> He created PulseAudio and Avahi. Neither of those are reliable, stable pieces of software

They weren't when they were first created, but since they became mainstream and other people joined their teams, I've found them to be just as stable* and far more featureful than the things they replaced

* not perfect, just not worse than what came before

They're still not. For instance, the latest PulseAudio release includes a patch that contained incomplete changes that broke various resamplers but was included anyway. It'll be reverted in the next major release which will include other half-baked changes. (It turns out that wasn't even the cause of the crashes I was experiencing; even before that patch it was broken in a way that meant just having the official volume control application open caused random crashes. I think anyway - unfortunately, all the variable names are misleading and the comments non-existent, so it's impossible to be sure what the code's meant to do.)

A minor disagreement, that software is perfectly reliable with decent hardware and perfect 3rd party software, it just does not degrade gracefully (to say the least) when things it interoperates with misbehave. If you're not doing networking or audio, it doesn't matter if buggy hardware makes PA go bonkers or weird service defs make avahi behave weirdly. A lack of error handling and error recovery.

Its a violation of the old unix principle of be liberal in what you accept.

The most likely outcome of that philosophy installed in everyone's init system is going to be a heck of a lot of systemd crashes, although it'll all technically have a root cause of some other software or perhaps dodgy hardware. None of those crashes would occur with an init system that handles errors better, so systemd is going to catch a lot of blame.

Actually, the flat volumes feature that's on by default in recent PulseAudio versions (and cannot be disabled without editing a config file as root) relies on entirely unspecified behaviour of the audio hardware, namely the timing relationship between volume changes and sample data. It has a hack that avoids deafening people unexpectedly on most hardware, but at the expense of very audible volume glitches every time it decides to fiddle with the global volume.

(Also, thanks to inconvenient analog issues like component tolerances, the most reliable way for hardware to deal with PulseAudio's flat volumes is to provide no volume controls except for a single digital one that simply undoes all of PulseAudio's weird fiddling. This volume control would provide no functionality that couldn't just be implemented in software, but PulseAudio handles devices without one really badly.)

Many open source projects are no particularly reliable. PulseAudio and Avahi are if anything well above average stability wise. They also have many more developers than Lennart in case you are fighting some sort of personal war against him.

I went to Lennart's systemd talk at Linux Plumbers conf a few years back. He's seems mostly technically competent, and he's definitely made big contributions, but his attitude ("my way or the highway") leaves much to be desired. That being said, it may be an (over)reaction to all the hate he's gotten. And I very much agree, his attitude on portability is going to come back to bite him (and many others).

Oh, good! Based on my experience with attempting to run Linux as a first-class citizen, rather than in a VM hosted by an OS which can deal with actual hardware, I feel sure we can all look forward to another decade of "You can't get it working? Oh, you must just be an idiot, then." followed by "Why does no one use Linux on the desktop?"

Because you blame Apple when OSX doesn't run on your Windows laptop?

It is not the fault of the kernel maintainers, or any of the periphery developers, that you can't put Linux on your Windows computer. You want a computer that runs Linux, buy from thinkpenguin or system76 or Dell or HP a laptop or desktop meant to run it, or build your own and do the research on what vendors provide driver support.