As the end of the year approaches, Twitter has a hash-tag named #Code2014 that encourages people to tweet which programming languages they’ve been using over the last 12 months.
I joined in, but afterwards felt compelled to explain my thoughts on each in more than 140 characters. Hence this blog post.
My main language, and the one I am gainly employed to write in. I found myself at the Scala Exchange conference again this year, and was pleasantly surprised by the attention that Scalaz was receiving. A lot of people were still unconvinced, but the Scalaz talks were very well attended.
I have also found myself immersed in Twitter’s Scala-Ecosystem in the last quarter of the year. I’m thoroughly enjoying it. I have become more convinced that Akka is middleware, and Futures (rather than Actors) are more than sufficient for the vast majority of use-cases.
First-class functions are so incredibly versatile, that I now get frustrated when using Java.
On a related note, this tweet amused me this year:
OH: "But functional programming is too complicated." pic.twitter.com/ksPeGgvMgE— Bryan Maass (@escherize) December 21, 2014
For a greenfield project, I’m not sure why I would ever choose Java over Scala again.
It has been said that Scalaz is a gateway drug from Scala to Haskell, but it is the blurred lines between OOP and FP in Scala that occasionally irritate me. As a pure language, Haskell will hopefully not have the same kind of problems. I have now finally finished reading LYAHFGG and I will start writing more things in Haskell next year.
I was incredibly excited by Clojure at the beginning of the year, but my excitement is beginning to wane. I would love to love Clojure; the collections libraries are magnificent, it is incredibly concise, and the premise of s-expressions is compelling. However, the lack of static typing constantly annoys me. Typed Clojure and/or Prismatic Schema may alleviate my concerns though. I need to spend more time investigating them.
I have been working my way through The Little Schemer, and apparently they had renamed Scheme to Racket. I’m not convinced it is as nice as Clojure, but might be worth investigating Typed Racket if the libraries to provide typing in Clojure don’t turn out very well.
The “Good Parts” are constantly shrinking though, as Douglas Crockford revised his advice:
Python is my default scripting language. I enjoyed PayPal debunking some myths of Enterprise Python, but Python is slow and lacks good concurrency support in comparison to Scala.
I have written a couple of utilities and one website in Golang this year. It could potentially replace Python as my default “operations” language once I learn more of the standard library.
It is annoying that it doesn’t have a
map function though, and without immutable data structures, I probably won’t ever trust it to build anything complicated.
Lua is a language I will spend more quality time with in 2015. Having previously worked with Lua when embedding it in Redis, I was excited to discover OpenResty and being able to embed it into NGINX. I also want to investigate using Lua for CUDA GPU Programming.
More In 2015?
Aside from those already listed, I need to spend some time with Erlang next year. I have now read a couple of books about it (including the excellent Learn You Some Erlang), and I really need to write some to experience it firsthand.