Re: Why called diffraction?

["John S. Denker" <jsd@MONMOUTH.COM>]

On 04/24/2003 04:14 PM, John Mallinckrodt wrote:
> ... "single slit diffraction," "edge diffraction," or "circular
> aperture diffraction."  Nothing that I would want very much to call
> "scattering" goes on in those cases and they seem to me to be pretty
> much the poster children for the phenomena called "diffraction."

Point taken.

Without disagreeing with the above, I would add
diffraction gratings and crystal Xray diffraction
patterns to the set of poster children.

So, how about this:
  diffraction is defined to be:
    a change in direction of a wave that cannot
    be explained in terms of ray optics, requiring
    instead physical optics.
    (And sometimes including the raylike zeroth-
    order diffracted beam as a trivial case but
    a natural generalization.  Call it zeroth-order
    diffracted or undiffracted; I don't care.)

=============

> Of course, all "diffraction" and "interference" phenomena (as well as
> "reflection" and "refraction" phenomena for that matter)

and everything else in wave mechanics

> can be
> considered to be the result of "Interference" (with a capital I).

Right.  We agree this is a truism.  Hence it has
no discriminatory power, and mentioning interference
in a definition, as I tried to do earlier, is pretty
much a waste.


> I would say that "interference" is what happens when you
> combine waves from a finite number of sources and diffraction is what
> happens when you combine waves from an "infinite" number of sources.

Suppose I have two pointlike scatterers.
Delta function potentials.  Born approximation
and all that.

To my taste, I would like to call the result
a diffraction pattern.  The fact that two is
a finite number does not discourage me at all.

If pressed, I think I could be persuaded to
call the scattering from a _single_ pointlike
scatterer diffraction, although I'd prefer to
just call it scattering.

> Then, perhaps we may be forgiven for considering gratings and
> crystals to produce "diffraction" effects if only because the number
> of waves being combined, while finite, is certainly "very large."

I don't think the numerousness of the lines in
the grating that is the key idea.  I think the
_spacing_ between the lines is what makes it a
diffraction grating.  Eight or ten lines is enough
to do a pretty good job (although not maximally
convenient to set up).