Re:Project Oberon
by vlovich123 on Tuesday September 15, 2020 @01:30AM
(#24478148)
That still leaves you up a creek if you wanted to reimagine what syscalls looked like or if you wanted to build a mechanism that was totally different. That's ignoring that the Linux syscall surface area is quite large & complex with lots of legacy cruft on its own.
[ Parent
| on HN ]
You can always patch the kernel to try out your ideas.
But really, the kernel is not where the innovation potential is. The way it works is really more or less decided by the way the hardware works. That's why the only real "novel" idea in the kernel space is microkernels, which date back decades.
Worth noting that the microkernel revolution kind of happened on Android, which is a re-think-computing-from-scratch kind of project. The drivers are all moved to userspace in Android, using their own IPC mechanism. The Linux kernel is left with less to do.