Hacker News new | past | comments | ask | show | jobs | submit login

Writing libraries, especially implementations of well-known algorithms and data structures, is a bad way to learn most languages. Technically, it's possible, but I think it skews the experience strongly toward languages that are similar to what you've worked with before and languages that are more permissive.

I suggest writing applications, or libraries closer to applications (e.g. a specific library that parses a special data format important to your domain), where you have a clear goal but not a clear path to the goal in mind. In other words, hacking.

Let's say the author had finished the Trie. What good experience could he possibly have? It's not going to magically run faster than the C++ version or use less memory. It's a well-known algorithm so there probably aren't subtle bugs in the C++ version that rust would have caught. The best you could possibly say is that it was faster to write -- which is exactly why it skews the experience toward familiar and permissive languages.

I am trying to learn rust (albeit very slowly) by doing some hacking to make it work better for postgres extensions. I know postgres, and I think it would be cool if it were easier to write postgres extensions in rust. Of course you can today, but you'd have to use a lot of unsafe FFI calls.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: