An interesting view and it’s certainly nice to see behavior-effects quantified.
At the same time, the Berkholz outlines a one-dimensional model where I’d see something like two or three correlated dimensions. There’s “assholeness”, the charisma, and “not suffering fools” index. Which is to say some people tell everyone off and some yell if someone doesn’t satisfy their criteria. Obviously, Linus himself is famous for being to the people who aren’t up to his standards - and to an extent I suspect this behavior actually can give him charisma with some number of people who are up to his standards. None of this is to recommend the approach, merely to say that it works.
Another thing to note: If you look at who Linus really yells at, it tends to be people who are up to his standards, and, in his opinion, should know better.
I think the Linux Kernel is the worst example for a general discussion because it is an outlier. Most contributors of the Linux Kernel are paid to work on it, there’s a huge amount of companies with a vested interest in it and also many forks that are not touched by Linus. On top of that, Kernels are not easily replaced in practice. As much as “run your company like Steve Jobs did” is bad business advice, “it works for the Linux Kernel” is bad advice for someone running an open source project. Obviously, there are different forces at work.
The only capital of a fully volunteer-run project is their reputation and the talk puts that quite nicely: a reputation is hard to gain and easy to destroy.
An interesting view and it’s certainly nice to see behavior-effects quantified.
At the same time, the Berkholz outlines a one-dimensional model where I’d see something like two or three correlated dimensions. There’s “assholeness”, the charisma, and “not suffering fools” index. Which is to say some people tell everyone off and some yell if someone doesn’t satisfy their criteria. Obviously, Linus himself is famous for being to the people who aren’t up to his standards - and to an extent I suspect this behavior actually can give him charisma with some number of people who are up to his standards. None of this is to recommend the approach, merely to say that it works.
Another thing to note: If you look at who Linus really yells at, it tends to be people who are up to his standards, and, in his opinion, should know better.
I think the Linux Kernel is the worst example for a general discussion because it is an outlier. Most contributors of the Linux Kernel are paid to work on it, there’s a huge amount of companies with a vested interest in it and also many forks that are not touched by Linus. On top of that, Kernels are not easily replaced in practice. As much as “run your company like Steve Jobs did” is bad business advice, “it works for the Linux Kernel” is bad advice for someone running an open source project. Obviously, there are different forces at work.
The only capital of a fully volunteer-run project is their reputation and the talk puts that quite nicely: a reputation is hard to gain and easy to destroy.