This post does seem to be summarizable as: "we should maybe start looking at the fact that these drivers crash (and recover) pretty frequently". Which is, like,

This post does seem to be summarizable as: "we should maybe start looking at the fact that these drivers crash (and recover) pretty frequently". Which is, like, a kind of low bar. :) It's interesting seeing how open-source faces the hurdles of change. X vs Wayland, OSS/ALSA/PulseAudio/PipeWire come to mind as examples, or systemd in a variety of SysV-ish shell scripts. Being willing to do the rewrite, to change the stack for good technical cause has been an interesting fulcrum, is a much harder to find similar purchase for when you're also supporting so much legacy. The task of making organizations care about the software is hard. Orgs everywhere are so driven by "shiny new things," new markets, new features. How & where we find the caring, the will to go improve things because they're in a bad way is a stunningly hard case to make apparent, to get folks onboard with. It feels like it doesn't affect the bottom line. Or put another way... company's dont really know how close they are to the Thermocline of Trust[1][2], don't recognize when instability & chaos & inconsistency overtake the organization or product. [1] https://twitter.com/garius/status/1588115310124539904 https://twitter.com/garius/status/1588115310124539904 [2] https://every.to/p/breaching-the-trust-thermocline-is-the-biggest-hidden-risk-in-business https://every.to/p/breaching-the-trust-thermocline-is-the-bi...