>It's currently > >: */ >r * r> / ; > >? Should not calculate this with double precition? Eg. something like > >xchg eax,[esi + 4] >cdq >mul [esi] >div [esi+4] >add esi, 8 > >? I ask because I currently write on a program that could benefit from the >higher precition. > > As always, there's a balancing act. In RetroForth, I choose to error on the side of simplicity (math is limited to 32-bit signed integers), with the knowledge that some potential drawbacks (lower precision) come as a result. I do want to explore more complex math functionality with TetraForth though and double precision is high on the list. Also unsigned math and fixed point math. All of that will take me quite a while to implement though. For those who need this in RetroForth, here's the definition for you: : */ [ $044687 3, $99 1, $26f7 2, $0476f7 3, $08c683 3, ] ; -- Charles