[Touche]

A lot of people say this and I'll repeat the answer: go look at npm's source code and tell me you'd like to "patch in" a large change such as the one Yarn takes on.

No seriously, go look at their source code. In particular go look at the caching code, one of the big things that Yarn fixes. You think their time would have been better spent hacking on that?

[throwanem]

Maybe? It'll be interesting to see how many bugs turn up in Yarn that've already been fixed in npm. Until we can perform that kind of comparison, I'm not sure it is possible to know whether building a new package manager from scratch was more worthwhile than improving the one that everyone already uses.

[agnivade]

Exactly my point. I am not totally against reinventing the wheel. Sometimes its fun and sometimes interesting results do come out of it. I am especially looking at the servo project.

But considering the amount of collaboration effort it took to build this (fb said they collaborated with multiple companies in different timezones), wouldn't it be better to spend a little more time understanding npm's codebase and get it patched ? Maybe I'm wrong, but yea just a thought.

[Touche]

> But considering the amount of collaboration effort it took to build this (fb said they collaborated with multiple companies in different timezones), wouldn't it be better to spend a little more time understanding npm's codebase and get it patched ?

Better for you? You know that open source work is free labor, right? Why should they prefer to spend weeks trying to understand undocumented, opaque and "clever" code (much of which has never been refactored) when they can start from scratch with a codebase of their own design?

What I'm asking is, what is the value to them? Or to you (assuming you ever do open source)? Unless you are getting paid for your OS work, you're under no obligation to do things any way other than what works for you.