Just let me debunk some of your "facts". ;-) First, the 5% you cited is for web developers - i.e. techies. If you look at the operating systems market share among average users, you'll notice that Linux has only about 1% (http://www.netmarketshare.com/operating-system-market-share.aspx?qprid=8). OSX has five times as much. Second, writing applications for Android has nothing in common with programming for Linux - completely different environment, language, API etc - and so the developers group don't really overlap. Since you've demonstrated that you know nothing about portability between unix-like systems, you're probably not a programmer, but you could at least do some reading here. Here's a homework for you: figure out how many pieces of software are there in the AppStore, how many are there in Debian, and then compare. And that's not mentioning the fact that in the last few years lots of Open Source developers replaced Linux with OSX, which is rather evident on Open Source-related conferences.

Sun didn't compete in operating systems. Neither did IBM. There is no business there - lots of companies tried and failed; only one succeeded, and mostly because of IBMs historical mistake. Suns' mistake had nothing to do with Linux or Solaris - it was all about their management staff. They avoided wasting some money on Linux, but I doubt it was amount that could make a difference. Linux kernel / Solaris userland wouldn't make porting software from Linux any easier compared to normal Solaris; from the programmer point of view, differences between unix-likes are mostly in userland. Solaris kernel / Linux userland would, though. Apple ditched Linux, because they needed something better. And that's one of the reasons OSX has several times more desktop users - and more developers - than Linux.

1. Trying to deprecate AIX was a mistake. Sun kind of did the same (remember McNealy saying how they love Linux?), but they realized their mistake a little sooner than IBM. Still, this had nothing to do with their business failure - Sun simply didn't have a clue on how to do business. 2. Linux kernel + Solaris userland doesn't make any sense at all, since technologically Linux is inferior (that's the point with IBM benchmarking, btw). Solaris kernel and Linux (GNU) userland would be better, though. 3. Apple affair with Linux was a total failure. That's why they ditched it and never looked back, instead going with their own operating system, not borrowing a single line of code from Linux kernel.

Not sure where you've spent last five years, but IBM is pushing its proprietary systems again, and with pretty good results. Did you notice that for big machines, their always publish server application benchmarks under AIX, not Linux? High-end Linux benchmark results from IBM are limited to things like HPC.

The best place to investigate a problem that manifests itself on a production machine and cannot be easily reproduced on a development environment may be that machine - especially when doing it is safe. With DTrace, it is. With e.g. SystemTap - it's not.

You're missing the fact that DTrace is safe to use, so it's impossible to crash the system just by tracing it.

You can do the same with DTrace. Read about DTrace feature called "speculation".

DTrace is Open Source, Free Software (FSF certified), thus, the fact that it's owned by Oracle doesn't really matter much. You don't need Solaris to use it; DTrace is fully functional in MacOS X and FreeBSD (in the latter, userland dtracing is available from 8.2).

Oracle can't sue, because CDDL license guarantees the right to use the patented techniques that come with CDDL-ed source code.

That doesn't include ZPL layer. Basically, it's not even a filesystem, in that you can't put any files on it.

CDDL license grants the user the right to use patented techniques. In other words, it guarantees that Oracle can't sue people using ZFS code.

NetApp did sue Sun. Their patents are on their way to being thrown out in the court, FYI.

Actualy, CDDL is years ahead of Btrfs _wishlist_. Looks like Btrfs developers didn't even start thinking about e.g. hybrid pools yet, or proper support for SMB shares (including NFSv4 ACL support), or deduplication.

Looks like some folks try very hard to notice a very simple fact: that in every case involving license incompatibility one of the parties is always GPL. There are GPL-compatible licenses, there are GPL-incompatible licenses, but there are no e.g. Mozilla-incompatible licenses other than GPL. In other words - in the Open Source world license incompatibility problem just doesn't occur when not dealing with GPL. Also, GPL-incompatibility is not the only problem with that license. GPL code cannot be incorporated into software licensed under most GPL-compatible licenses, such as BSD, Apache, Xorg, whatever. But I digress.

Might explain why Panasonic is replacing Linux with FreeBSD in their VIERA TVs (see license agreement at http://panasonic.jp/support/global/cs/tv/download/2010/down_navt.html).