HNHacker News
TopNewBestAskShowJobs

michael-lehn

2 karma · joined July 26, 2026

Mathematician and lecturer at Ulm University, Germany. Numerical mathematics, scientific computing, HPC, compilers, and computer architecture. I like teaching computing from first principles — from logic gates and CPUs to compilers and numerical software. GitHub: https://github.com/michael-lehn Web: https://www.uni-ulm.de/mawi/institut-fuer-numerische-mathematik/institut/mitarbeiter/mlehn/
submissionscomments
michael-lehn··on GEMM Tutorial and UlmBLAS
Ups, screwed up the title. Should have been "Show HN: ..."
michael-lehn··on Show HN: ABC v0.1 – A teaching language grows up a little
A small update on the naming rabbit hole. A few candidates I currently like: emsiel — pronounced roughly like “M-C-L”. I like how it sounds; several people independently said it sounds vaguely biblical. :-) The downside is that it feels more like the name of a personal project than something inviting people to build a community around.

uhu — “under the hood”. This fits the teaching philosophy quite well: the language/compiler is deliberately small and, more importantly, doesn't try to hide what's going on under the hood.

ijc (“itchsy”) — short for “it’s just C”. It stays conceptually close to C, while using the syntax I wish C had. There's also an accidental “I just see” hiding in the spelling.

At the moment I'm leaning towards ijc, but I'm still not quite ready to rename everything. :-)

michael-lehn··on Show HN: ABC v0.1 – A teaching language grows up a little
OK, you convinced me. :-)

And the timing is probably good. I'm just about to leave my basement and head to LA for a few days to enjoy some actual sunshine. So I'll either turn to ashes or come back with a creative new name. ;-)

michael-lehn··on Show HN: ABC v0.1 – A teaching language grows up a little
Yes, I know about the earlier ABC. :-) When I originally named mine, it was really just a small language for my own course, so I wasn't particularly worried about name collisions. And the wordplay was simply too tempting: students would use ABC ("A Better C") to write their own compiler, which I jokingly called ABC ("A Bloody Compiler"). The student compiler has since grown into a separate project and is now called "not-abc". Until recently I still thought the language itself was small and obscure enough that there wasn't much risk of confusion. But now that I'm starting to develop it beyond its original teaching purpose, I think the naming issue is a fair point. You're also not the first person to bring it up since I published v0.1. So yes, I may have to come up with a new name. :-) I'm not convinced by LehnABC yet, but suggestions are certainly welcome.
michael-lehn··on Anonymous GitHub
Ok, but why? I can host my own git server for using git just by myself. Why should I want to share it and loose any credit?
michael-lehn··on Ask HN: What's the endgame of the AI comments buried in every post?
To get some upvote, to increase karma
michael-lehn··on Ask HN: Is GitHub Fried Today?
Today it's mostly working. Checkout the status here: https://www.githubstatus.com
michael-lehn··on Best code editor without AI?
Yes, neovim is great.
michael-lehn··on Ask HN: Where to learn software engineering fundamentals?
One thing I would add to the excellent suggestions above is that it can be surprisingly useful to go one level below what is usually considered “software engineering fundamentals.” I teach mathematics students, and over the years I've become increasingly convinced that understanding some of the machine underneath makes many higher-level concepts much easier to understand later: memory and addresses, a simple CPU, assembly, calling conventions, and eventually how a compiler translates a programming language into something the machine can execute. There is a pedagogical principle I like: don't learn the solution before you know the problem. For example, pointers become much less mysterious once memory and addresses are concrete. Many programming-language features also make more sense once you have encountered the problems they are designed to solve. This is one reason I like the Computation Structures recommendation above. I've been experimenting with a similar bottom-up approach in an undergraduate course at Ulm University. We eventually go as far as writing a small compiler, and then rewriting a smaller version of it in its own language so that it becomes self-hosting. I've started putting the material in English here:

https://github.com/michael-lehn/not-abc