The recent fiasco involving Linus Torvalds running Linux like the privileged man that he is and entirely missing the central points of the arguments made by people who don’t enjoy his kind of privilege reminds me that there are some OSS projects that have become so large, so important, and so monolithic, that the behavior of their project leadership becomes indistinguishable from a monopoly/monopsony.

Turns out, OSS is not immune to monopoly/monopsony after all, and I think one big part of a possible solution to this problem is the same one that should be applied to monopolies/monopsonies in capitalist markets: forcing interop of smaller components.

When any project becomes a monolithic tangle, it becomes enormously expensive—impossible even—to fork. In other words, it becomes anticompetitive. To keep the ecosystem healthy, such projects must be broken into interoperable pieces that are small enough that it becomes tractable for a small group of people to fork one component at a time.

Breaking up a monolithic project into sub-projects that interoperate using standard protocols is what makes UNIX so resilient and long-lived. We need to apply that philosophy up and down the stack.

tl;dr: When an organization or project becomes “too big to fail”, that is a sure sign that it should be broken into smaller pieces, and that includes OSS.

Maybe monopoly/monopsony is actually an emergent feature of the way humans self-organize, and not merely a side effect of capitalism after all. Maybe we should be dealing with it in all organizations, not only capitalist ones.