AW: Harmonix content formula

[Martin Czech]

> >Sometimes numerical solvers are completely wrong (not in the FFT case).
> >So using numerical methods requires that you allready know how
> >the result will look like in order to reject wrong "solutions".
> 
> -hmm, couldn't we even argue that numerical solutions are always, one some
> level, "wrong", because they necessarily involve approximations and, being
> done in computers, finite-resolution numbers? i wonder how much progress has
> been made in getting computers to perform symbolic (and hence truly
> "correct") operations on the level of fourier transforms.. be interesting.. 
> 
Of course numerical solutions are only aproximations.
What I meant is absolutely wrong, going into the wrong direction,
not even close to the real solution, not even of the same shape.
This can happen quite often.

m.c.