Sorry, I need a good measure of both, and was thinking I could just use the SINS number conversion methods for binary files, since my mesh-generator uses numbers in the same range as SINS for the "Save-Place" file...

My Mesh-output already truncates to 6 places (4 really, for cleaner, smaller files), but it would be for saving numbers exactly:

If I remember my CSE classes correctly, it was hotly debated which bytes should save data on which side of the decimal point, and how many bytes to devote to storing a number exactly (and where to fudge when things don't fit)

I was thinking two bytes on either side of the decimal would do it, since I only use four decimal places precision, but I could use three bytes in front so the numbers could be large enough to handle certain things...

I'm looking for a parsed format since I have to write the program to un-parse and reconstruct the numbers, so I can reload my saved-data file directly into my program...

Example: 65535.9999

byte1=int(int(number)/256)

byte2=int(number)-(byte1*256)

temp=number-((byte1*256)+byte2)

byte3=int(temp*256)

byte4=((temp*256)-byte3)*256

Something like this is what I'm looking for...