[KH]

On 3/20/2015 6:53 AM, wrote:
On Thursday, March 19, 2015 at 8:02:33 AM UTC-6, Robert Peirce wrote:
The point was that no scientific theory is necessarily true. In fact,
to be a scientific theory the possibility must exist to prove it false.
It may in fact be true, in which case it never will be proved false.

I cannot cite papers or other evidence for what follows. In many cases
it comes from something I read, possibly years ago. For that reason it
could be wrong. Consequently, it should be taken for the idea, not the
details.

snip

Yes to be a theory, it has to walk like a theory, talk like a
theory and act like a theory. Hanging a "Theory" sign around a
an incoherently babbling immobile drunkard is not less of a
strawman than insisting on Nyquists theorem being a "theory".

Very true, and Nyquist/Shannon comes under the category of theorem, not
theory, and it is easily proven mathematically.

Many, many implementations can be based on Nyquist, and fail to achieve
sonic accuracy - and fail miserably. But that's an engineering failure,
not a falsification of the underlying Theorem. If the theorem says that
samples taken at 1/2 fz intervals are sufficient to reproduce the
waveform - no "missing data" no "stair steps", but some real world
implementation of digital based on that theorem fails to achieve the
expected accuracy, that says nothing about "digital" per se, nor about
the general theorem.

Only *after* the engineering methods used for that implementation are
shown to be sufficiently rigorous and accurate do the results have a
wider applicability.

Your real world example of a pathological DAC is the poster child.

Keith