peerreynders
malloryerik
Here’s a Hacker News thread from a guy who’s done just what you’re talking about (he says he doesn’t know OTP/concurrency stuff yet at all and just dove right in to Phoenix).
The site in question: ficdb.com. Nicely executed, right? He’s kindly open sourced it. So what you’re talking about can be done. I imagine that if his site were focused on live interactions between people for example that he’d have taken a step back and studied Elixir and OTP/concurrency more. And I might be wrong but I’d guess that he’s starting to learn Elixir, OTP, and the rest of it more deeply now.
3
AstonJ
it was made by 
Cruz
I’m on the same boat as you, started learning Phoenix just a few months ago whenever I could find time. So, I’m still learning Elixir, Phoenix, Ecto, and the ecosystem in general.
I also thought that I could learn Phoenix without much of Elixir. However, after spending a little time with the framework, I came to the conclusion that even if I could, I didn’t want to do that anymore. There are many reasons for that, but these three come to mind at this moment:
-
Elixir is awesome. I haven’t felt this way about a language since the early days of Java - don’t judge, if you’ve spend your days programming DB based applications in C++, you’d have loved Java as well
I found it impossible not wanting to learn more of the language -
The BEAM and especially OTP are remarkable pieces of engineering. I just regret not finding about it ten years ago. Not only you can build something that performs and escalate, but you do that knowing that the end result will be reliable, fault tolerant even. Elixir makes it a lot easier to use.
-
If you really want to build just “web sites” there are plenty of languages and tools for that. But, if eventually you want to move to something more complex, you probably want to follow the “Phoenix is not your application” method (or mantra if you wish). By doing that you’ll find yourself thinking about your domain problem right at the beginning of the development cycle. And, by the time you bring Phoenix into the picture, you probably feel pretty confident about the system you have built.
This is why I’ve put my Phoenix book in the shelf, and I’m up to my third book in Elixir
Good luck
4
axelson
Scenic Core Team
If you have trouble coming up with a problem that is interesting and feasible for you then you might want to take a look at exercism.io they have a good Elixir track and will help you right idiomatic Elixir code for fun little problems. I like the problems better than something like project Euler because they aren’t overly mathematics focused.
2
bjunc
I’m a bit of a “jump into the deep end” kind of guy, but I started by reading Programming Phoenix (I also have quite a bit of front-end experience). While the book focuses more on web apps, you absolutely get introduced to core Elixir concepts along the way, but you get to build something tangible too. I even subbed out the controller logic of the examples for a GraphQL schema (check out the Absinthe library). The Elixir documentation is great, so you’re really only a Google search away from digging into pattern matching, what a tuple is, guards, etc..
In my opinion, way too much “check this language out” discussion about Elixir is low level performance, and zeroes in on the nuts and bolts of concurrency, self-healing, etc.. While that stuff is GREAT, I think for most people new to the language, it’s better to start the productivity aspects.
Lastly, here’s an article I wrote touching on this a bit more:
https://edgewise.engineering/choosing-elixir-5c1f4052dd28
and a podcast where I go into it a bit more:
https://devchat.tv/elixir-mix/emx-014-choosing-elixir-with-bobby-juncosa/
1