


Matz But with Haml, ERB and, the topic at hand, JIT based in LLVM, your results have been incredible.
Kokubun They just happen to be increasing speed.
Matz I see, thank you very much.
You have just begun work on your LLVM JIT compiler but as a Ruby committer do you have any work with Ruby planned for the future?
Kokubun Well I was in charge of ERB but I’ve already closed all my tickets there. There’s nothing left for me to do. I finished my task of speeding up rendering. It’s now twice as fast as when I joined.
Matz Double is incredible.
Kokubun But now there’s nothing left to do. As for Ruby’s core, I’m very interested in the area of VMs. So I started playing around with JIT and after I made a version of LLVM, and a more conservative version of Vlad’s (Vladimir Makarov) MJIT. And I found that with Vlad’s work, I wanted to release a lot myself. Vlad’s work had a lot of techniques but I found that not all of them were effective. And I thought that some of them simply made the program more complex and that it would be better for everyone to leave those out. So I wanted to work on a simple as possible JIT compiler.
Matz I see.
Kokubun I think I could help out there.
Matz I guess since he is also an experienced programmer, he may be planning too far ahead and complicating things.
Kokubun Yes, so I think someone needs to step in there and point it out.
Matz Actually Vlad’s branch has a lot of pull requests these days.
Kokubun Yes, they are very active.
Matz I think the more attention a project has, the more eyes on it, the better it’s developed.
Kokubun I think so, yes.

Kokubun Yeah, that’s about it. I’m very interested in that endeavour.
Matz I see.
Kokubun If you’re interested in your work, your productivity increases.
Matz Yes, of course. I think that’s definitely true.
Kokubun And it shows in the results.
Matz Yes, I understand.
It didn’t come up during the Kamizono interview, but the two of you are coworkers, correct?
Kokubun That is correct.