Some random thoughts after discussions at the FP Dag yesterday:

We need more learning resources based on pracitcal projects like https://learn-haskell.blog/, and the practical project should not be compiler related. Perhaps the Haskell School of Expression is also a good example of this but it is a bit outdated.

One objection to teaching Haskell through more practical Haskell projects is that it wouldn't show off the benefits if all you're working with I/O almost all of the time. I asked Jaap van der Woude who was sitting next to me - and is a fervent proponent of program design by calculation (e.g. as described in this book) - the question: "Could you calculate a game from its specification?" His immediate response was no, but later he did explain that he used to give students an assignment to specify the game Snake, which was already often too much for his students.

I'll have to try doing that exercise myself sometime. I still think that if we ever want to get to a future where we can write correct and reusable software, then we'll have to find a way to make it easier to formally specify our programs.

#haskell