharris001_at_aol.com> wrote in message news:20040709090846.29796.00001159_at_mb-m18.aol.com...

> For instance, one constant aggravation is the lack of integration between front
> office and back office operations. Why no integration? Because they were
> written as standalone modules, all containing redundant information. If the
> database program itself forced the designer to point to one and only one unique
> instance of a thing, the integration of back and front office would have
> happened automatically as system layers were added.

This issue has nothing to do with the logical model of data used in the front office and the back office. It's the result of the fact that the two DBMSs do not exchange data. Your solution assumes there is only one DBMS between them; if that situation had existed in your earlier example, there would be no problem either.

> The method of normalizing data by adding more tables with greater degrees of
> detail seems to me to be the very source of the lack of integration problem.

Not at all; the source of the problem is lack of cooperation among federated DBMSs. If you have federated DBMSs that don't cooperate, then you'll still have the problem even if you use Neo's model.

Marshall Received on Fri Jul 09 2004 - 11:16:01 CDT