Re: Can relvars be dissymetrically decomposed? (vadim and x insight demanded on that subject)

Tony D wrote:
> Bob Badour wrote:
> > Your last statement above overlooks logical independence. A relvar is
> > not a piece of memory, and it can be a mathematical, logical,
> > referentially-transparent-placeholder-shorthand-for-an-expression
> > variable. What, after all, is a view if not a relvar and not a
> > mathematical, logical,
> > referentially-transparent-placeholder-shorthand-for-an-expression variable?
> >
>
> Fair point. I forgot that, in the Introduction to Database Systems at
> least (I don't have TTM here), relvars are noted as coming in two
> varieties.
>
> > Since algebra.
> >
>
> Since I was in the context of a 3GL, I didn't think this objection
> would come up. Silly of me. I would amend that sentence then to read
> "Since when was '+' defined over memory references ?"
> (Casts, BCPL and C-derived lunacy excepted. "*s++" indeed.)
Yes. I would add a word of caution which is that abstract thinking about relations should avoid decutiveness based on implementation. As a rule of thumb, I am not sure about some point in RM, I always get back to math. Received on Tue Jul 18 2006 - 12:16:14 CEST