>>>It would be interesting to see implementations of this model if anyone is intrigued enough to post their take. (I'd especially like to see Calvin's stab at it as he's looking for more interesting uses for VFP :-)
>>
>>It seems to me that the main challenges - as long as you stay with Newtonian mechanics - is in the presentation. UI, letting the user select planet sizes, problems with the scales (perhaps, the user should be able to select both time and distance scales), etc.
>
>Presentation and performance, of course.
>
>How long should it take the program to model the revolution of the earth around the sun? We know in reality it takes 365 days, but it would be nice if the program showed us the orbit in alot shorter time than that! Being able to derive 365 days from the time the model actually took to run is what was meant in my original post by calculating how much time has elasped. So your part in paranthesis about selecting time scales is exactly what I was getting at there.

I think the following should be user-selectable options:

(1) At what intervals are the velocities recalculated. Recalculating every second will give greater accuracy than recalculating every day; this is more relevant when an object passes close to another one, where the gravity vector changes from one moment to another. And, of course, recalculating more often will make the program slower.

(2) Acceleration factor for display: for example, for the planets' movement, 1 day per second, or even several days per second, might be appropriate.

Apart from that, it might be interesting to have options that change x- and y-coordinates of the screen center, as well as the zoom factor, dynamically (that is, while the animation works). This would allow me (the viewer) to concentrate on the inner planets, or have a more ample view, of the entire Solar System.