Re: Normalization, Natural Keys, Surrogate Keys

Ed Prochak <edprochak_at_magicinterface.com> wrote in message news:<3CF7812F.3CB96FA0_at_magicinterface.com>...
> And with that training do you at least see that you transformed a
> relational database into a Network model database? It is not really
> simpler.

How have I turned it into a network model? First of all, I don't believe there is that much difference between the relational and network models. From all I can remember they both have 1-M and 1-1 relationships. If I recall the only real differnce is that in a relational DBMS I can create my own keys whereas I cannot in a heirarchical. Also a relational DBMS cannot create a network database and vice-versa. Maybe you could explain it to me.

> And when you get into problems with duplicates and other data
> integrity issues, you have to go back to those compound keys to solve
> it. But trust me, it won't be pretty.

Well, so far I haven't had to do this. As of this writing I have developed or guided the development of many enterprise-scale databases which all seem to be humming along quite nicely.

Of course, as I stated earlier, I don't design the database in a vacumn. When I do a database design I'm doing so with several factors in mind - performance, maintainablity, programablity, report generation, and ease of upgrades. Since I am a project leader, it is up to me to ensure that the database designed for a given applicaiton is suitable. I must also ensure that my programmers are able to easily work with the database given the requirements.

Now, I don't believe we should never have a unique constraint or that natural keys should never be used. For a bank application or a student application, I would absolutely ensure that something about my entity (a student, an account, a transcript, whatever) is unique.

Ahh, I'm too tired. It's friday and I've got a serious case of paint fume inhalation from remodelling my guest bathroom.

Have a great weekend all!

Go USA! (At 4:30 in the freakin morning..)

KP Received on Fri May 31 2002 - 17:13:01 CDT