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

Same here, I think small comments every 3-10 lines adds a lot of value, especially to large code bases. Being able to quickly scan through comments visually when you only need a high level overview of a chunk of code can take a 5 minute task of reading and understanding 50 lines of code into a 15 second task of reading and digesting a few small comments.

Also, I'm curious to see how your experience was working on the firefox code base. I haven't checked out the code base myself, but have built from source a few times. Curious to see how the Rust progress has been going on the code base and which parts of firefox are currently implemented in Rust.




Firefox is still largely C++ and Javascript and that's not going to change anytime soon. But it makes sense, because mostly you're adding or changing stuff that's already deeply connected with all the existing code.

I think you see Rust when there are really big projects that can be treated as a library. For instance the network team is writing their QUIC implementation in Rust – called neqo – and I can imagine that this was possible because it's low level enough and separate enough from the rest of Firefox.

It's of course also a resourcing question. For instance the download-manager code in Firefox is a big, big mess and it's not really owned by anyone. But it works, is safe, and handles a lot edge cases. Rewriting this thing (even again in Javascript) would be an enormous effort, and that's true for so many parts of the code base that might benefit from rewrites. A rewrite would be nice, but there probably will always be more important things to work on.




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

Search: