Open Source Can’t Coordinate
I was taking a shower this morning, and was pondering yesterday’s problem, where I suspect that I have an outdated version of hotspot Linux profiler, but I can’t just go and download a fresh release from GitHub, because hotspot is a KDE app, and I use NixOS. And NixOS isn’t a problem — it’s a solution.
Linux on desktop is a rickety tower of competing libraries, protocols and standards, which is always in an Escheresque sort of perpetual motion, taking off but simultaneously falling, and the best way to enjoy it is to take a photo, a frozen snapshot in time.
The underlying force there is the absence of one unified baseline set of APIs for writing desktop programs. There’s no single entity that can coordinate the API, in contrast to Windows and MacOS.
But then, how can Linux exist? How does that square with “never break the user space?” I’ll let you ponder the question, but let me first tell you a story from a domain where I consider myself an expert.
Better LSP Than Never
The past ten years saw a big shift in how we are writing software: baseline level of “interactive static analysis” became the norm, go to definition is universally available. The surprising fact here is that the shift occurred a decade too late!
The shift was caused by Microsoft releasing its Language Server Protocol specification. But there’s little value in the protocol itself. Its implementation is mediocre, it was strictly worse than the state of the art at that time, and its governance is abysmal. The only great thing about LSP is that it exists!
If you were anywhere near JetBrains a decade ago, it was blindingly obvious that the absence of broad availability of basic IDE features leaves a lot of the value on the table, and that the multi-process IPC architecture is the way to go (JetBrains did IPC for Rider). But it is also clear why JetBrains didn’t do LSP — why would they? While the right solution on the technical grounds, you aren’t going to get paid for being technically right. As sad as it is, some amount of deadweight loss is needed to capture some of the value you are producing, and you need to be able to capture value to invest in things! So the world had to wait for Microsoft to pick up the slack here, when they decided to gobble up the entire developer ecosystem as an investment.
There was a decade of opportunity for OSS to coordinate around an IDE protocol, but that didn’t happen, because OSS is bad at coordination.
Why Linux?
But then, why and how does Linux exist? I think part of that is a rather unique governance structure, where there’s a centralized control over the API area and strong commitment to the public interfaces. But the bigger part is POSIX. The reason why we have Linux, and BSDs, and XNU is that they all provide the same baseline API, which was defined from the outside. The coordination problem was pre-solved, and what remained is just filling-in the implementation. But there was no one to coordinate Linux on desktop.