OpenBSD 3.7 Released

pgilman writes "It's official: OpenBSD 3.7 has been released. There are oodles of new features, including tons of new and improved wireless drivers (covered here previously), new ports for the Sharp Zaurus and SGI, improvements to OpenSSH, OpenBGPD, OpenNTPD, CARP, PF, a new OSPF daemon, new functionality for the already-excellent ports & packages system, and lots more. As always, please support the project if you can by buying CDs and t-shirts, or grab the goodness from your local mirror."

Comments

Forget that!

[Bobdoer]

Just download the new openBSD song. It's guaranteed to be 5 times more fun than the software!

Yes, you are a fanboy

[xbsd]

Linux is very ad-hoc. It just sort of "grew." It was developed in many places by many people, few of them working together with the big context of "the Linux system" in mind.

This is the typical response of a BSD fanboy when comparing his/her BSD with "Linux", not with a Linux distro. Let's do a real comparison. I'll use RedHat Linux and Debian in most examples.

OpenBSD is the opposite. People working on OpenBSD core packages have a specific kernel, userland, config script, etc., etc. in mind. There is a concept of "the OpenBSD system" and it is fairly consistent.

You can say EXACTLY THE SAME about the Linux distros I mentioned. Both RedHat and Debian have their own "generic kernels", core pkgs, etc.

The fact is, OpenBSD just does things the Right Way. People say OpenBSD's big strength is security, but that's slightly missing the point. OpenBSD's strength is correctness. From correctness yields stability, security, and all around ease of use.

Well, let see where's the hype...
Google, one (if not the most) popular search engine in the planet depends on Linux. So does Amazon.com, Earth's largest library, and MerrylLynch, one of the world's leaders in financial investments. In all cases, the stability and performance required are state of the art, and needless to say, these 3 institutions have more things to keep secure and more things to worry about than all institutions using OpenBSD combined. Just take a look at the testimonials in the OpenBSD website: http://www.openbsd.org/users.html

Now it's time to use the 2nd most popular argument of the fanboys: they use linux because of the hype.

Let's assume that three of the most powerful companies on the internet invest millions of dollars in a technology fad. Let's see what the experts are using:

The University of California, Berkeley, the alma mater of the BSDs does not use OpenBSD. Actually, they barely use FreeBSD because most computers use Debian Linux. So does the MIT, which uses mostly Red Hat Linux and Athena, its own distro. Same thing in Stanford and CMU.

NASA's Jet Propulsion Laboratory uses Linux to build better spacecraft and make accurate calculations, such as the on-board navigational computers of space probes and airborne Scanning Radar Altimeter to study hurricanes. http://www.linuxjournal.com/article/3936. They use it in the Institutional Navigation System Software (INSS) in all flight projects (Galileo, Cassini, Mars, DS1, Stardust, etc.) It contains 4.5 million lines of source code. Guess what? They use RedHat.

The U.S. Army manages personnel records for 1.2 million U.S. Army soldiers, and they access those records reliably and securely anytime, from any place via a Web interface. They use RedHat, not OpenBSD. http://www.redhat.com/solutions/info/casestudies/u sarmy.html

I can go on and on forever, but this is useless. Most of the OpenBSD fans are amateurs reading crypto books, not security professionals.

Re:Yes, you are a fanboy

[xbsd]

This is not the same. Red Hat and Debian mostly pull from upstream sources which do not develop together. For most of OpenBSD userland, the upstream is the same as the package maintainer.

I think you need to learn a bit more about the way Red Hat or Debian are integrated and how much they influence and contribute to the "upstream sources". Alan Cox is one example, he's a key kernel developer and he serves Red Hat interests, just as many other Red Hat employees or Debian devs helping out Linus. The same applies to different projects.

By the way: for every Linux distro I've used, the default kernel always lacks something or doesn't work in some way, and I always end up building a custom one. With OpenBSD, the default kernel is much better than any default Linux kernel I've seen.

That's cool, but I hope you concede that your situation it's not common. That the vast majority of Linux users do not need to recompile kernels and, as a matter of fact, kernel recompilation is way more common in BSDland, http://www.freebsd.org/doc/en_US.ISO8859-1/books/h andbook/kernelconfig-custom-kernel.html

As for your last argument, about how many people use Linux: This proves nothing. I can just as easily say, "Look how many people use Microsoft Windows! Obviously, it must be better!"

You missed the point. I never said that Linux was better because more people use it, I said that experts and corporations that really need performance, security, and overall a well written OS choose Linux over OpenBSD and I gave you plenty of examples that you are free to compare against the testimonies in the OpenBSD website. http://www.openbsd.org/users.html