Le 05/04/2025 à 23:09, William Peckham via Dng a écrit :
> I regularly coded in FORTH up until a couple of decades ago.  RPN
> makes perfect sense at that level, and in a calculator, if you think
> in terms of stacks.  It is not only fast and efficient, it encourages
> clear planning and thinking about the data.

    I understand it. When I was a student, at some stage, we had
exercises (and controls) of numerical calculations. It necessitated
either the logarithmic tables or the slide rule. It was necessary to
spread the steps of the algorithm on a , well, spread sheet. We didn't
use the pocket calculators yet, but RPN pocket calculators would
certainly have goten their place in this framework, and they probably
got it few years later. However the calculus to be done was presented in
the form of a mathematical formulas, and the job to be done was what,
today, we ask the computer to do.

    Although I don't know it, I bet the algorithm used to convert any
mathematical formula to an RPN algorithm is the same in all compilers of
all languages implementing formulas, and I doubt it is covered by 
copyright. I don't claim, though, that RPN calculation in general, and
FORTH in particular aren't an interesting exercise.

--     Didier