In December 2004, the NetBSD Project released the feature-rich NetBSD 2.0. Even after such a masterpiece, developers kept working on improvements, new features, and new ports following the new development roadmap. Federico Biancuzzi recently interviewed them to find out what they are working on and how they plan to promote their project in the near future.
“My own opinion is that most of those companies and a lot of smaller ones around the planet keep choosing Linux and other GPLed software because it’s “trendy,” not because they have found it better than BSD.”
C’mon, give Linux some credit please.
It is too bad that not too many people are aware that NetBSD is an excellent alternative to Microsoft Windows. It definitely does not get as much press as Linux or FreeBSD, but it can definitely hold its own when it comes to being a stable and secure OS. With its clean design, solid package management, and a long solid track record of security, it makes for an ideal choice for academic and commercial production servers. It also makes for a solid desktop experience as well. I have been happily using it as my primary desktop OS for the past 3 years.
come on, its not a simple thing that
NetBSD 2.0 bet Linux 2.6 in MySQL database.
so, ultimately it has much power than linux.
also, Linux community is very very large compared to NetBSD in both money wise and count wise and also many BIG giants support Linux.. but no gain…
so that clearly shows not money not community size makes an OS better. its ideas !!!. thats where NetBSD proven it..
also, all the communities NetBSD, FreeBSD, OpenBSD, DragonflyBSD, Plan 9 (i love their ideas and creativeness), Linux … start supporting L4-Hurd .
its also a great OS which is going to prove the power of Free Software …
http://www.gnu.org/software/hurd/hurd-l4.html
I’ve been using Linux for several years, and NetBSD 2.0 for one month.
I just wanted to say that for an Airis Laptop where I tried to install Knoppix, Debian Woody and Sarge, Ubuntu, trying both linux2.4 and linux2.6, I always had problems with the display, on console and X. My first attempt with NetBSD has clean display with both console and X, just as Microsoft Windows installed in other partition.
So for those boxes where Linux refuse to install properly, give NetBSD a try.
I assume you’re talking about this test?
http://software.newsforge.com/article.pl?sid=04/12/27/1243207
I’d actually say that NetBSD got demolished by Linux 2.6 in MySQL here. True it was ahead in one of the lightly-loaded cases, but it was also more significantly behind in the other light-load, 1CPU test (the sysmark one), and utterly fell on its face when trying to handle a more substantial load. Even worse than OpenBSD, which is pretty “impressive”.
Be aware that it is not clear whether the author set PTHREAD_CONCURRENCY or not. If not, he was practically doing the 2 CPU tests with only 1 CPU.
BTW. Shortly after that benchmark the policy changed to automatically set PTHREAD_CONCURRENCY when there is more than one CPU.
NetBSD simply doesn’t attract the same volume of funding, support and attention as Linux, and I would argue that this has had a lot more to do with popularity than with technical merit. But popularity is what made Windows what it is, technical merit is what made Apple where it is. Go figure.
Where must this PTHREAD_CONCURRENCY variable be set? /etc/rc.conf?
I have a quad-processor SPARC running NetBSD 2.0 so I’d be interested to know… 🙂
thats what i mean…
NetBSD 2.0 bet Linux 2.6 in 1 CPU. NetBSD SMP developement is very slow compared to Linux SMP and theres very less companies to support SMP on NetBSD.
so if the same effort given to Linux SMP is shifted to NetBSD, then think how much performance we can get thru NetBSD. but to popularity and style GUI stuffs, many turns away from NetBSD.
Where must this PTHREAD_CONCURRENCY variable be set? /etc/rc.conf?
It is just a shell variable. In current it is documented in the pthread manual page:
PTHREAD_CONCURRENCY
The number of concurrent threads to be run. This value should at least be 1, and smaller than or equal to the number of CPUs.
the reason of high increase in Linux is purely due to easy GUI installation procedure(XFree86). this XFree86 made many end users in the starting stages to try Linux which in turn made many companies to support Linux.
GPL is the main reason for the developers to love Linux.
XFree86 is the main reason for the end users to love Linux.
which became overwhelming of users which made companies to support Linux.
IF a easy GUI installer is provided for any Free UNIX OS, sure it would have spread very higher than and before Linux.
know the truth !!!
If I were creating a router or something for a company that didn’t really want to contribute back, BSD would definitely be the way to go. Linux doesn’t really offer advantages in this situation.
As an end user, I use Linux. I use it because it has better hardware support on my crappy computers, because it is easy to install and configure for a decent desktop, and because there are a lot of companies/groups that support its use as a desktop. That’s not to say that a BSD couldn’t do all those things, but there is no BSD that installs a nice Gnome or KDE desktop system with applications like Mandrake, Ubuntu, Suse, or Fedora. ACPI support on my laptop is also hard and even a bunch of Linux distros choke on it.
I’m a big fan of Debian so I’d really like to see Debian/GNU/BSD making progress because that would offer a BSD-based system that is easy to install and configure as a desktop with a great binary package management system.
thats what i mean…
NetBSD 2.0 bet Linux 2.6 in 1 CPU.
No, that’s what I was talking about. The single CPU case. NetBSD only won one out of the 4 single CPU tests, and by a fairly small margin (~10%).
NetBSD SMP developement is very slow compared to Linux SMP and theres very less companies to support SMP on NetBSD.
so if the same effort given to Linux SMP is shifted to NetBSD, then think how much performance we can get thru NetBSD. but to popularity and style GUI stuffs, many turns away from NetBSD.
Well maybe if NetBSD was better than Linux _before_ anyone started putting money into Linux, they would have put their money into NetBSD instead.
Daniel,
should concurrency be set for Intel’s hyperthreading procesor?
I do think so, but ask on port-i386 if you want to be sure.
Where must this PTHREAD_CONCURRENCY variable be set? /etc/rc.conf?
It is just a shell variable. In current it is documented in the pthread manual page:
PTHREAD_CONCURRENCY
The number of concurrent threads to be run. This value should at least be 1, and smaller than or equal to the number of CPUs.
For the MySQL case; MySQL isn’t started from an (interactive) shell, but via rc.d. Must this variable then be set in /etc/rc.d/mysqld ?
> Where must this PTHREAD_CONCURRENCY variable be set? /etc/rc.conf?
>
> It is just a shell variable. In current it is documented > in the
> pthread manual page:
>
>
> PTHREAD_CONCURRENCY The number of concurrent threads to be run. This
> value should at least be 1, and smaller than or equal to the number
> of CPUs.
—————-
Where to enable this system-wide for daemons that are not started from my shell, and don’t obey my variables?
“Well maybe if NetBSD was better than Linux _before_ anyone started putting money into Linux, they would have put their money into NetBSD instead.”
Unfortunately, due to legal issues with AT&T, and right around the same time Linux was released, people were scared to invest time and money in an operating system that may be dissolved. Fast forward to a few years later and Linux has now garnered a great deal of traction with the Open Source community and companies are starting to invest heavily into it. Before this BSD was pretty much the only game in town.
It’s not as cut and dry as you would have everyone believe.
NetBSD is a very nice operating system but I think Linux will overtake in the long term in most of the comparable feature areas. The reason for this is because the BSD licence causes NetBSD to become other systems whereas the GPL enforces the changes to be made to Linux itself and not a closed derivative.
I think both GPL and BSD licences are valuable but in terms of operating system performance, for example, MS could take schedular improvements from BSD and integrate them with windows eliminating competition between the two systems in that area. The GPL means that severe algorithmic improvements to the Linux kernel are less likely to directly aid commercial operating systems.
I don’t think you are correct in saying that the BSD license will be the downfall of NetBSD. For example, just take a look at FreeBSD: it has enterprise class features, many of which were the result of commercial vendor support, notwithstanding the BSD license. Just also look at IPv6 support: no one doubts that the BSD family (thanks to KAME) has the most comprehensive and reliable IPv6 features. Look also that the NetBSD cross-compile build ability: there’s no Linux distribution that offers such wide platform support all cross-compilable from one source tree on one platform.
So please, the GPL and BSD licensing issues are far more complex that your simple claim.
In fact, if you stand back and look, you’ll see that in parallel with Linux and GPL strength over the last several years, equally have the BSD operation systems performed in terms of offering features and other qualities. And there seems to be no indication that there’s a decline, if anything, the factors show that BSD is going from strength to strength, which renders your GPL/BSD claim entirely untrue based upon recent experience, unless you can provide any more additional indication about factors in the future you think will make a difference.
Hopefully 2.0.1 will fix the usb2.0 issues. I loved netbsd, but I love my iPod more.
I use NetBSD on a p120 system; only because my modern Linux installation cd’s would not function. (Obsolete drivers no longer in livecd kernel???!!!??!)
One thing NetBSD can improve on its installation program, which made no sense to me, something like FreeBSD’s /stand/sysinstall would be a great improvement in my opinion. I prefer no installer, like gentoo, than a poorly designed/implemented one like NetBSD and Debian.
I think the BSDs could use a user friendly graphical installer. Changing the logo to appear more “professional”, but having a archaic installer is overlooking the issue.
I prefer no installer, like gentoo, than a poorly designed/implemented one like NetBSD and Debian.
Well, actually I like to install NetBSD the “Gentoo way”. It’s not that complicated: just boot from the installer cd/floppies, ^C to the shell, disklabel, newfs, mount, ftp, untar, fstab, MAKEDEV, installboot, passwd, /etc/localtime, reboot. (A more detailed explanation can be found here: http://www.mclink.it/personal/MG2508/nbsdeng/ap-inst.html)
IMHO, an installer shouldn’t be fancy. Its task is to copy the bits and make the system bootable. Everything else (further configuration, installing packages, adding users, …) can/should be done after the reboot. And then you have sushi(8), if you like.
OpenBSD’s installer is nice, too. It’s just a shell script, it fits on one floppy, and just asks you some simple (mostly y/n) questions (except for the partitioning).
An installer dosn’t nessicarily have to be fancy to be functional, like Slackware’s and FreeBSD’s. However a fancy, user-friendly and functional installer is a salute to good design and provides a pleasent interface like Mandrake.
If I can use Mandrakes installer , which in my opinion is both well designed and functional, rather than NetBSD’s poorly designed installer, I would choose Mandrakes.
Many people do judge a book by its cover; when I see NetBSD’s installer I see something almost DOS-like in appearance; Mandrake though looks modern — even though Mandrake really dosn’t have any superior technology when it comes to graphical user experience in comparison to BSD.
If the BSD’s are willing to change their time honored and respected logo’s in the community in persuit of becoming more “professional”, perhaps they should change their installers or atleast provide an option for a user-friendly graphical one.
I don’t disagree with you foo, in fact if NetBSD supported DVB, then it would be running on this box instead of slackware. My point was that Features that BSD systems have; things that make BSD great can legally be copied easily. In terms of code, things are obviously more difficult. In theory, there’s nothing to stop closed source operating systems from using Linux’s algorithms as long as they don’t steal exact code. It’s just easier in the case of BSD because you can focus on integration with your own code rather than a rewrite.
I’ve read there are plans to integrate pkg_select http://cvs.sourceforge.net/viewcvs.py/pkgsrc-wip/wip/pkg_select/DES… (once it matures a bit) into either sysinst or sushi. I’d vote for sushi — let the installer first do its job and then launch sushi to do all the necessary after-installation configuration. Personally, the only part I’d change in sysinst is the partitioning tool — currently it’s too confusing when compared to the equivalents in other operating systems.
…easier to use BSD code and integrate it I mean
You’ve got a point there, but imagine NetBSD did have a fancy GUI installer. After the reboot, you’d be dropped at the commandline anyway, so users who get disappointed by non-GUI systems, would only get disappointed at some later point…
So a “better” installer wouldn’t “improve” NetBSD’s presence, IMHO.
<u>Unfortunately, due to legal issues with AT&T, and right around the same time Linux was released, people were scared to invest time and money in an operating system that may be dissolved.[/i]
By the time that lawsuit had ended, Linux was at version 1.1. Long before any corporate interest started in Linux.
And in fact, the first “corporate” interests that started investing in Linux were Linux based startups like RedHat.
Fast forward to a few years later and Linux has now garnered a great deal of traction with the Open Source community and companies are starting to invest heavily into it. Before this BSD was pretty much the only game in town.
It didn’t get left behind because of lawsuits, I can tell you that.
It’s not as cut and dry as you would have everyone believe.
Interesting fact for you: before IBM’s deciding to spend a billion dollars on Linux, they had far more free BSD gurus on staff than Linux guys.
Yeah right. Just watch where GPL will be 10 years from now.
I’ll tell you right now: “6 feet under!”
GPL is an insult to professional software development (and professional developers!).
I’ll give you all an example as to why this will happen. Imagine, just for a second, that car tires were GPL.
So I want to build a car. So I design everything from the engine, the seats, electrical system, etc.
After I’m done, I have to give EVERYONE my schematics and plans, because the tires are GPL.
Is this clear now?
GPL won’t be around corporate markets for long.
It sure ISN’T around our company anymore
Anyone who thinks all software should be free (RS) is a nut.
As to NetBSD being slower that Linux (LOL), NetBSD is now faster than FreeBSD, so go read the truth here: http://www.feyrer.de/NetBSD/gmcgarry/
Is this clear now?
Clear as mud.
As to NetBSD being slower that Linux (LOL), NetBSD is now faster than FreeBSD, so go read the truth here:
And what, exactly, does that have to do with NetBSD versus Linux?
Installing NetBSD on a MacPPC is a nightmare, I couldn’t get it working. Could they *PLEASE* concerntrate on getting the install working nicely – how about including the HFS utils, for example?
“And what, exactly, does that have to do with NetBSD versus Linux?”
FreeBSD is faster and performs better than Linux. So do the math
>FreeBSD is faster and performs better than Linux.
really ? maybe in 1997 …. but not now
With its clean design, solid package management, and a long solid track record of security, it makes for an ideal choice for academic and commercial production servers. It also makes for a solid desktop experience as well. I have been happily using it as my primary desktop OS for the past 3 years.
Hmz when I compiled the “native” mozilla from pkgsrc I was not able to _compile_ addons like flash or java. Only these stupid bin pkgs. So finially I gave up. I needed the box so installed FreeBSD again.
GPL is an insult to professional software development (and professional developers!).
This is why the majority of Linux commits are done by paid developers. Also this is why IBM, Red Hat, Sun, etc all contribute to GPL or LGPL software.
The GPL is not an insult to anyone, except to your intelligence.
Your patient enough to compile Mozilla from source, but no patient enough to give NetBSD a shot? The bin packages are the linux binaries so you can use flash, and java, which are also linux binaries. How do you think flash on FreeBSD works? Its not like you notice a difference between the native and linux versions.
@Adam (IP: —.nap.wideopenwest.com)
BSD and GPL have very different spirits. The first one is strongly academical (making the source available with no strings attached, just requiring the user to give credits where they’re due), the latter is strongly political (anti-proprietary, and openly communistic since it aims to abolish private property as far as software is concerned).
Professional programmers usually prefer to be *free* to stick a price tag on *their own* coding work whenever they feel like. That’s why professional programmers usually favor FreeBSD (or other *BSDs) over Linux.
IBM, Red Hat, Sun, etc are supporting Linux because 1) software isn’t their core business and 2) they have to fight the Microsoft monopoly. I really don’t think there can be a different conceivable reason for their supporting a GPL’d software.
@Anonymous (IP: —.org)
Would you please stop trolling?
Your aggressive tone and your angry attitude have nothing to do with the BSD community.
FreeBSD is faster and performs better than Linux. So do the math
Err… not quite. Nice try though.
After I’m done, I have to give EVERYONE my schematics and plans, because the tires are GPL. Is this clear now?
No, it’s not.
Plenty of companies are using GPL software as a base for their in-house projects. They tweak the code to their needs, but since they never distribute their changes (for free or profit), they have no obligations to make their source available.
In this respect, the GPL provides a large pool of free software to work with and modify, but is toothless to force companies to make public their modifications.
GPL won’t be around corporate markets for long.
As long as volunteer developers are willing to contribute under it, I can’t see companies refusing to use something that’s free. You assume that everyone who uses GPL code is in the business of producing commercial software while ignoring the fact that many companies simply want to leverage GPL and other OSS to accomplish tasks for them (network services, powering e-commerce, engineering/research, etc.)
The bin packages are the linux binaries so you can use flash, and java, which are also linux binaries. How do you think flash on FreeBSD works? Its not like you notice a difference between the native and linux versions.
So I _did_ compile Firefox from source, right.. And so I want to see a Flash Plugin And JDK to work with _that_ install. And sure I know it’s Linux stuff :: in FreeBSD I would use www/linuxpluginwrapper and java/jdk14, they’re both derived from Linux.
Well, I kind of liked NetBSD but there is no single reason to use it as a desktop OS over FreeBSD. Sure I can tweak a lot _in the end_ but that doesn’t say I like the initial setup of things too much.