decfloat -- again

[Pete]

rational arithmetic has it's place, but the fractions quite often become really huge

They do, indeed. Of course you can always waste a lot of time doing GCF iterations to reduce those fractions, but I found it is faster with string math to just let those huge fractions grow. That is especially true when you get complex fractions that hardly reduce at all, but a lot of cpu time is spent trying to see if they can be reduced. At the end of a routine, the numerator is simply divided by the denominator to display a decimal representation, so it really doesn't matter if even the end resulting fraction gets reduced. For ongoing calculations, that might slow things down to a point where a GFC reduction should be utilized.

Pete