4 ms·

I've been writing Haskell professionally for 5 years now and as a hobbyist even longer and this presentation just doesn't resonate with my experience and view t

I've been writing Haskell professionally for 5 years now and as a hobbyist even longer and this presentation just doesn't resonate with my experience and view towards the future. That said I've taken my focus off working in it professionally. It's such a chore to have to constantly justify it to newly-externally-hired management and inevitably get a top-down directive to rewrite. I'm done relying on managers and corporations to drive Haskell. My focus now is freedom & independence. I've got plenty of hugs and impactful personal projects & Haskell advancements I'd like to undertake in the next 5-15 years, and nothing even sniffs Haskell as a tool for me building without thinking or exerting energy. Maybe my issue is I love Haskell but I am abjectly not a community member.

linuxlizard 6y ago
What kind of production systems are a good fit with Haskell? I've wanted to learn Haskell but keep getting distracted by other shiny things that are more directly applicable to my job.
- JoelMcCracken 6y ago
  Services. Websites. Here is a reasonable breakdown for more info https://github.com/Gabriel439/post-rfc/blob/master/sotu.md https://github.com/Gabriel439/post-rfc/blob/master/sotu.md
- whateveracct 6y ago
  General purpose "cloud" software is what I've worked on. Compilers and interpreters. They don't have to be CLIs though. You can create languages to configure said cloud software :)
- marcosdumay 6y ago
  Haskell won't give you high interactivity (on the milliseconds of response time), or C-like performance. It will make it hard to control low level details of your memory and execution path. It will also not bring a lot of gain if you only have homogeneous data (like in machine learning). If you have a problem where intermediate performance (on the level of Java and .Net, better than Python or JS) is ok and people won't notice 100ms extra here and there, Haskell is viable. If it's not focused on low level IO, and deals with complicated stuff, odds are it's your best option right now.

linuxlizard 6y ago

What kind of production systems are a good fit with Haskell? I've wanted to learn Haskell but keep getting distracted by other shiny things that are more directly applicable to my job.

JoelMcCracken 6y ago
Services. Websites. Here is a reasonable breakdown for more info https://github.com/Gabriel439/post-rfc/blob/master/sotu.md https://github.com/Gabriel439/post-rfc/blob/master/sotu.md
whateveracct 6y ago
General purpose "cloud" software is what I've worked on. Compilers and interpreters. They don't have to be CLIs though. You can create languages to configure said cloud software :)
marcosdumay 6y ago
Haskell won't give you high interactivity (on the milliseconds of response time), or C-like performance. It will make it hard to control low level details of your memory and execution path. It will also not bring a lot of gain if you only have homogeneous data (like in machine learning). If you have a problem where intermediate performance (on the level of Java and .Net, better than Python or JS) is ok and people won't notice 100ms extra here and there, Haskell is viable. If it's not focused on low level IO, and deals with complicated stuff, odds are it's your best option right now.