Lightnews — Scholar-powered news

Reposted by Chris Penner

Simon @hojberg.xyz · 2d

I wrote about AI and programming.

About the identity challenges we are facing as programmers and the potential erasure of craft.

There's a fun little easter egg in the header as well :)

hojberg.xyz/the-programm...

The Programmer Identity Crisis ❈ Simon Højberg ❈ Principal Frontend Engineer

On AI, Creativity, and Craft

hojberg.xyz

4 10 18

Chris Penner @chrispenner.ca · 2d

P.s. try clicking the HAL9000 in the header;

3

Chris Penner @chrispenner.ca · 2d

I'm very pleased to call Simon ( @hojberg.xyz ) my co-worker, team-mate and friend, go check out this wonderfully written take.

hojberg.xyz/the-programm...

The Programmer Identity Crisis ❈ Simon Højberg ❈ Principal Frontend Engineer

On AI, Creativity, and Craft

hojberg.xyz

2 4

Chris Penner @chrispenner.ca · 2d

IMO we shouldn't be trying so hard to solve "the expression problem"; We should make it easy to update existing functions when types change, but it shouldn't always "silently work" with the new type; oftentimes existing functionality needs a human to look over it when there are changes!

Chris Penner @chrispenner.ca · 3d

Error-as-value;

Yes, all tasks should be cancelled, `ki` takes care of that in my case.

hackage.haskell.org/package/ki-1...

ki

A lightweight structured concurrency library

hackage.haskell.org

1

Chris Penner @chrispenner.ca · 3d

hackage.haskell.org/package/ki-1...

Ki

hackage.haskell.org

1

Chris Penner @chrispenner.ca · 4d

What do folks out there do for error handling in concurrent tasks in #Haskell?

For now I've settled on having an error `TMVar`, then use a `ki` structured concurrency scope which forks all my jobs and ends in an `(Right <$> awaitAll) <|> (Left <$> readTMVar errVar)`

2 3 8

Chris Penner @chrispenner.ca · 4d

TIL callCC's return value isn't rank 2 polymorphic (but you can just route it through Void and use `absurd` to make it rank 2 polymorphic)

Spent too long staring at the error wondering why it wasn't working haha.

#Haskell

github.com/haskell/mtl/...

Why not use rank 2 polymorphism for `callCC` · Issue #127 · haskell/mtl

I think rank 2 polymorphism is a must-have for callCC, or this code will not compile: f :: Int -> String f x = (`runCont` id) $ callCC $ \exit -> do y <- case x of 0 -> exit "error: x == 0" _ -> pu...

github.com

5

Chris Penner @chrispenner.ca · 7d

Huh, TIL that `Text.lines != Text.splitOn "\n"` when the string ends in a newline.

Reposted by Chris Penner

Unison @unison-lang.org · 9d

You can now spin up Unison Cloud clusters on your own infrastructure!

✅ Build elastic distributed systems and services in vastly less code
✅ Fast, typed RPC
✅ Deployments in seconds
✅ Free to get started

youtu.be/0sZqI1XoGLY

Unison Cloud on your infrastructure

YouTube video by Unison Language

youtu.be

8 10

Chris Penner @chrispenner.ca · 13d

Stay tuned and you won't be disappointed, Arrows and Category classes are next!

5

Chris Penner @chrispenner.ca · 13d

Oh whoops, lol that’s an important “not”!

3

Reposted by Chris Penner

Andrew Jarrett @ahrjarrett.bsky.social · 13d

Interesting article by @chrispenner.ca about why monads aren't the solution to everything, and a functor I've never crossed paths with called selective applicatives.

Still absorbing it, but great read as always. Check it out if that's your thing.

chrispenner.ca/posts/expres...

Monads are too powerful: The Expressiveness Spectrum

Monads are a useful tool, but what costs do we pay for their expressive power?

chrispenner.ca

1 3

Chris Penner @chrispenner.ca · 13d

Haha, I told my friend you couldn’t write an echo server, he said sure you can; you just enumerate every string one-by-one… good way to get your LOC up…

1 1

Chris Penner @chrispenner.ca · 13d

New blog post!

#Haskell is built on a foundation of Monads, but are they really the optimal way to sequence effects or should we keep looking for something better?

What's the big deal with Applicatives and Selective Applicatives?

Read on!

chrispenner.ca/posts/expres...

Monads are too powerful: The Expressiveness Spectrum

Monads are a useful tool, but what costs do we pay for their expressive power?

chrispenner.ca

4 8 30

Reposted by Chris Penner

Unison @unison-lang.org · 14d

✨ The new dependents view shows everywhere a definition is used, broken down by category, as a separate card in Unison Share and the Unison Desktop app.

Your codebase is a bidirectional graph! You should be able to walk it as such.

3 8

Reposted by Chris Penner

Unison @unison-lang.org · 17d

📽️ New to the Unison workflow? We’re making short videos to walk you through the basics.

Follow along here: www.youtube.com/watch?v=QJ1h...

Install and upgrade a library dependency

A walk-through of the process for installing a library dependency and upgrading it. * `lib.install` downloads the library into the `lib` namespace. * `ls lib` will show the current dependencies in the project * `upgrade` with no arguments opens up a version picker. Select the old version first, then the new one. Simple changes can be propagated without developer intervention; more complicated upgrades will prompt the UCM to create a branch for the upgrade process and open up the impacted terms in a scratch file. The resolution process follows the standard workflow for updating Unison code, described here: https://youtu.be/8lCDSCm_rGA.

www.youtube.com

2 5

Chris Penner @chrispenner.ca · 21d

I'm pretty stoked on how well WASM seems to be working for languages that target it (not that there aren't difficulties there).

Even more stoked for these updates!

Thomas Steiner @tomayac.com · 22d

In #WebAssembly spec news, #Wasm 3.0 is complete 🎉! Read all about it in Andreas Rossberg's article: webassembly.org/news/2025-09....

Wasm 3.0 Completed - WebAssembly

WebAssembly (abbreviated Wasm) is a binary instruction format for a stack-based virtual machine. Wasm is designed as a portable compilation target for programming languages, enabling deployment on the...

webassembly.org

3

Chris Penner @chrispenner.ca · 24d

It may in fact be a VeryBadIdea™ but I'm extremely tempted to duct tape my soldering iron to my 3D printer and use it to wood-burn the grid lines on the Go board I'm building...

1

Chris Penner @chrispenner.ca · Sep 8

We all know AI is unreliable, but what's the worst case you've seen?

#zig
This one made me exhale sharply through my nose today:

Zig doesn't have a built-in "defer on error" construct, but there are several patterns you can use to achieve similar functionality:
Method 1: Using errdefer (Zig's error-specific defer)
Actually, Zig does have exactly what you're looking for! It's called errdefer:

2 1 8

Chris Penner @chrispenner.ca · Sep 4

I'd be curious about a comparison which charts number of repos vs age of the language :)

1 4

Chris Penner @chrispenner.ca · Aug 22

Programming languages really need a way to expose functions as "you can import this within this project and tests, but it's not public to other libraries"; somewhere in-between fully public and fully private.

3 4

Chris Penner @chrispenner.ca · Aug 21

Here's the source of the quote btw: sdiehl.github.io/typechecker-...

Language Design - Typechecker Zoo

Typecheckers built for fun

sdiehl.github.io

1

Chris Penner @chrispenner.ca · Aug 21

I personally love how low friction it is to just try it, I’ll run jj absorb, it says exactly what it did, you can just undo if you didn’t like it, no harm done!

3

Reposted by Chris Penner

Haskell programming language @haskell.org · Aug 20

The GHC developers are very pleased to announce the availability of the first alpha prerelease of GHC 9.14.1.

Highlights:

* Major improvements to the Specialiser
* SSE/AVX support in the x86 native backend
* Significant improvements in the GHCi debugger

blog.haskell.org/ghc-9-14-1-a...

#Haskell

GHC 9.14.1-alpha1 is now available | The Haskell Programming Language's blog

blog.haskell.org

14 41