My old friend the SINC function (sci.math)
==========================================
In article
http://groups.google.nl/group/sci.math/msg/08dff005bdbdeb07?hl=en&
Randy Poe writes:
> Real, physical quantites have uncertainties. That is
> one of the fundamental properties of physics. And it
> isn't just due to quantum considerations. Take an
> average metal bar. It has no exact length, not down
> to the precision of an atomic width or so. There are
> temperature fluctuations and small forces from
> Brownian motion which will cause that bar's length
> to fluctuate. The atoms themselves are in motion which
> is another cause of inexactness.
As Randy describes himself, there are _fluctuations_ of all kind that
will cause x = 0, when conceived as a physical quantity, to fluctuate.
Meaning that x = 0 is actually to be _interpreted_ as x = 0 +/- delta,
where delta is the uncertainity.
But we don't need Randy's argument in order to see that there is kind
of an uncertainity in _every_ realization of the real numbers. Take a
look at the floating point quantities in a digital computer:
0.000000000000000000000000000000123597059137504570...
|----------------------------||---------------------> oo
material ideal
Then, inside the computer at hand, _this_ value of x cannot possibly be
distinguished from x = 0. But suppose we buy a somewhat better computer
with "extended precision", such that:
0.000000000000000000000000000000123597059137504570...
|-------------------------------------||------------> oo
material ideal
Then suddenly the value which was supposedly equal to zero becomes just
_close_ to 0 instead. We conclude that, in the material world, a real 0
cannot be distinguished from its (supposedly very small) environment.
Now consider an old friend of mine, the function:
sinc(x) = sin(x)/x for x <> 0
= 1 for x = 0
But NO ! We insist upon our Enduring Freedom and, as _mathematicians_,
we like to define, instead, the following:
S(x) = sin(x)/x for x <> 0
= 2 for x = 0
So I have the question: which of the above functions, S(x) or sinc(x),
has an empirical counterpart? Is it possible that S(x) will _ever_ have
an empirical counterpart? Just answer the question, please. Yes or No.
Jean-Claude Arbaut, Jiri Lebl and Randy Poe all have been clever enough
not to give a straight answer to this simple question.
Han de Bruijn