Re: ROCKET ERROR

[brian whatcott <inet@INTELLISYS.NET>]

At 11:39 12/29/99 -0500, David Bowman wrote:

> Regarding Ludwik's recently reported results for a rocket problem:
> > ...

[Ludwik]
> > I did this and the new results are in very
> > good agreement with the theory...

[David]
> I'm sorry to be so pedantic here, but your results *are* the theory
...
> Numerical simulations are not compared to theories in the sense that
> experimental results are compared to theories (i.e. as tests of the
> theories, the experimental results, or some combination of both).
> Rather, they are (to the extent the numerical techniques and algorithms
> are reliable solution techniques) predictions *of* those theories....

> David Bowman

That's what I like best about Ludwik.
He may take a lickin' but he keeps on tickin'   :-)
(to coin a Timex tag).

Persisting through critiques is a quality I associate with
some of the greatest investigators.

On the question of numerical simulations, I admit to hungering for
the time delta as the iterated quantum rather than efflux mass in
Ludwik's model.
 The assumption was a constant mass flow rate so this was I suppose
an implicit simulation in time.

On the question of simulations - can they ever said to be in
agreement with theory? - I recall in bulletin board days a person
would sometimes post puzzles of conditional probability.

 Here's one in the general format.
 "I add 6 blue marbles to a barrel containing 200 red marbles and 15
yellow ones. What are my chances of pulling at least one yellow marble
and no blue ones in three tries, selecting three at a time, and
throwing them back?

There was of course a formal approach. But I persisted in running
Monte Carlo trials to obtain an answer to the nearest tenth percent.
  I suspect that here I was not mechanising any theory of conditional
probability - I was running an actual experiment.




brian whatcott <inet@intellisys.net>
Altus OK