range error

[Dar Scott]

On Thursday, April 8, 2004, at 09:56 PM, Stephen Messimer wrote:

> Type	Operators *: range error (overflow)
> Object	Bed Adequacy
> Line	else return theNumber * factorial(theNumber -1)
> Hint	factorial
>
> Is there a work around for this?
>
> OSX 10.2.8, Rev 2.1.2, 512 MB of RAM

I have 10.2.8, too, and for many calculations I get 00000Inf.

The "double" floating point numbers used by Revolution use about 11 
bits for the exponent.  A few special codes are used, but that is 
approximately the range.

Try 2^1000 and 2^2000 in the message box.

Try length(2^1000).  Or 169^169.

I'm making a list of what kinds of libraries I might build soon and on 
my list are several variations of enhanced numbers and associated 
arithmetic.  I am interested in what ranges folks would find 
interesting or useful.  I might find something else more fun or 
lucrative, though.

As for your workaround...  You might look at your math.  If you are 
dividing a factorial by a factorial, you might want to change your 
function to multiply the number n to m instead.  You might convert your 
math to work with the logs of your values.

If you really need extended math you might build your numbers out of 
multiple numbers.

Dar Scott