Don’t get me wrong I support Phoenix that is why I am here. But scaling is a little overrated, Facebook still uses PHP, Instagram still uses Django. And for those with less traffic they will be fine managing scalability. Phoenix has to mature more in terms of the size of community, available batteries, and learning resources. This forum however is one the best in helping Phoenix grow.
jslearner
dimitarvp
Instagram is also owned by Facebook – and Facebook can afford to literally shovel dollars in a furnace in order to avoid hiring programmers that work with other languages and not PHP.
Many people aspire for their companies to grow to the scale of a FAANG corporation (Facebook / Amazon / Apple / Netflix / Google) but the truth is that the problems they have are vastly different compared to the problems of the other 99.9% of companies in the world.
FAANG’s methods should not be emulated unless you grow as big as them. So count me strongly not impressed by their gross resource inefficiency management.
Spend enough time around here and you’ll see this is a divisive topic. Elixir’s creator is literally a former Rails contributor and he also helped a lot with Devise (Rails’ authentication framework). Many people working with Phoenix do not want it to become too bloated and coming by default with a lot of batteries included.
I get the sentiment; sometimes I wish Phoenix came with an authentication generator out of the box with a bunch of options – like do you require re-sign-in after you change the email of your account or not? – and sometimes I do get grumpy that I have to do a lot of stuff manually. But then I remember that this way I am in control; not only of the features and possible bugs introduced by the extra batteries, but also of the performance.
Phoenix’s scalability and lean resource usage are actually far down my list of its favourite traits. They are something that sweetens the deal to a degree but it’s not what I love about it. The fact that it steps on Erlang’s OTP and thus it is extremely predictable under load has to be in the top 3 though. Many other frameworks very quickly buckle under pressure, and non-FAANG companies cannot afford to just spend hundreds or thousands bucks extra in monthly hosting bills to dodge the effort needed to fix the scalability and lag problems.
4
jslearner
jslearner
dimitarvp
Definitely not, such absolutism leads directly to nowhere. 
There are a lot of projects out there that have a ton of developers in them, the company is bound by enterprise contracts to use certain hosting and/or technology, etc. etc. – the real world has a lot of political constraints that we the programmers find boring and don’t take into account.
But if you have the choice and the blessing of the higher-ups, definitely go with Phoenix. I was frustrated while doing several projects with Phoenix but in the end I found out that the degree of control I have gives me everything I need to solve every problem in the project. And that’s not the case with other frameworks.
And, as mentioned earlier, the smaller hosting bills are a very nice bonus.
Finally, Elixir’s resilience and predictable multicore performance are something that a lot of other languages and frameworks are just now starting to try to emulate. Erlang/Elixir are miles ahead – in the purely technical sense – of pretty much 99% of everybody else.
4
keathley
jslearner
I agree Phoenix has its strength. It’s just hard to justify a Phoenix project to clients when even Wordpress will do. For my own projects I’m using Phoenix though.
0
dimitarvp
Eh, but WordPress solves entirely different problems.
WordPress => shovel websites, get paid by the kilogram. 
Phoenix => handcraft serious web applications.
They are not comparable at all. It’s like comparing a static website generator – like Hugo or Gatsby – with Phoenix.
4
dimitarvp
jslearner
@dimitarvp I’m not saying your Elitist, I’m just saying you should not belittle other frameworks because it turns away those wanting to join the Phoenix community. Btw it’s pretty unfair to say Wordpress developers does not create serious web applications.
0