Randomize a List problem

[Albert Chan]

(09-17-2020 12:20 PM)John Keith Wrote:  A few more random observations about the original (incorrect) RANL program:

The first iteration does nothing because the argument for ROLLD can never be greater than 1.
The program would run faster starting 2 t FOR n...

Your optimization works, as long as the list has more than 1 item.
With a starting of 2, this meant it is possible to hit with 2 ROLLD, even if t=1

For most other programming languages, this is not a problem.
However, HP-50g has a FOR loop quirk, that *always* run the loop at least once.

≪ 2. 1. FOR n n NEXT ≫ EVAL       → 2.

To see the actual problem, lets save RANL "optimization" under 'RANL2'

≪ LIST➝ ➝ t
≪ 2. t FOR n n RAND × CEIL ROLLD NEXT t ➝LIST ≫ ≫

Starting with an empty stack:

{10.}
RANL2       → "ROLLD Error: Too few Arguments", now stack = 10, 2
{20.}
RANL2       → 10., 2., {20.} ok
RANL2       → 10. 20., {2.} ???