Good software isn't like a real-world monument; people don't spend effort to go back and revisit it later and gawk at it. Good software either gets championed by someone, or it fades forever into obscurity. (I wrote a really amazing hack years ago ... in COBOL. Maybe you've heard of me?)

Likewise, the people that are contributing to BSD right now are doing so because they believe in what they are working on. Although some of them are true cross-platform developers, many others would never make the switch. It would be agony trying to work in Linux if you're, say, Theo de Raadt. Seriously, I wonder if that might be his own personal version of hell.

I use a lot of both -- Debian, an Arch-like thing, OpenBSD, FreeBSD, and have for years. The only reason that I know that iptables has serious shortcomings is because I've used pf. If I had never used the BSDs, I would have no idea what I was missing out on. If I had twenty more hours in a day, I might even be tempted to make iptables more pf-like.

pf is kind of my go-to for BSD cheerleading, but it shines in other core areas too. For instance, Linux is really only just now getting stuff that looks like a distributed or cluster file system; but BSD has had this for quite a while already. Heck, one of the best options for a while in this area was HAMMER, developed by one guy for his personal BSD. And if you wanted something a little bit more mainstream, FreeBSD has supported ZFS for a long time -- Linux is still catching up there. Or, there was the new malloc that OpenBSD developed a few years ago, that ended up revealing old memory-related bugs in all kinds of software (e.g. http://undeadly.org/cgi?action=article&sid=2008070815522...).

There's simply no way that all those people would be doing the same kind of work in the sort of environment that Linux is fostering right now.