Re: About Entity Relation Diagram

On 18 Dec 2004 21:36:26 -0800, "Mark D Powell" <Mark.Powell_at_eds.com> wrote:

>I just grabbed the first online reference I could find that would allow
>quick access to the point, since my computer books are all still
>packed away as I just moved on the 6th of Dec. But the older the
>source the better from my point of view.

Because you are wrong and the old sources tend to be worse at this point than the modern ones.

> I see you have an IBM email
>address. My early training was all based on IBM materials.

It is very deprecated to learn the fundamentals using vendor's materials.

Among other things, they always try to hide the product's flaws sacrificing the correctness, or plainly cheating the readers.

> A great
>many programmers seem ignorant of the fact that Dr E. F. Codd created
>relational algebra to describe his new theory of Relational databases
>while he worked for IBM.

But not the readers at this group.

> Later researches at IBM created SQL to be
>used with databases built on Dr Codd's theories.

But many people seem ignorant of the fact the result was a botch-up because they were not faithful to the model, and Dr Codd was very dissapointed.

>But the border between the logical design and the physical design is
>really pretty thin and is likely going to be heavily slanted toward the
>features of the RDBMS in use.

The border is cristal clear, but with the current SQL products the options for physical design are very narrow (specially with the cheaper ones). The most part of the physical design is automatically created by the DBMS and there is no way to change it.

With the current SQL products, the only way to change many aspects of the physical design is to change the logical design, and this is one of the sources of the spreaded confusion.

Regards Received on Sun Dec 19 2004 - 12:12:57 CET