Was it wise to start this project in Elixir? Note - I have no idea what this project does, it may be a tremendous success, I just think it's a pretty wide gamble to start it in a language that has such a tiny user base and steep learning curve.
Elixir is, more or less, a thin shell on top of Erlang. As long as Erlang is maintained, Elixir can be pretty easily carried along with, as far as maintaince goes.
There’s an entire web framework for Elixir called Phoenix.
You’re painting Elixir as this obscure language that no one knows and while it’s undeniable that it may not be as widespread as JS, PHP or Python, it is not so marginal.
I live in a European city that’s not a tech capital and a couple major startups use it, and I’ve seen it used in major apps across the web.
Never used it myself but read how it excels at concurrency and messaging so why not use it for this?
It is pretty much obscure, I don't think a couple of major startups is gonna cut it. Elixir never got the kind of adoption that will last for years and years after the hype is gone. Instead it had a nice spike a couple of years ago and i's declining already. Ruby can live for another 20 years based on the 10 good years it had, Elixir never had those years it seems. Yes, it got tremendous attention on hackernews and blogs, but actual adoption? Seems like startups are more conservative than one might guess.
Even worse, it's functional. So if a PHP/Ruby/Python dev wanna work on this cool OSS project, when it's in Elixir that's just another big hurdle.
See for yourself https://insights.stackoverflow.com/trends?tags=elixir
Ruby and PHP are also declining (well, relative to Stackoverflow questions), but Ruby still gets 10x more questions than Elixir and PHP 100X(!) more. Now sure, lots of PHP questions are from noobs who don't work in the industry. But if we start looking at jobs we're gonna see pretty similar results.
Some examples of companies using Elixir / Erlang in production: Discord, WhatsApp, WeChat, Bet365, Pinterest (notification system).
This is hardly "obscure" - these are applications used by billions of users all over the world.
Potential developer market for your business or OS project is a factor in engineering, sure. But it's certainly not the only one, and there's a utility threshold for how useful a large developer base can be - maybe I don't care that there are only 10,000 good Elixir in my region if I only need 3 or 4, and I can entice them with good conditions (salary, or a prestigious OS project)
World-class CTOs and engineers choose Erlang and Elixir for their working characteristics as programming languages, a point which you've chosen to completely ignore.
> Even worse, it's functional
I wouldn't consider myself a functional programmer (although if a language offers FP facilities I often use the hell out of them over imperative and OO constructs) but if I built something in FP because I thought it was be the best-suited paradigm for the task at hand, I'd happily weed out people who can't be arsed to learn the rudiments.
We'll have to agree to disagree. I think the numbers clearly show Elixir is obscure and declining fast and that in general that can't be good for OSS projects, you seem to think otherwise and that's OK.
