Re: Entity vs. Table

"Alfredo Novoa" <alfredo_at_ncs.es> wrote in message news:40d164f6.538133_at_news.wanadoo.es...
> On Tue, 15 Jun 2004 09:23:35 -0400, "Alan" <alan_at_erols.com> wrote:
>
> >> The Relational Model says nothing about the physical level.
> >
> >I am talking about taking a completed relational model design and
> >implementing it. You know, I've designed all the tables and now I'm
typing
> >SQL to build them in the RDBMS.
>
> SQL mixes the logical and the physical levels, but the most part of
> the physical design is created by the SQL DBMS automatically.
>
> It is perfectly valid to have redundancy in the physical tables.

Yes, if you denormalize to less than 3NF. This is done often for performance purposes, or ease of reporting, as well as some other reasons. But the argument was about 3NF and redundancy.

>
> >> But where is the rule?
> >>
> >> A double oval says that an attribute is derived but it says nothing
> >> about the derivation rules.
> >>
> >> You are losing the business requirements.
> >
> >No one said that the ERD models ALL business requirements.
>
> But you said that it can represent most of them, what is clearly
> false.

So you keep saying. That doesn't make it so.

>
> > For the 50th
> >time, it models relationships among the data.
>
> Only a very little part of the requirements, and it does that worse
> than The Relational Model.
>
> There are not precise definitions for "entity" and "relationship". The
> distinction is arbitrary. The ERD is a backward step and heavily
> oversold.
>

I guess the often ranked number 1 university in the U.S. (by U.S. News & World Report) in I.S. grad programs is wrong and you are right. Use whatever tools you want to use, just leave me alone. Received on Thu Jun 17 2004 - 15:17:50 CEST