DragonFly BSD Announces 1.0RC1

CoolVibe writes "Matt Dillon announced the availability of DragonFly BSD's 1.0 Release Candidate #1. Get it at Dragonfly BSD's site (please use a mirror or post mirrors as comments). Changes and features include: variant symbolic links, UDF support, lightweight kernel threads, message passing, GCC 3.4 in the tree, binutils 2.14, Kernighan's awk 2004-02-07, BIND 9.2.4 rc4, CVS 1.12.8, libpcap 0.8.3, tcpdump 3.8.3, less 381, MMX/XMM kernel optimizations are now on by default, greatly improving bcopy/bzero/copyin/copyout performance for large (>4K) buffers, XIO, acpica5, new AC'97 codec support, network stack revamping, long standing bug fixes for wide variety of support and stability issues, and way, way, way more. A new installer is also in the works that uses DragonFly's new CAPS IPC mechanism. The installer beta is available from LiveBSD. (Not updated to RC1 just yet, but it gives a nice idea of the progess made)"

Comments

Exciting stuff

[Dan Ost]

I think this project is a good model for how large projects should be handled.
They published their design and a roadmap for implementing their design. This
makes it easier for a lurker who is watching the project to actually jump in
and contribute to it.

At least, it seems that way in theory. Anyone have any idea how responsive the
community has been to this project?

Re:Exciting stuff

[Anonymous Coward]

Yes, it is exciting. Consider the events which led up to this accomplishment, and why Matt Dillon chose the path he followed.

Take a look at NetBSD, for instance. NetBSD is a much cleaner architecture than FreeBSD. It is increasingly apparent that NetBSD has room to grew whereas architecturally the FreeBSD 5.x series is at the end of the line. This is why Matt Dillon started the Dragonfly project. Matt attempted to re-architecture some of the worst cruft in FreeBSD. Sadly for Matt, some of the most clueless bitch slapped him unmercifully.

The case made by Matt was that the FreeBSD architecture had been hacked with undocumented junk sprinkled throughout the source tree. A change in one area ripples through the whole tree. Matt rightly said that only one or two people really understand the internals of FreeBSD anymore. Jordan Hubbard and Mike Smith both pointed this out earlier. The FreeBSD source tree is becoming more and more unmaintainable. PHK is perhaps the only person left who fully understands FreeBSD internals. Unfortunately he is considered part of the problem by many.

And the problem started when FreeBSD started chasing feature checklists trying to match Linux. This is probably the root of all the ugly hacks. In practical terms, it was Linux which was driving FreeBSD development, rather than some rational architectural plan. FreeBSD got sucked into a resource expensive "arms race" which drained the "treasury" so to speak. NetBSD stayed out of the fray, choosing instead to follow its own vision. And it has payed off nicely for NetBSD.

In any case, if there is to be a FreeBSD 6.0 someday, it will probably look like Dragonfly. I would say that future is now. Dragonfly 1.0 == FreeBSD 6.0.

Suffer fools gladly

[greenhide]

Uh, anyone want to give an idiot like me a concise and clear reason why DragonFly BSD is superior to the other BSD variants? What specific applications is it more suited towards?

No, I'm not trolling, but thanks for asking.

Re:Suffer fools gladly

[dodell]

Not only did you not read anything in the links in the /. post, you appear to have not even read the post. Go do this. In short, we're about performance.

Secondly, you appear to be looking for a reason that we are ``superior'' to any other BSD variant. When you find clear reasons why one operating system is superior over another for any given application, please let me know.

Re:Suffer fools gladly

[jazman_777]

And really, the post itself lists a TON of features DragonFly has. I'm not going to list them all again.

He wanted a "clear and concise list", not a TON of reasons. Your TON of reasons is nothing like clear and concise. This is your chance to do some nice marketing to a community of geeks. How hard could that be?

Re:Suffer fools gladly

[dodell]

And, as stated BEFORE I said that, there's a clear and concise list at dragonflybsd.org. What more do you want?

Re:Suffer fools gladly

[greenhide]

I actually did read the features that were listed in the article, but many of them were over my head.

Like I said, treat me like an idiot. Kind of like if I were to go to an auto dealer, talking to me about pistons, catalytic converters, and the like would be of no help. I'd want to know about fuel efficiency, overall emissions, power when going up hills, how much you could fit in the trunk and the back seats, and its overall safety record.

So, in operating system terms, I'd want to know things like: this operating system is designed for highly secure environments, or environments where speed and system response is most crucial, or where stability is exceedingly important, or where you deal with extremely large file sizes. Would using this OS for a web server or firewall be overkill? Is the system designed to be easy to maintain -- it did look like it has a package system that they're going to make similar to Debian's, and that would be a nice feature.

I don't know hardly *anything* about SuSE or Debian, but if I were asked, I'd respond, "Well, I know that Debian has extremely easy-to-use package management. A great number of people swear by Debian as being extremely easy to install. Other people say it's slower. SuSE comes out of Germany, so it's especially popular there as a lot of the documentation is available in German. However, if you don't know German it may be difficult to find support, as it's possible a lot of the forums will be in German rather than English." If they were interested in specifics, I'd then say, "I'm sorry, I don't know enough. Check out Google for reviews/benchmarks."

People raise "SuSE vs. Debian" type questions all the time -- it's a normal question, and it's often much easier to read responses that have been distilled from personal experience than poring through each distribution's website. I *did* go to the DragonFly website, read through the intro, but didn't understand it fully -- it was definitely written for people very familiar with BSD, and with the apparent misguided path of FreeBSD-5. I also read the FAQ. I didn't realize that I should check out the "Goals" section in order to realize what DragonFly offered differently, and then again "iomodel" is not necessarily the most appealing link name for a noob luser like myself.

Since this is one of the early appearances of DragonFly BSD, I think asking for a general explanation of what makes it so damn special is not out of the question.

Re:Suffer fools gladly

[Dan Farina]

So bleeding edge development, perhaps? I was pretty unawares of what DBSD was all about, although, to be fair, I didn't follow it. I think this is a good idea to breathe some new life into the BSDs...I'm sure successful features from DBSD will be borrowed by the others once they prove themselves useful and stable. Maybe even ease the linux troll's complaints about how slow development is in the BSD world.

Re:Suffer fools gladly

[Anonymous Coward]

There is a huge difference in developer support, which is the main reason that FreeBSD is the only fork that has reasonably modern performance features.

Sorry, but that's bullshit. With people like Poul-Henning Kamp nuking stuff they don't consider useful (see the recent ibcs/sysv compat thread on freebsd-current) and doing damage to the team work others have put so much effort on, I see more and more people migrating to NetBSD (alpha support is almost EOL'd for FreeBSD too) and DragonFlyBSD.

NetBSD's scheduler activations already outperform FreeBSD's KSE, and work on most architectures (as opposed to FreeBSD's x86-only KSE)

Sorry, but NetBSD is already ahead of the game, sans the flamers like Poul-Henning and David O'Brien.

Maxim Hermione

Re:Suffer fools gladly

[dodell]

Good for Linux. We aren't Linux, neither do we aim to be. We're BSD and by nature provide a complete OS, not just a kernel. We'll be implementing our own SSI stuff and we won't be taking any of it from Linux, of course. When our SSI work is done, we'll be able to do that too.

What's the big deal though? Who says because Linux already can do it, that it's pointless for us to implement it as well?

*shrug*

Re:Suffer fools gladly

[DashEvil]

Devin has never in my experience, been anything than nice.

Chances are, if you don't understand the technical stuff at Dragonflybsd.org, then it probably isn't for you. This isn't an insult, just a fact.

Anyhow, I don't understand why so many of you people want to be walked through this. I do think it's hilarious that when a person fails to hold your hand you claim he's an elitest, or hot headed. No, he's neither. He just has more important things to do with his time than tutor/babysit you. heh.

Re: Welcome to the same hell

[Anonymous Coward]

Linux doesn't do any of those, since it's just a kernel. Some distros might hack in something like PAX, but its nowhere near as pervasive and as well tested as on OpenBSD - it takes toolchain support for randomised shared libraries, and OpenBSD also fixed a lot of bugs in Propolice while integrating it, showing noone else has seriously tried it to use it before across multiple architectures.

Oh I didn't even mention stuff like StackGhost on sparc...