HN Explorer

Ask HN: Should I learn only one language until I know it all?

1 points by devcheese ↗ HN
So I pretty much want to learn a lot of things. I want to learn Ruby, Go, Swift, Elixir, (few more things) I haven't really mastered one language, and I want to be able to do that. I've got some advice saying that it takes time and to keep learning.

Should I put my focus into mastering one language or have fun with the languages I want to learn and find which one I want to master?

8 comments

[ 4.1 ms ] story [ 24.8 ms ] thread
hawkice
Pick one where you have a friend who knows it well, and can help you through the pitfalls. Then build something complicated enough that you aren't entirely sure how to do it yet.

"Mastery", "until I know it all", "find which ones I [like]" all kinda fit into this process, so you end up doing them all at once.

[dead] viraptor
Depends what interests you, where you want to work, for how long, etc.

You can probably survive with Java or C# in corporate world for the rest of your career. In webdev you can probably stick to one language, but will have jump between frameworks as new ones become more popular and there's no work in old services anymore. Although that's still likely more than one language (php, js, sql for example)

If you're anywhere near systems administration, you'll have to deal with anything that comes by anyway, so shell, perl, python, ruby are going to be your friends.

If you just enjoy the programming and you want to learn more languages, then of course it's a good idea. Each new language will allow you to understand the other ones better and stop you from doing things only one way because you only know one way.

So the answer is as usual - it depends.

informatimago
It is definitely very worthwhile to get to the depth of at least one language and one framework, because with the scope of what you'd learn about these ones, you will have a frame to more easily understand and learn to use the others, even if you never become expert with the other languages and frameworks you will have to use.

On the other hand, since it takes ten years to become expert, you may not want to wait before starting to learn alternatives. But as a newbie, you don't want to switch to other things too early either. You should have written a couple of non trivial programs (eg. a real application) with a language before you switch to another.

angersock
Who cares how many languages you "know"? What counts is the things you've built in those languages.

That usually means that you should be focusing on one or two languages, building enough to have "used them in anger", and then figuring out what to do.

p1esk
Instead of wondering about this, pick the language you know best now (or if you don't know any, go to udacity/coursera, and take intro to programming with whatever language they use), and start working on a project. After completing a few projects, you will know what to do next.
MichaelCrawford
I regard it as vitally important that everyone learn the assembly code of at least one instruction set architecture. Not so much because you will ever need to write any, but because all programming languages ultimately result in machine code.

If your java or php or sql or haskell or python suck, the reason your users can tell that they suck, is because your machine code sucks.

The simple awareness of what goes on under the hood will make you better at any language.

Aside from that its helpful to learn several distinctly different paradigms, like python, lisp, c++ and smalltalk.

codygman
I do plan on understanding the STG, C--, and Assembler code Haskell compiles to... eventually :)
p1esk
If by "sucks" you mean "slow", it's probably because of an inefficient algorithm. Knowing assembly won't make you realize you're doing O(n^2) where you can do O(n). These days it's hard to justify learning assembly for someone who's not planning to write high performance low level code, and even there, your time might be better spent learning GPU programming or multithreading.