I made the decision to stop using Haskell for things I consider "professional" use a couple years ago. No matter how much I love the idea of the language, there

I made the decision to stop using Haskell for things I consider "professional" use a couple years ago. No matter how much I love the idea of the language, there's a couple superficial things it gets wrong and one fundamental thing, or at least I believe it's fundamental, perhaps some academics have proof one way or the other. The superficial things are things that are not really part of the language, but more of its standard library, part of the 'prelude', or things the community has agreed upon: - the I/O monad (and others) can throw errors - prelude is full of bad practice like use of strings and linked lists - there's something wrong with how function names can't be namespaced There's more than this, and they're not just things I came up with, they've been fought over in the community at length. Other preludes exist, I/O libraries that don't throw errors, even whole different systems for doing I/O exist. It's not that solutions are hard or impossible, they're there, it's just that the way the language is managed it's not likely that they'd ever be applied in a way the entire community, and especially those drawn to it by any marketing would experience them. Maybe I'm making a fool of myself and things changed in the past couple of years, things were really developing at a rapid pace when I left, but it seemed all the improvements happened around haskell, not to the language itself. When Stack and LTS-Haskell emerged it was the most amazing thing that I saw happening to Haskell in like 10 years, but given the lackluster response Snoyman received from the core team it seemed like they didn't even realize it even though the entire community had like an upheaval. In my opinion, Haskell's marketing is fine. It's being marketed to every college student anywhere that goes to a good enough university that it teaches functional programming. In addition to that there's raving fans (myself included) that extoll its virtues to anyone who would hear. What it needs is a better retention rate. It needs that people who learn it, immediately realise it's fit for production use, that it's not just a fun experiment and a educative way of looking at things. For that we need something like a Haskell 2.0, designed by the community to incorporate everything we learned about Haskell's flaws and make it an absolute powerhouse of productivity. Maybe Haskell should be split into two flavors, one the teaching language with the linked lists, strings and funny pattern matching, and one that's got a prelude that will set you up with the perfect base to build the next compiler, database, service or web application. If all those smart people who learn Haskell and tinker with for a bit, actually stayed around and used it to build awesome stuff, Haskell would truly have an enormous foot print on the world. (not really relevant but the fundamental thing is that Haskell's performance is really hard to predict, this is true for many garbage collected languages, but for Haskell it seems to be extra bad because of the way its paradigm just doesn't map 1:1 to how computers are built)