GHC has slow compile times because otherwise the #Haskell programmers would be far too productive. Balance must be maintained.
06.08.2025 21:01 โ ๐ 19 ๐ 2 ๐ฌ 0 ๐ 0@sjoerdvisscher.w3future.com
#haskell #categorytheory #lumatone ๐ซ @boekencurator.bsky.social ๐ https://w3future.com ๐ผ https://tweag.io ๐ฃ๏ธ https://strijpskamerkoor.nl
GHC has slow compile times because otherwise the #Haskell programmers would be far too productive. Balance must be maintained.
06.08.2025 21:01 โ ๐ 19 ๐ 2 ๐ฌ 0 ๐ 0foldMap!
04.08.2025 22:02 โ ๐ 3 ๐ 0 ๐ฌ 0 ๐ 0Something like this would be nice (especially if it would perform as well): foldMap \(Sigma tag val) -> Pi \tagโ -> if (tag == tagโ) then [val] else []
04.08.2025 21:58 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0The Clowder Project: An online resource for category theory and mathematics Logo is an associator pentagon with space cats.
From the "Progress" page: Eventually, they too grew too large for their format, as had happened before. Organizing them in PDF form kept getting more and more difficult. For instance, during its last year, compilation was taking 8 hours, and LuaLaTeX had to be recompiled specifically to increase certain built-in limits. Indeed, the last compiled PDF of the project ran at 10566 pages.
10,566 pages of category theory notes (that took 8 hours to compile) now in process of converting to Stacks Project format: www.clowderproject.com/index.html ๐งฎ
25.07.2025 17:23 โ ๐ 11 ๐ 3 ๐ฌ 1 ๐ 0It's always traverse, as per @clementd.wtf
24.07.2025 14:01 โ ๐ 2 ๐ 1 ๐ฌ 0 ๐ 0Honestly, don't bother learning about par and pseq, parTraversable is the only thing you need. #haskell #parallelism
24.07.2025 12:22 โ ๐ 7 ๐ 0 ๐ฌ 1 ๐ 0Pure parallelism, my favorite form of parallelism!
23.07.2025 16:48 โ ๐ 4 ๐ 0 ๐ฌ 0 ๐ 11. Researching this weekโs column, I felt that I understood at last just why, on so many issues โ benefits, Gaza, environmental and planning laws, anti-protest laws, bank deregulation, fiscal rules etc - Labour is out-Torying the Tories. Why it has become so cruel, so destructive and so spineless. ๐งต
23.07.2025 07:52 โ ๐ 781 ๐ 352 ๐ฌ 41 ๐ 57I believe proarrow equipments are a great setting to study optics in! Read more in my mastodon post:
types.pl/@sjoerd_viss...
As a โsenior man in techโ, I can only wholeheartedly support this article. The observations and conclusions very much match what I have seen and concur with the experiences I heard from other women. We (= the men) need to do better!
19.07.2025 10:32 โ ๐ 2 ๐ 2 ๐ฌ 0 ๐ 0I believe proarrow equipments are a great setting to study optics in! Read more in my mastodon post:
types.pl/@sjoerd_viss...
Who is doing the most exciting work at the intersection of PL and LLMs right now?
17.07.2025 10:32 โ ๐ 7 ๐ 1 ๐ฌ 1 ๐ 0A simple formalization of alpha-equivalence. ~ Kalmer Apinis, Danel Ahman. arxiv.org/abs/2507.10181 #ITP #Rocq
16.07.2025 08:26 โ ๐ 2 ๐ 2 ๐ฌ 0 ๐ 0The videos for Zurihac 2025 are online. Our own Cheng Shao, main author of GHC's Wasm backend, taught everything you need to know to build your own Haskell Wasm app in five lessons. www.youtube.com/watch?v=9zv0...
What about you? What was your favourite presentation?
Profunctor Laws dimap id id = id: OK 27 tests completed dimap composition: OK (62.03s) 14348907 tests completed
Smallcheck fun: Let's just check the (->) profunctor instance on functions on a tiny datatype, say data Three = A | B | C.
Hmm, why is it taking so long?? Oh, there are 27 functions of type Three -> Three, and the composition test takes 5 of those, that's 27^5 :-)
Having been subjected to Javascript for over 20 years, I still only got 16/28. This really is insane!
11.07.2025 21:16 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 0I scored 9/28 on jsdate.wtf and all I got was this lousy text to share on social media. This is insane @rich-harris.dev ๐คฏ
11.07.2025 20:04 โ ๐ 5 ๐ 1 ๐ฌ 3 ๐ 0Dank! Het heeft zijn voors en tegens zie ik nu.
10.07.2025 20:00 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0De lopend 10-jaarsgemiddelde lijn is nog steeds behoorlijk grillig, zou een nog langer lopende gemiddelde lijn erbij interessant zijn?
10.07.2025 10:41 โ ๐ 0 ๐ 0 ๐ฌ 1 ๐ 0There's a (very recent) Xkcd for that!
nos.nl/artikel/2573...
xkcd.com/3106/
You should search for "mandelbrot effect", including quotes, on google right now (with AI turned on).
02.07.2025 11:52 โ ๐ 1 ๐ 0 ๐ฌ 1 ๐ 1does anyone know, like, a LOT about nix, and would be willing to help me clean up nix build stuff for a huge project at work? it's OSS, and i could pay u :3
26.06.2025 19:53 โ ๐ 7 ๐ 3 ๐ฌ 3 ๐ 0Really proud to live in the Netherlands seeing this list.
20.06.2025 10:15 โ ๐ 4 ๐ 1 ๐ฌ 0 ๐ 0Wij hebben hier ook net onze wandeling erop zitten om dezelfde reden!
20.06.2025 09:56 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0TIL Haskell's bitshift-right operator works on strings, too!
main = do
putStrLn ("22" >> "1")
putStrLn ("891" >> "2")
How is this different from `fromJust` being in base?
15.06.2025 11:19 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0I.e. only separately functorial in each argument, they only allow the funny tensor product between categories.
04.06.2025 11:51 โ ๐ 0 ๐ 0 ๐ฌ 0 ๐ 0I wrote some words and made some renders on what the New Design Language of iOS, iPadOS and macOS might be. Physicality: The new age of UI. www.lux.camera/physicality-...
03.06.2025 15:07 โ ๐ 116 ๐ 22 ๐ฌ 14 ๐ 10You can hate on LLMs all you want, but with help from ChatGPT I was able to implement a bit of quantum computing, something I wouldn't have been able to do on my own. github.com/sjoerdvissch...
31.05.2025 10:41 โ ๐ 3 ๐ 0 ๐ฌ 0 ๐ 0It warms up my hands.
29.05.2025 20:02 โ ๐ 1 ๐ 0 ๐ฌ 0 ๐ 0