Some recommendation. Reference here. I think mostly it has to be between Haskell or OCaml if we go straight into FP, but Elixir for practicality. I think this is great because it gives you rooms of reproducibility and automation, a more low-level control over state machine, store and transform.
With low-level control also comes with better performance, which is great because you don’t have to come back and optimize it often anymore, and that’s why it is great and important to pick the most lazy ones with not much abstractions or frameworks (good ecosystem and tooling is a must, but not meta frameworks).
The days of scaffolding and boilerplate or the kings of frameworks are behind us. Moving forward, only thing that matters is system and design.
Making money as a developer is seemingly antagonistic towards technicality (?), which is all about getting that sweet PMF and providing values for the customer. I am not against that obviously, but I believe that has to be able to be reconciled to some extent. Great craftsmanship and quality of engineering can and will absolutely become the guardrail of services, (see Benchling).
The main difference is really the application code vs the “code” code, where everything else is functions. It is a big stretch thinking you can make everything functional and nice and clean and expected when you are shipping features that your customer wants, and you do not get to control what they want.
A bit of meta-programming would be great too. I love how Clojure or lisp handles programs, treating everything as data transformations and return output based on the input calls.
I think the future will just be a giant function that takes an input (initiation/idealization/birth) and gives an output (observation/outcome/death). This is critical in d/acc because if we can do this in a sustainable, democratic and decentralized way and having sounded institution as guardrails to prevent tampering (e.t. Daron Acemoglu, James A. Robinson, Simon Johnson), we can achieve a good ecosystem & products that lasts long.
What problems do you want to solve exactly? You clearly care about DevX, FP and quality of engineering A LOT. Besides consumer/B2B SaaS, is there anything that you can extract value or provide solutions to some problems? Even games can solve some creativity starvation that I have.
Bluesky, ATProto and FP, is there anyways to aggregate everything? Is the internet the way you wanted it to be? It is not democratic and anything can be taken down in a heartbeat. We need a way to duplicate the internet so hard right now to avoid tampering, and I cannot think of anything else except a distributed database/internet that has asymmetric signature are able to doing so.
Also Look at Phil Bagwell’s work and Eric Steven Raymond (ESR)‘s work on catb (how to become a hacker).
Scala OCaml Haskell elixir Elm PureScript Idris 2 Clojure Lean 4 Factor
Notes for mostly-adequate-guide-to-fp
Resource
- How to Design Programs: Preface
- fp-ts: learning resource
- Mostly adequate guide to FP by @DrBoolean
- The State monad by Paul Gray
- Functional Programming with TypeScript by Sahand Javid
- fp-ts Tutorial series on YouTube by @MrFunctor
- Effect: Why Effect? | Effect Documentation