sourcedemos/julia/julia.xtl
1⍝!/usr/bin/env xetal
2⍝# Julia sets and the Mandelbrot set from one function, u:i_terate.
3⍝# Both iterate z becoming z * z + c and count the steps z stays
4⍝# within 2 of the origin. For the Mandelbrot set, c is the grid of
5⍝# points and z starts at 0; for a Julia set, z starts at the grid and
6⍝# c is one number, which scalar extension spreads over the whole
7⍝# grid: the step is written once and works both ways. X_eTaL has no
8⍝# complex numbers yet, so z and c are each a real and an imaginary
9⍝# part (see the demo's README Workarounds). Run it with "xetal run
10⍝# demos/julia/julia.xtl"; it prints the grid's shape, a
11⍝# Julia set (c = -0.8 + 0.156i) drawn with characters, and the
12⍝# Mandelbrot set from the same function. The live page lets you pick
13⍝# c on the Mandelbrot map and watch the Julia set change:
14⍝# https://softwarewrighter.github.io/X_eTaL-demos/julia/
15
16⍝## The view
17
18⍝# rows, cols: the picture's size. cx, cy, w: the view's center and
19⍝# width. k: how many steps. aspect: how tall a cell is for its width
20⍝# (2.0 for terminal characters).
21rows ← 22
22cols ← 64
23cx ← 0.0
24cy ← 0.0
25w ← 3.2
26k ← 40
27aspect ← 2.0
28
29⍝# Numbers centered on 0 (u:c_entred n): 0 1 2 3 becomes
30⍝# -1.5 -0.5 0.5 1.5.
31ᵘc̲entred ← { n → (f̲loat o̲ffsets n) − (f̲loat n − 1) ÷ 2 }
32
33⍝## The grid of points, by broadcasting
34
35⍝# re, im: one row of real parts and one column of imaginary parts.
36⍝# grid: both parts spread over the whole grid with t_able, as a pair
37⍝# (gr, gi): every point's c, for the Mandelbrot set, or every
38⍝# starting z0, for a Julia set.
39re ← cx + w × (ᵘc̲entred cols) ÷ f̲loat cols
40im ← cy − aspect × w × (ᵘc̲entred rows) ÷ f̲loat cols
41grid ← ((o̲ffsets rows) 'r̲ight t̲able re, im 'l̲eft t̲able o̲ffsets cols)
42(gr, gi) ← grid
43
44⍝## c u:i_terate z0: k steps of z * z + c from z0
45
46⍝# c and z0 are each a pair (real, imaginary): two numbers (scalar
47⍝# extension) or two planes of the grid, taken apart by the patterns
48⍝# on both sides (D7, landed). The state adds a count of steps spent
49⍝# inside; a point outside keeps its z.
50ᵘi̲terate ← { (cr, ci) (zr0, zi0) →
51 ⍝ One step, as a local function (s_tep, bare and scoped to this
52 ⍝ lambda, PN6): the same arithmetic as mandelbrot.xtl's u:s_tep.
53 s̲tep ← { (zr, zi, counts) →
54 inside ← f̲loat 4 ≥ (zr × zr) + zi × zi
55 a ← (inside × cr + (zr × zr) − zi × zi) + (1 − inside) × zr
56 b ← (inside × ci + 2 × zr × zi) + (1 − inside) × zi
57 (a, b, inside + counts)
58 }
59 k 's̲tep p̲ower (zr0, zi0, 0.0 × zr0)
60}
61⍝ -- end of the core ---------------------------------------------------
62
63s̲hape gr
64
65⍝# julia: a Julia set, c = -0.8 + 0.156i the same for every point, z0
66⍝# the grid.
67(jzr, jzi, julia) ← (-0.8, 0.156) ᵘi̲terate grid
68s̲hape julia
69(1 + f̲loor (9.0 × julia) ÷ f̲loat k) s̲elect " .:-=+*#%@"
70
71⍝# mandel: the Mandelbrot set from the same function: c the grid,
72⍝# z0 = 0.
73(mzr, mzi, mandel) ← grid ᵘi̲terate (0.0 × gr, 0.0 × gi)
74(1 + f̲loor (9.0 × mandel) ÷ f̲loat k) s̲elect " .:-=+*#%@"