Re: Towards a definition of atomic

"Jan Hidders" <hidders_at_gmail.com> wrote in message news:3cdcbc6b-cdc6-47dc-8af7-4d0e1ed10839_at_j78g2000hsd.googlegroups.com...

(quote)
On 1 feb, 19:55, Marshall <marshall.spi..._at_gmail.com> wrote:
>
>
> The concept we are running into here is the same as "alpha
> equivalence"
> from lambda calculus.
>
> Consider these two programs:
>
> Program 1:
> let x=5;
> let y=3;
> x+y;
>
> Program 2:
> let a=5;
> let b=3;
> a+b;
>
> Are these programs identical? It depends on what we mean by
> "identical." Certainly they calculate the same value, but certainly
> that's not enough to consider them identical.
>
> If we consider the two programs as strings of symbols, then
> obviously they are different. However clearly they are very closely
> related, and we can formalize this perception, and this formalization
> is called alpha equivalence. Roughly, the two programs are alpha
> equivalent because they have identical structure if we ignore the
> specific choices of variable names. From this perspective, there is
> no significance to the specific choice of "x" or "a" or whatever as
> variable names, but that doesn't mean the names are meaningless,
> because the distinct names encode identity within the structure
> they are embedded in.

This is the usual explanation for why and how abstract identifiers can have meaning. So does that mean you think that abstract identifiers can sometimes actually have meaning and be useful? (end quote)

I would make a small change in the wording above.

Instead of "can have meaning" I would have said, "can convey meaning". The variable named "a" or "x" has no meaning because of its name (unless the name has mnemonic value which, in this case, it doesn't).

However, if an observer somehow attaches meaning to the variable "a" in one of its uses, that meaning can then be transferred to all of the uses of the same named variable (within the scope of the name). Received on Sun Feb 03 2008 - 14:50:33 CET