> CRUDy web API things, because frankly that's most of all software that gets written
If that statement is true (I couldn't say), I might argue it's incomplete, because the elephant in the room is JS/TS. Haskell-in-the-browser is in horrible shape, which means that options for tight integration with UI code are severely limited. If there really is endless work for UI-less, JS-foreign web tech then that's great for Haskell ... maybe.
Because ... look what you've done to have an incrementally better experience writing CRUD apps:
1. You've got existential bus risk in your senior haskeller, and you live or die by their quality and experience, which since there's strictly less of them, is probably off the beaten path.
2. You've chosen a stack that has a massively smaller hiring pool. The benefit of "oh they'll take less pay because of the shiny" is maybe great at first, but if you have any adoption and attendant perf problems, you'll need those senior devs. Even if they are willing to learn, they can't use the tools they know best so at a minimum you're losing momentum.
3. This is more typical of smaller langs but Haskell has it's own spin on it: you're risking not having library support for a million things when you need it. Haskell has the false comfort of doing C FFI well, but that just means that when your core integration stops being maintained, now your haskellers better be C-savvy as well.
> dire and antagonistic view of things
If that's true, it's because the promise is so much greater than vanilla apps. I loooove runtime languages and Haskell should be the best of them: better-than-average GC, incredible concurrency support, great FFI, flawless refactoring, perfect DSLs, solid apps. It makes it worth forcing programmers to learn an unusual language if you can really conquer anything, and for a bright moment in the 2010s it seemed like you could.
Now, Haskell has lost that momentum and mindshare to Rust. This is literally true with some core Haskell library maintainers straight-up defecting. There are signs in the Rust community now of slowing momentum, and for similar reasons. There might be a boom-bust cycle for advanced languages with shaky industry support, I don't know.
Haskell leadership needs to change. There needs to be a laser focus on tooling, on browser compilation toolchain, on industry adoption. Avoiding success at all costs was a clever idea for a minute (ok for 3 decades) but post 2010s, there seems to be an active desire to shut off industrial adoption to keep it safe for the tinkerers. (End rant)
If that statement is true (I couldn't say), I might argue it's incomplete, because the elephant in the room is JS/TS. Haskell-in-the-browser is in horrible shape, which means that options for tight integration with UI code are severely limited. If there really is endless work for UI-less, JS-foreign web tech then that's great for Haskell ... maybe.
Because ... look what you've done to have an incrementally better experience writing CRUD apps:
1. You've got existential bus risk in your senior haskeller, and you live or die by their quality and experience, which since there's strictly less of them, is probably off the beaten path.
2. You've chosen a stack that has a massively smaller hiring pool. The benefit of "oh they'll take less pay because of the shiny" is maybe great at first, but if you have any adoption and attendant perf problems, you'll need those senior devs. Even if they are willing to learn, they can't use the tools they know best so at a minimum you're losing momentum.
3. This is more typical of smaller langs but Haskell has it's own spin on it: you're risking not having library support for a million things when you need it. Haskell has the false comfort of doing C FFI well, but that just means that when your core integration stops being maintained, now your haskellers better be C-savvy as well.
> dire and antagonistic view of things
If that's true, it's because the promise is so much greater than vanilla apps. I loooove runtime languages and Haskell should be the best of them: better-than-average GC, incredible concurrency support, great FFI, flawless refactoring, perfect DSLs, solid apps. It makes it worth forcing programmers to learn an unusual language if you can really conquer anything, and for a bright moment in the 2010s it seemed like you could.
Now, Haskell has lost that momentum and mindshare to Rust. This is literally true with some core Haskell library maintainers straight-up defecting. There are signs in the Rust community now of slowing momentum, and for similar reasons. There might be a boom-bust cycle for advanced languages with shaky industry support, I don't know.
Haskell leadership needs to change. There needs to be a laser focus on tooling, on browser compilation toolchain, on industry adoption. Avoiding success at all costs was a clever idea for a minute (ok for 3 decades) but post 2010s, there seems to be an active desire to shut off industrial adoption to keep it safe for the tinkerers. (End rant)