Newest at the top
| 2025-03-30 05:54:43 +0200 | <haskellbridge> | <thirdofmay18081814goya> neat thanks for all the references! |
| 2025-03-30 05:54:28 +0200 | <EvanR> | now there's ghcjs so you can use whatever is in the browser |
| 2025-03-30 05:54:07 +0200 | <EvanR> | three penny gui was an early web-based thing |
| 2025-03-30 05:53:09 +0200 | <EvanR> | ... and gtk ... |
| 2025-03-30 05:53:02 +0200 | <EvanR> | there are bindings to SDL, glfw |
| 2025-03-30 05:52:40 +0200 | <EvanR> | there's the diagrams library |
| 2025-03-30 05:51:02 +0200 | <geekosaur> | *relatives |
| 2025-03-30 05:50:57 +0200 | <geekosaur> | gloss and relativrd |
| 2025-03-30 05:49:16 +0200 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn |
| 2025-03-30 05:47:48 +0200 | <haskellbridge> | <thirdofmay18081814goya> or, how do people usually do 2D rendering in haskell |
| 2025-03-30 05:47:22 +0200 | <haskellbridge> | <thirdofmay18081814goya> is there a standard wrt rendering 2D in haskell? |
| 2025-03-30 05:47:13 +0200 | <haskellbridge> | <thirdofmay18081814goya> speaking of which |
| 2025-03-30 05:41:32 +0200 | harveypwca | (~harveypwc@2601:246:d080:f6e0:27d6:8cc7:eca9:c46c) |
| 2025-03-30 05:39:06 +0200 | infinity0 | (~infinity0@pwned.gg) infinity0 |
| 2025-03-30 05:38:18 +0200 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 245 seconds) |
| 2025-03-30 05:37:34 +0200 | harveypwca | (~harveypwc@2601:246:d080:f6e0:27d6:8cc7:eca9:c46c) (Read error: Connection reset by peer) |
| 2025-03-30 05:37:29 +0200 | <EvanR> | at which point you stop |
| 2025-03-30 05:37:18 +0200 | <EvanR> | instead of "human eye" test I wonder if you could instead specify the limit of precision by saying lines and shapes have a minimum "pen size" so if you keep trying to draw the mandlebrot set you'll just start getting the same blob over again |
| 2025-03-30 05:36:00 +0200 | infinity0 | (~infinity0@pwned.gg) (Ping timeout: 246 seconds) |
| 2025-03-30 05:33:14 +0200 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) merijn |
| 2025-03-30 05:32:51 +0200 | <EvanR> | (if necessary) |
| 2025-03-30 05:32:08 +0200 | <haskellbridge> | <thirdofmay18081814goya> yeah |
| 2025-03-30 05:31:52 +0200 | <EvanR> | it would be better if you do the approximation at the end of the program than at the beginning |
| 2025-03-30 05:31:42 +0200 | Eoco | (~ian@128.101.131.218) Eoco |
| 2025-03-30 05:31:18 +0200 | <haskellbridge> | <thirdofmay18081814goya> then you can do stuff by function composition |
| 2025-03-30 05:31:14 +0200 | Eoco | (~ian@128.101.131.218) (Quit: WeeChat 4.4.2) |
| 2025-03-30 05:31:04 +0200 | <haskellbridge> | <thirdofmay18081814goya> right, well the point where mathematical and cs-theoretical perfection coincides is to consider "approximate" itself as the circle, instead of "approximate n" for a given n |
| 2025-03-30 05:29:28 +0200 | <EvanR> | straight line segments are really bad at approximating curves btw, at least use parabolic segments |
| 2025-03-30 05:28:41 +0200 | <EvanR> | ultrafinitist confirmed |
| 2025-03-30 05:28:31 +0200 | <haskellbridge> | <Bowuigi> Fractals in general do that AFAIK |
| 2025-03-30 05:28:24 +0200 | <haskellbridge> | <thirdofmay18081814goya> actually, there is an "n :: Natural" such that "approximate" cannot designate a circle that is physically closer to any other approximation (due to the size of atoms, or alternatively your screen resolution) |
| 2025-03-30 05:27:14 +0200 | <EvanR> | (and then it would probably miss a lot in the end) |
| 2025-03-30 05:26:48 +0200 | <EvanR> | that's a good one, the mandlebrot set seems to need an infinite number of steps, to keep adding lobes |
| 2025-03-30 05:26:47 +0200 | <haskellbridge> | <thirdofmay18081814goya> i secretely will use "approximate (100!^100!)" and will tell you it's a circle and you won't be able to tell it's secretly a polygon, so you'll agree with me 😈 |
| 2025-03-30 05:25:22 +0200 | Eoco | (~ian@128.101.131.218) Eoco |
| 2025-03-30 05:24:35 +0200 | <EvanR> | actual circles have been accepted as "geometric" since forever |
| 2025-03-30 05:24:05 +0200 | <EvanR> | a theory of approximations wasn't called for |
| 2025-03-30 05:23:51 +0200 | <haskellbridge> | <thirdofmay18081814goya> then "circle = approximate" :D |
| 2025-03-30 05:23:31 +0200 | <EvanR> | Mandelbrot "set" is actually a Locus so yes |
| 2025-03-30 05:23:28 +0200 | <haskellbridge> | <thirdofmay18081814goya> uh, the contrary, "approximate (n + 1)" has lower error than "approximate n" |
| 2025-03-30 05:22:58 +0200 | <EvanR> | this is gratuitously low quality xD |
| 2025-03-30 05:22:57 +0200 | <geekosaur> | 😈 |
| 2025-03-30 05:22:57 +0200 | <haskellbridge> | <thirdofmay18081814goya> with this idea, you can specify a circle as a function "approximate :: Nat -> CircleApproximation", where "approximate n" approximates the circle more closely than "approximate (n + 1)" |
| 2025-03-30 05:22:53 +0200 | <geekosaur> | are (visualizations of) Mandelbröt sets geometric diagrams? |
| 2025-03-30 05:22:34 +0200 | merijn | (~merijn@host-vr.cgnat-g.v4.dfn.nl) (Ping timeout: 260 seconds) |
| 2025-03-30 05:22:18 +0200 | Eoco | (~ian@128.101.131.218) (Quit: WeeChat 4.4.2) |
| 2025-03-30 05:21:47 +0200 | <haskellbridge> | <thirdofmay18081814goya> EvanR: you can't draw a circle, but you can approximate a circle with abritrary precision (including to an extent indistinguishable to the human eye) |
| 2025-03-30 05:21:20 +0200 | <haskellbridge> | <thirdofmay18081814goya> so if the list is "[(0,0), (1,1), (2,2)]", we have two lines |
| 2025-03-30 05:21:18 +0200 | <EvanR> | ok you can't draw a circle |
| 2025-03-30 05:20:56 +0200 | <haskellbridge> | <thirdofmay18081814goya> well yeah: the ordering of the linked list specifies which lines exist |