QUOTE: a sound server is not really competition for ALSA or OSS
I'm not sure whether you think I said it was, but just in case: I didn't. It's competition for other sound servers, though, which is why I mentioned ESD (and I would have mentioned Jack, too, but it slipped my mind).
Ultimately, the way the Linux developer community has in effect chosen to pursue continuing development has come to resemble a game of darts more than a thoughtful pursuit of solid development goals. In many cases, some choices were made that were very smart in isolation, but embedded within the culture of Linux development at large they ended up getting incorporated into what looks like a plate of spaghetti rather than a software architecture:
http://blogs.adobe.com/penguinswf/files/penguinswf/linuxaudio.pngThe fact that some hotheaded self-important so-and-so decided to go off on a wild-hair expedition to produce exactly the *wrong* thing for improving the overall Linux sound architecture -- yet another component bolted into the middle of the already growing tangle of backed up plumbing rather than doing some thoughtful work on how to better tie together what is already there and organize things so needed functionality can be included in some kind of plan -- is bad enough. What's worse is that even when he (Lennart Poettering) was telling people what he had written was not production-ready, the whole community was scurrying around racing to be among the first distributions to incorporate PulseAudio into their systems. It's really just endemic to the culture now, from what I've seen. I suspect the seeds of this started with the way the BSD Unix and Linux paths began, relative to each other, where the people working on what would ultimately lead to current BSD Unix systems were trying to ensure they had a good architecture in place before throwing it to the wolves, while Linus Torvalds basically just tossed code into the wind to see what would happen. The people tired of waiting for BSD Unix on x86 jumped onto the Linux kernel and started bolting together Frankenlinux creatures, while the people who really wanted more stability and actual design either stuck with the wait for BSD Unix or eventually moved back toward it, at least in some cases.
As a result, after a couple decades of gestation, we are where we are now: Linux-based systems tend to look more "advanced", but be a lot more slapdash and disorganized in practice. That's not to say there isn't good work being done in the Linux world, of course. There are a lot of smart people doing some interesting things there. The user-facing culture (and a lot of the forefront of the developer culture as well), though, is getting to the point where compatibility with things that are already known to work well, portability of design, user-controlled system behavior, well-tested principles of architecture, and transparent organization of system operation are being explicitly and publicly denounced by the up-and-coming generation of the Linux community.
Compared to countless hours wasted debugging configuration oddities of Linux sound subsystems over the last decade, it took me about ten minutes to get sound doing everything I needed the first time I encountered it not doing what I wanted immediately after install on FreeBSD -- a fifteen second operation -- back in 2005, and since then that has only changed because of the fact that the configuration changes I made back then started being a part of the default setup.
It helps that DragonFly BSD, FreeBSD, NetBSD, OpenBSD, and smaller BSD Unix projects actually share code back and forth much more freely, productively, and regularly than any two Linux distributions I've ever seen, despite the fact that these BSD Unix projects recognize each other as separate OS projects while the Linux community seems to regard "Linux" with its hundreds of different OSes as a single operating system just because they all base things off the work of the Linux kernel project.
On the plus side, I'd still much rather use a Linux distribution (as long as I get to pick it, at least) over MS Windows or MacOS X.