User blog comment:Dr. Satl, M.D./Petit Computer 3 Update?/@comment-14600979-20130928232156/@comment-1595298-20130929233244

I wouldn't be so sure about that. It's entirely possible to read in decimal digits and convert directly to fixed point, and I have a feeling that might be exactly what they do. A little test I performed just now seems to imply that they only handle 6 digits after the decimal point, and ignore the rest.

The smallest positive number representable in their fixed-point format is 1/4096, which in decimal is 0.000244140625. However, if I type in that entire decimal value to Petit Computer, I actually get 0. It doesn't give me something greater than 0 until I increase the number to 0.000245.

Besides, even if they did use floating point format to read in values, they would need to use double precision to have the 32 bits of precision necessary for their fixed-point format. I guess I can't really say for sure either way unless I could see their source code, though.