Why aren't large numbers limited to 64 bit values?

[Geoff Canyon]

On Thu, Dec 10, 2015 at 1:46 PM, Peter TB Brett <peter.brett at livecode.com>
wrote:

> When numbers are too large to represent exactly, LiveCode automatically
> shifts to using 64-bit IEEE floating point representation.  Floating-point
> numbers store a number in three pieces:
>
> 1) a sign bit
> 2) an exponent (11 bits)
> 3) a mantissa (52 bits)
>

​Ah, thanks, I'd forgotten about the division between the mantissa and the
exponent. So any calculation up to about 10^308 should be of the right
order of magnitude, but beyond that all bets are off.