Re: Natural keys vs Aritficial Keys

On May 30, 3:12 am, paul c <toledobythe..._at_oohay.ac> wrote:
> David BL wrote:
>
> ...
>
> > When you said the idea of sub-variables is mysterious, I thought
> > that's great because my main interest in posting was to see whether
> > the idea made sense and I wanted to have a discussion about it. I
> > thought the functional dependency idea was quite interesting, and I'm
> > curious to understand how it relates to the concept of a derived
> > relvar. Unfortunately you don't seem interested in engaging in a
> > conversation. Please, I would rather if you did.
>
> I'm more interested in progess than conversation. The first thing you
> need to do is reduce your vocabulary so that it's clear what you mean,
> best to stick with a relational vocabulary to start with (and I don't
> mean that silly dictionary that used to get updated here, growing like
> topsy, just a crutch for people too lazy to think). As for the above,
> functional dependencies have to do with relations

Granted, you are saying that I shouldn't talk about functional dependency between variables because the relational community have reserved the generic term "functional dependency" for something specific to do with relations. That's fair enough, but I'm not mixing up concepts. I'm well aware of the distinction.

What about the following definition: Let x and y be variables (in the sense described by C.Date). x is said to be /derived/ from y if there exists a function that at all times maps the value held by y to the value held by x.

> , relvars have to do
> with language implementation. Not all languages require relvars, eg.,
> functional languages, so as such relvars aren't essential to the minimal
> theory.

Minimal theory of what? A relational database by definition has relvars. Relvars are essential to relational database theory.

The topic of discussion concerns the First Great Blunder which by definition concerns the equation relvar = object class. Received on Sat May 30 2009 - 03:55:08 CEST