Re: [boost] Constrained genericity at last? (any interest in an operator traits/concept traits library?)

From: Daryle Walker (darylew_at_[hidden])
Date: 2004-10-02 19:29:47

On 9/25/04 6:16 AM, "Victor A. Wagner Jr." <vawjr_at_[hidden]> wrote:

[SNIP]
> I don't mean to pick on you in particular.  I just got tired of hearing
> people repeat the "mantra" of "you can subtract two pointers, but not add
> them".  It's simply NOT true.

Actually, it IS true... From a certain point of view.

That's because pointers are NOT (conceptually) numbers, they're LOCATIONS!
Addition among them makes no sense.

Technically, subtraction between them would be equally senseless.  But
memory locations are treated as being linearly arranged in an order.
Therefore we can define subtraction as the OFFSET one location is linearly
from another.

So when you take the midpoint of some pointers...

    T  *a, *b, *c;

..you're taking the average of their offsets (from a defined origin).

    T * const  base = a;
    T * const  average = base + ((a - base) + (b - base) + (c - base)) / 3;

> It has also been somewhat of a disappointment that the HLL folks (and C in
> particular (you can program right down to the sand) have managed to ignore
> this)
[TRUNCATE]

Or maybe they didn't ignore it to the point of an arbitrary mantra, but had
a rationale for not treating pointers as simple numbers.

-- 
Daryle Walker
Mac, Internet, and Video Game Junkie
darylew AT hotmail DOT com