(48G 49G 50G) Runge-Kutta Diff Eqn solver

[Giuseppe Donnini]

(06-14-2021 10:24 PM)rcthacker Wrote:  I doubt the firmware writers designed the input arguments to allow for this.

Before commenting on the intentions of the firmware designers, I would first listen to what the very head of the firmware design team, Dr. William Wickes, has to say about it: on page 415 of his HP 48 Insights, Part II: Problem Solving Ressources (of which a free copy is available here), we may read the following:

A nice property of HP 48 Solve is that the current “equation” does not literally have to be an equation. You can use programs, expressions, and equations almost interchangeably for solving purposes. In effect, HP Solve always solves f(x) = 0. In the case of an HP 48 algebraic expression object, f(x) is the expression represented by the object, where x is the name of the unknown variable. For an equation object representing g(x) = h(x), HP Solve solves f(x) = g(x)-h(x) = 0. For a program, f(x) is the expression that is equivalent to that (RPN) program. Programs used for this purpose must be equivalent to an algebraic expression, taking no arguments from the stack and returning one real number or unit object.

And an equivalent, albeit shorter, explanation is already given on page 97 of his HP-28 Insights, which you claim you have read.

So, yes, this is indeed a central characteristic of the RPL design.