Raku: Python:19 extreme math – Steve Roe

Interesting. Tried it in Pharo. With floats it derails at the same point the Python version derailed in the article. But if you change 4.25 to 17/4, Pharo uses rational numbers and it gives a correct result as far as I bothered to try (it bocomes slow after around 35), iteration 38:

x := [ :n | n = 0 ifTrue: [ 4 ] ifFalse: [ n = 1 ifTrue: [ 17/4 ] ifFalse: [ (1500 negated / (x value: n - 2) + 815) / (x value: n - 1) negated + 108 ] ] ]. Transcript clear. 1 to: 38 do: [ :n | Transcript show: n; show:' | '; show: (x value: n) asFloat; cr ].

Results:

1 | 4.25 2 | 4.470588235294118 3 | 4.644736842105263 4 | 4.770538243626063 5 | 4.855700712589074 6 | 4.910847499082793 7 | 4.945537404123916 8 | 4.966962581762701 9 | 4.980045701355631 10 | 4.987979448478392 11 | 4.992770288062068 12 | 4.995655891506634 13 | 4.997391268381344 14 | 4.998433943944817 15 | 4.999060071970894 16 | 4.999435937146839 17 | 4.999661524103767 18 | 4.9997969007134175 19 | 4.999878135477931 20 | 4.9999268795046 21 | 4.999956127061158 22 | 4.9999736760057125 23 | 4.999984205520272 24 | 4.999990523282228 25 | 4.99999431395856 26 | 4.999996588371256 27 | 4.9999979530213565 28 | 4.999998771812312 29 | 4.999999263087206 30 | 4.9999995578522585 31 | 4.999999734711332 32 | 4.99999984082679 33 | 4.999999904496072 34 | 4.999999942697642 35 | 4.999999965618585 36 | 4.99999997937115 37 | 4.99999998762269 38 | 4.999999992573614