Re: [Phys-l] computational physics activities (was: universal gravitation ....)

[Rick Tarara <rtarara@saintmarys.edu>]

As someone who has written 'lab simulation' software, let me defend such (to 
a point).  First, let me emphasize that what I've written was not initially 
designed to substitute for live labs--rather the goal was to provide 
pre/post lab tools with the possibility of doing a fully simulated lab with 
students who 'missed' lab.  OTOH, the number downloads of these simulations 
to countries where one can assume a lack of funds for equipment (and 
comments back to me) suggests that many ARE using them as 'lab experiences'.

In my mind, the key to keeping the simulations semi-true to life involve two 
things.  1)  The software provides animations and data collection that VERY 
CLOSELY mimic the live experiments.  That is, the computer does NOT collect 
or record the data and the lab simulation programs do no analyze the data 
(although often might need to transfer collected data--from lab book--back 
into a spreadsheet program for full analysis.)  So, the simulation is ONLY 
of the physical experiment and most of the measurements--(timers and 
sometimes length measurements).  To be sure, I don't simulate 'weighing' 
exercises (although that too could be part of the programs if desired).
2)  There are built in uncertainties in the data generated.  That is, I 
build in some random uncertainties in the values that will be measured. 
While it is not really possible to get a systematic error (again I could 
have built in that possibility, but these programs are not really intended 
for science/engineering students--at least not as fully simulated labs), the 
exact same starting parameters for any trial will result in slightly 
different values.

The point being that through design, one CAN incorporate _almost_ as much of 
the 'live lab' as desired.  That being said, I strongly recommend against 
using software simulations for any lab experiment IF that experiment can be 
done (in some form) with actual equipment.

Clearly, exercises like the lunar lander cannot actually be done and if that 
kind of exercise is deemed useful--go for it with software.  When I had 
three semesters with our Chem majors (now only two), I used to have them 
write a program to animate the motion of 3-5 free charges (motion in 2 
dimensions) teaching enough TrueBasic to accomplish this (a two page 
tutorial was enough).  That exercise IS useful in having them break the 
physics (vector forces, accelerations, velocities, displacements) down into 
programmable chunks.

Rick

***************************
Richard W. Tarara
Professor of Physics
Saint Mary's College
Notre Dame, IN
rtarara@saintmarys.edu
******************************
Free Physics Software
PC & Mac
www.saintmarys.edu/~rtarara/software.html
*******************************

----- Original Message ----- 
From: "John Denker" <jsd@av8n.com>
> You will notice I've been speaking of computational physics
> "activities."  I just can't bring myself to call them "labs".
>
> I'm a card-carrying (and scar-carrying) experimentalist, and to me
> there is something sacred about a real physics lab.  That's where
> the truth comes from, or as near to truth as we know how to get.
>
> Computers can be used to advantage IN CONJUNCTION WITH real
> honest-to-goodness lab work ... but when they are used instead
> of lab work, something indescribably important is lost.
>
> If *I* am modeling physics on a computer, I will make the physics
> come out right, because that's what I do.  Alas if you turn kids
> loose on a computer, they generally don't /want/ the physics to
> come out right.  Usually they would rather have a transporter-room
> that zaps them directly up to the moon, without having to mess
> with complicated physics like universal gravitation.  The computer
> is too powerful.  It is too unconstrained.
>
> The computer allows fiction.  There is no fiction in a real
> lab.