[Coco] Re: CoCo gcc project

[KnudsenMJ at aol.com]

In a message dated 11/2/03 9:04:09 PM Eastern Standard Time, james at skwirl.ca 
writes:

> My main question is, why would we need floating point support?  It 
>  seems to me that fixed point arithmetic would suit the needs of most 
>  applications.  Not that I'm trying to poo poo the idea, I just want to 
>  understand what application it would be required for.

Astronomical simulations.  Even simple animations will overflow all too 
easily in fixed point.  And if you're trying to predict the position of Mars in 
2008, errors can pile up fast.

Anyone remember "Planet Engine", an astronomical predictor program that one 
or two guys wrote in B09 and sold for a few years?  I may have it somewhere.

BTW, B09's floating point numbers are a byte less precise than RSBASIC's, at 
32 versus 40 bits.  C can do 32 or 64 bits, the latter being handy if you time 
travel back to King Arthur's day and need to time an eclipse just right :-)   
I think we could get my on 32 bits for now. --Mike K.