Dragonfly BSD 1.8.1 has been released. “Security updates for BIND, File, libmagic, and TCPDUMP; X.Org added to various paths, including periodic directories for cron and manual paths; the dynamic loader now properly searches objects, solving problems with a number of pkgsrc applications; the fwe network interface is now properly dependant on Firewire; a bug in Vinum was fixed; update the EST module (CPU voltage/frequency reporting); the virtual kernel now properly handles spurious SIGTRAPs; MFC a bug fix for SMBFS which fixes a kernel panic.”
Now as they implemented many new things its high time to document them properly in their handbook.
For example NOTHING there about they great new feature Virtual Kernel Support.
Now their handbook is nothing more then FreeBSD handbook with about three pragraphs changed.
That fact need to chenge because all their revolutionary features are useless without good documentation.
You could kill Linux with this saying
I agree.
In general, the best system with the worst documentation is useless.
“In general, the best system with the worst documentation is useless.”
This is correct and has the following effect: Developers will avoid this (best) system. Documentation is the Alpha and Omega in OS design and implementation. (I’m a regular user and developer on FreeBSD and Solaris systems, so I know how helpful good documentation is.)
Good to see DragonFlyBSD making interesting progress. I wish they could get more attention. I’m a FreeBSD user, by the way.
I’ve always kept a bit of an eye on Dragonfly, although it’s not enough to make me switch, because FreeBSD does everything it should and very stable.
The only real places I’ve seen some problems is with MySQL performance, and coming in Version 7 and the tests, it scales very well and that’s been greatly improved.
I might check out Dragonfly at some stage, but you really can’t beat FreeBSD in terms of Stability, and they’ve got one of the best handbooks around.
Edited 2007-03-29 03:24
I agree that the handbook needs updating, among other things. It’s on my todo list (to send in patches/revisions), but alas, I have not yet had the time.
If you’re interested in setting up a vkernel you can try “man vkernel” … or follow this link:
http://leaf.dragonflybsd.org/cgi/web-man?command=vkernel§ion=AN…
A couple caveats regarding the man page instructions: 1. At one point you’ll have to use disklabel manually to create a partition in the root image file — although not difficult there is no explanation as to how to do this; and 2. If you want to use the virtual ethernet device you might have to make /dev/tap on the host system.
A small update, its going to be interesting to see which reaches the update in a cleaner and quicker way – although FreeBSD does have a bigger developer community, at the same time, if DragonflyBSD yields improvements with a smaller development team, I think one thing that might need to be analysed by the FreeBSD development team was whether their SMP approach was wise.
Dragonflybsd is still very immature, but I’d say once it reaches possibly 2.x or 3.x, it’ll be mature enough, and feature complete, to be able to see it being an alternative operating system for not only server but desktop use as well.
I think one thing that might need to be analysed by the FreeBSD development team was whether their SMP approach was wise.
Dragonfly have completely different goals long term (clustering support) that require a different approach. FreeBSD took, evidently, a wise road as is shown by the performace that we can see now. It just took a bit longer to get there than expected.
DflyDSB has some very interresting SMP thoughts but that does not mean it will turn out to be a better approach in the end, but then again it may be a better solution. Time will tell. FreeBSD works here and now for production work and for most of us that is what matters.