3 ms·

Elixir – Erlang didn't change, the world did (2017)

8 points by mindB 9y ago

mindB 9y ago
I'm not a user of Elixir or of Erlang. I stumbled upon this article, and I'm now thinking about giving it a try. The "guns" analogy was an interesting take, and I wanted to hear what HN thought.
- The_suffocated 9y ago
  The author does have a point. Data immutability is nice and it does make your program safer. But that "guns" metaphor is cliché. One may also say that Erlang is more dangerous because it is not statically typed. I don't think crediting/discrediting a language based on simply one language feature is justified. (Actually, Erlang has mutable data too if you use mnesia, ets or the like, but you may just opt not to use them.) In the article's chart, the author also classifies Go as one of the most dangerous languages. Now I'm a novice Elixir programmer and I don't know Go, but if Go were really so dangerous, I should have heard many horror stories about it (you see, Go is a vastly popular language, much more so than Erlang and Elixir combined). Yet, while you can find a significant number of articles criticising how primitive Go as a language is, those "shoot yourself in the foot" type of complaints are surprisingly few (especially if you compare the situation with C or C++). I think this tells us something about Go's safeness. And he says that "To an Erlang programmer Haskell is an impractical language". I wonder how many Erlang programmers out there would be so snobbish. I have only dabbled with Haskell, but my impression is that Haskell is an unbelievably clean language (at least when compared to OCaml or Elixir), to a point that if I write a Haskell program, I would be very confident that what I've written is exactly what I intended in mind. How is that impractical? If you really want to see some in-depth discussions about Erlang, I recommend the following two interviews: 1. Joe Armstrong and Simon Peyton Jones discuss Erlang and Haskell (https://www.infoq.com/interviews/armstrong-peyton-jones-erlang-haskell https://www.infoq.com/interviews/armstrong-peyton-jones-erla...) 2. Ralph Johnson, Joe Armstrong on the State of OOP (https://www.infoq.com/interviews/johnson-armstrong-oop https://www.infoq.com/interviews/johnson-armstrong-oop)

8 points by mindB 9y ago

I'm not a user of Elixir or of Erlang. I stumbled upon this article, and I'm now thinking about giving it a try. The "guns" analogy was an interesting take, and I wanted to hear what HN thought.

The_suffocated 9y ago
The author does have a point. Data immutability is nice and it does make your program safer. But that "guns" metaphor is cliché. One may also say that Erlang is more dangerous because it is not statically typed. I don't think crediting/discrediting a language based on simply one language feature is justified. (Actually, Erlang has mutable data too if you use mnesia, ets or the like, but you may just opt not to use them.) In the article's chart, the author also classifies Go as one of the most dangerous languages. Now I'm a novice Elixir programmer and I don't know Go, but if Go were really so dangerous, I should have heard many horror stories about it (you see, Go is a vastly popular language, much more so than Erlang and Elixir combined). Yet, while you can find a significant number of articles criticising how primitive Go as a language is, those "shoot yourself in the foot" type of complaints are surprisingly few (especially if you compare the situation with C or C++). I think this tells us something about Go's safeness. And he says that "To an Erlang programmer Haskell is an impractical language". I wonder how many Erlang programmers out there would be so snobbish. I have only dabbled with Haskell, but my impression is that Haskell is an unbelievably clean language (at least when compared to OCaml or Elixir), to a point that if I write a Haskell program, I would be very confident that what I've written is exactly what I intended in mind. How is that impractical? If you really want to see some in-depth discussions about Erlang, I recommend the following two interviews: 1. Joe Armstrong and Simon Peyton Jones discuss Erlang and Haskell (https://www.infoq.com/interviews/armstrong-peyton-jones-erlang-haskell https://www.infoq.com/interviews/armstrong-peyton-jones-erla...) 2. Ralph Johnson, Joe Armstrong on the State of OOP (https://www.infoq.com/interviews/johnson-armstrong-oop https://www.infoq.com/interviews/johnson-armstrong-oop)