One thing that really sticks out about DragonFly is that it feels a lot more like a research project than any of the other BSDs. The mailing list traffic puts me in mind of the GNU project's HURD; there is a lot of fairly high-level academic discussion about the merits of various approaches. In many ways, the two projects are mirror images of each other; DragonFly is trying to take microkernel concepts and apply them to UNIX, while HURD is trying to take UNIX concepts and apply them to a microkernel.
DragonFly had two major advantages over HURD:
- They inherited a mature, stable, and well respected code base in the form of FreeBSD 4.
- The leadership of Matt Dillon, a veteran of two decades of OS development, had the ability to match the vision.
Reading thought the mailing lists, FreeBSD users will find several familiar names. Just as many NetBSD developers contributed to both systems after Theo's fork, quite a few FreeBSD developers are still involved with DragonFly.
While still very young as an individual project, DragonFly has roots all the way back to UNIX and BSD in the '80s, both in terms of code inheritance and people. It is a mature operating system, and is being actively taken in some very interesting directions. Even if you don't use it today, it is well worth keeping an eye on.