Is Minix an academic toy?

http://minix.org/ is a simple static site seems to be hosted on a FreeBSD server (according to the HTTP headers). http://minix3.org/ is a less trivial site. It seems to be hosted on a Solaris server (http://searchdns.netcraft.com/?host=minix3.org&x=0&... ). As some other responses here hint (http://lwn.net/Articles/467937/ ), Andrew Tenenbaum does not seem to use it as a desktop.

Is the desktop usable? Let's look at the user guide:

* http://wiki.minix3.org/en/UsersGuide/WebBrowsers
* http://wiki.minix3.org/en/UsersGuide/EmailApps

Minix3 was released at October 2005: 6 years ago. If it were more than a toy, it has had time to mature. I guess it is at about the same point that Linux was when it was six years old (getting support for shared libraries). But Linux was already used for web servers at the time.

The last commit on the Subversion tree was 6 months ago (https://gforge.cs.vu.nl/gf/project/minix/scmsvn/?action=b... ).
The web site mentions that "ports to PowerPC and to ARM (Intel XScale) are underway". But I don't see any non-i386 code in trunk (https://gforge.cs.vu.nl/gf/project/minix/scmsvn/?action=b... ). Referring to ARM as "Intel XScale" shows how dated that claim is. It's so dated that the missing x86_64 port does not appear there.

And then there's one minor thing about the license: the license of Minix 3 is a four-clause BSD license. This is unlike the 3-caluse or 2-clause BSD license for the bulk of the licenses of the other BSD-s (right?). This puts it in the same boat as OpenSolaris as having a GPL-incompatible license (right?). Something with that vintage of 2005-OSes, I guess.