| There are a few factors. Most devs are trained in OO approaches so you have to learn to think about problems from a functional perspective. While a lot of people believe that the functional approach is a huge win (myself included), asking developers to both use a different language AND think about problems differently is a leap. Truth be told, I was blissfully ignorant of how many concurrency issues non-BEAM languages truly have until I learned Elixir. Now I can't unsee them. While there are a lot of polyglot devs out there who are happy to utilize a different language for its strengths, there are also a lot of developers and companies who are both time and emotionally invested in their language of choice. It's not exclusive to a particular language, it happens everywhere. One of the unfortunate things that accompany languages that are able to do more with less, is that the companies which succeed with them...need fewer developers. There are a lot of hugely successful companies using Ruby on Rails, just as an example. Even today, you will see people fight tooth and nail against using it either because: A) They want to use their stack of choice or B) They don't understand why it's a better choice for the job. In language terms, you win developer eyeballs with 1 easily quantifiable thing...benchmarks. X is faster than Y in benchmarks, a lot of developers will want to use X. Quantifying productivity is harder (Ruby). Quantifying advantages of concurrency, clustering, heap isolation, maintainability with very high productivity is harder (BEAM). Benchmarks are easy. Widely adopted language selling points as a comparison for what Elixir is up against: - Java is taught in schools, heavily adopted in enterprise, portable, fast. It's okay at everything. - Python is taught in schools, installed by default on Linux. It's okay at everything. - Javascript is the only language available in browsers and has a selling point of using the same language on the client and server. Widely adopted despite a huge chunk of people who hate it. - .NET gets the entire backing of Microsoft, usage on Windows and a natural selling point with the various Windows shops out there. It's widely useful on those platforms and has an invested and certified following. - Apple languages are similar to .NET selling points. - Go is portable, pretty fast has a low learning curve, fast compiler (productivity win) and the closest concurrency model to the BEAM as well as backing of Google which lets execs rest easy. I can definitely make a case for why Elixir is a better choice for most of the things you'd use these language for, except mobile apps. But even then, there is going to be an argument. People will defend their choices and you won't see much progress. And somebody will ask for a benchmark, which Elixir won't win because it's not designed to win straight line benchmarks. |