On Mon, Jul 09, 2012 at 01:36:11PM -0700, ddd mmm wrote:
> With 64-bit systems, is there any temptation to represent numbers
> differently in picolisp, using a single 64-bit machine word?
Yes, this is indeed the case.
> It's my understanding that Lua does this, using floats for both floats
> and integers. Would this violate something about picolisp that must be
> preserved? If not, would it make picolisp much faster for certain
> operations and therefore make it more versatile in its applications?
PicoLisp doesn't use floats, but the 64-bit version supports separate
encodings for bignums and short numbers. With that, arithmetic
operations on small numbers (up to 61 bits including the sign bit, which
are by far the most common operations) are a lot faster.