Re: Natural keys vs Aritficial Keys

From: David BL <>
Date: Fri, 29 May 2009 01:07:11 -0700 (PDT)
Message-ID: <>

On May 29, 11:18 am, paul c <> wrote:
> David BL wrote:
> > On May 29, 1:27 am, paul c <> wrote:
> >...
> >> If he (or anybody else) is
> >> proposing that a tuple programming and user interface is superior to a
> >> relation interface then they will have to go quite a few steps beyond
> >> addressing, they will have to suggest some tuple operators and why they
> >> would be better, more useful, et cetera. Would such operators replace
> >> relational ones or just augment them? If the latter, why would you need
> >> both?
> > Who is proposing any of that? Am I "he" in the above?
> Your words:
> "As I said, there is nothing implicitly wrong with a tuple-variable ..."

You dispute that? Would you say for example that a variable shouldn't record a (x,y) position?

> and
> "I find the concept of a "sub-variable" of a given variable
> interesting. For example, given a tuple-variable, there are sub-
> variables for each of the attributes. However with other ways of
> composing data structures, sub-variables may not exist. A relvar is
> one such example. ..."
> That is pretty mysterious stuff, tuple-variables have sub-variables,
> relvars may or not have sub-variables, relvars are structures (not
> logical pointers presumably?) maybe that comes from thinking about C++
> too much. (Nothing wrong with C++ in the riight context, but this isn't
> one.)

I thought the subject matter (which is the first great blunder) concerned the interplay between a database and some application that accesses the database. Many such applications are written in an imperative style. In fact, an imperative style is implicit in the instruction to perform an assignment to a variable, and a relational database is regarded as basically a set of variables that record relations. Whether an application be in C++ or not is hardly relevant.

I'm guessing you think the subject matter only concerns data modelling aspects. If that is the case I think you're confused about the meaning of the first great blunder.

I'm a little surprised that you find the concept of a sub-variable mysterious. By definition, if x,y are variables (i.e. they hold values) then x is a called a sub-variable of y if there exists a functional dependency of the value of x on the value of y.

I have no idea what you mean by "relvars are structures (not logical pointers presumably?". Please quote me rather than pretend to quote me. Received on Fri May 29 2009 - 10:07:11 CEST

Original text of this message