Slashdot Mirror
FreeBSD under the Penguins Shadow
An anonymous reader sent us an article about
FreeBSD, and
life in Linux's Shadow. Interesting article about
the culture difference between Linux and FreeBSD users.
Its a good one for you FreeBSD fans and you curious
Linux users alike. I wish more BSD stuff came down the pipe
here, but Linux just has the vast majority of the submissions
here too.
Couple of things you can use...
/usr/share/docs/handbook on an install of freebsd for setting this up.
/etc/ppp on recent releases of freebsd.
NATD already mentioned, it depends on functionality in the "IPFW" packet filter. There should be documentation in
The usermode ppp has NAT functionality built in, it uses the same internals as the natd code (libalias) and works well for "gotta have it now" nat on a dynamic IP ppp connection. Lots of examples in
If you've ever played with the sorta cross platform "ipfilter" (by darren reed) you can use this completely seperate packet filtering package and do NAT with it. The version that is integrated with the freebsd sources it a minor rev behind the latest release from Mr. Reed, but it does do nat, and in kernel code, as opposed to the previos two items. Much less documentation is available for this with an install of freebsd.
Currently I use Linux happily, and am thinking of giving *BSD a shot. However, I haven't found much in terms of support for my Gravis Gamepad (says it supports a standard joystick, but I want my 10 buttons :-)) Also, what about TV tuners? Does anybody have any info?
Yep, like the other said it does natd real well.
You can also use the userland ppp to do nat and
dial on demand for a network. I use it at home
with three Windows boxes and my FreeBSD desktop
machine. The server is nothing but a little 486
with 12 MB of ram and an old 450 MB HD.
It's never crashed and only goes down when I have
power failure. At one time it had an uptime of
105 days before a power glitch. It's also running
Apache.
I'd also like to know what the advantages/disadvantages there are about FreeBSD versus Linux. BTW, I'd also like to know about what the differences are betwee SysV Unices (like Linux) and BSD Unices, not merely advantages/disadvantages.
Fortunatly, thanks to VMWare betas/demo versions( www.vmware.com/) most Linux users could try the 2.2.x versions of BSD (FreeBSD 3.x in some latter versions of VMWare). (Would someone have a bootdisk with a kernel with the PCnet drivers so that a network install could be done ?)
Come on Linux users, stop disparaging FreeBSD and the other *BSD, we are all members of a same friendly family.
If you read my entire statement, I said I feel that the programmers should choose the license because they have that right. Hence, I said it was there business. As for the rest of your comments, thank you. I had no idea there was so much support. I tried an install of 2.2.8 when it was released and couldn't find most of the drivers you speak of in the kernel. Do you have to get them seperately and then install them like some kind of module or are they in fact in the kernel?
Here is the exact quote (with context! :):
."
:) In honest reality you could've done it on Linux, as we were using the linux binaries. I chose FreeBSD for its improved reliability and my happiness.
:-)
> Particularly if we can say something like "The Effects are x% more involved
> than those in Titanic (which used the popular Linux OS), but were achieved
> in y% of the time, due to FreeBSD's superior scalability. .
I doubt you can say that
I wish to point out that all freebsd committers I know look down on the attitude expressed in the quoted part. But, then again, that's from the advocacy list...
The story I know, there was 386BSD, and both Free and Net comes from that. I think that is also true of BSD/OS. And Free/Net weren't splitted "because of the usual internal strife". They were separate efforts around 386BSD, and when it became clear that 386BSD wasn't going to evolve, they originated both efforts. At the time, it was discarded joining efforts because the goals were much too different.
From my readings he certainly did use Linux, he even made a comment about FreeBSD's increased stability [over linux]. I personally routinely benchmark and test both systems and find FreeBSD to be about 20% faster than linux overall on the same hardware, and less prone to "glitches".
WinNT:Linux::Linux:FreeBSD
Standard answer is, "almost every book about Unix". FreeBSD is, of course, BSD. BSD books apply.
:).
Check out http://www.freebsdmall.com, though. There is a book session there. I find it a little bit incomplete, but it's a good start.
Also, check out The Handbook, on http://www.freebsd.org/handbook/index.html . It can also be downloaded in a number of formats, including pdf and ps, at ftp://ftp.FreeBSD.ORG/pub/FreeBSD/doc/.
Finally, http://www.freebsd.org/docs.html has everything (if you want to know what I mean by everything, check it out
Given that Linux has this walloping great amount of momentum and mindshare behind it, and that *BSD may have some technical advantages over Linux, why don't the *BSD people start contibuting to Linux and bring their mature code with them?
There's no need for "Linux vs *BSD" wars, we're all on the same side. If Linux has faults, fix them.
DG
I recently tried FreeBSD 3.0 on my old Pentium 200 and it's definitely cool (I learned Unix on DEC Ultrix, which was almost straight BSD), but there's one thing that turned me off about it: configuring the kernel.
Instead of Linux's nice TTY/ncurses/Tk driven system where each option can be toggled on/off/module with online help on each option you have to open up a text file, enter in the options you want, and pray you didn't typo. I realize people who've adminned BSD for years are probably pretty proficient at it, and you don't recompile the kernel too often typically, but it's still a sore spot in an otherwise very nice system.
Is there some tool for FreeBSD that makes this easier / less error prone?
I very much disagree. I am installing a FreeBSD box right now, having used linux for about 2 years, mostly off and on. fbsd install is not slick at all. RH allows > buttons, if you make a mistake, more help, and less likely to crash if you screw up the x-config. now that the fbsd is up, it's ok, but the install was rather rough!
A SysV nerd? Immature, definitely. ;-)
You have to configure a kernel with the drivers you want. Go to /sys/i386/conf, create a kernel file using GENERIC and LINT as reference, config filename; cd ../../filename; make depend && make all install. Done.
As an aside, I think there is a program that handles quickcam better than the driver.
FreeBSD, OpenBSD, NetBSD, etc.. They are all free and seem pretty similar at first glance.
Which one should I use? I think this is a major problem with BSD, like Linux (distros). I'd try BSD, but I am afraid of picking the "wrong" one. I'm sure each has its merits and faults.. It'd be nice if there was some sort of reference or guide to explain them..
This should be a goal of the BSD community. Explain this and I think you'll encourage more people to give it a try.
Could someone explain to me how the 150+ committers on FreeBSD are more centralized than Linus?
Mmmmm... so, we have a 20+ years code base, and we are to suddenly throw all that away to go fix Linux? And you want to know why not? Well, ok, some reasons beyond the obvious:
:-)
* FreeBSD code is good because it has evolved with those 20 years of use, not because BSD coders are better.
* Linux is too tied to GPL. GPL agenda just doesn't sit with a lot of us. In fact, some of us can't use GPL, because of business.
* Linux is too centralized. I'm not much willing to give up on my commit priviledges, and it doesn't seem to be any way of getting commit priviledges on Linux. Except, of course, creating Another Distribution.
* Linux is a kernel. FreeBSD is much more than that. What would the people who don't work on the kernel do?
* Hey, ok, let's give it a try. Would you mind replacing your VM system with ours? Ours is better. No cando? Oh, well, we tried...
Hop over to cheapbytes.com and for $13.99 get all
4 of the titles below:
NetBSD 1.3.3,
OpenBSD 2.4,
FreeBSD 2.2.8,
FreeBSD 3.1
I am a FreeBSD user, SysAd at an ISP and we are a FreeBSD shop. I bought the Linux Mondo pack for personal use from cheapbytes just to give Linux a fair evaluation for personal use and see which disto I like. Still haven't got to it as yet. Ironicially the Mondo pack has already been borrowed twice from me even though I have not used any of em as yet.
Get the pack and keep an open mind.
Reference? Like www.bsd.org?
FreeBSD -- optimized for Intel platforms
NetBSD -- optimized for portability
OpenBSD -- optimized for security
Of course, FreeBSD supports Alpha, Net and OpenBSD have very good Intel support, OpenBSD derives from NetBSD, and everyone keeps importing features from everyone else.
Dosen't that get a bit steep. What with having to fork over $80-100 for each Win9X setup?
Just curious.
Anyway we won't tell anyone your secret.
First, I'd like to point out that FreeBSD has, at the present, over 150 committers. These are the persons that have direct access to the source tree, and can directly make changes to it.
Second, if you want to contribute to something, FreeBSD has even a program to help you. send-pr.
If you still don't feel much reassured, please take a look at our contributors list. Unfortunately, we don't keep it up to date, so there is actually even more people whose contributed code has been imported into FreeBSD. You can find them in the cvs logs, though.
word up...
Maybe it is time to standardize. I think BSD would be a lot stronger if there was one flavor to advocate and improve. This whole net/open/free BSD thing is too confusing for most people.
BSD advocates argue that BSD not having distros like Linux is a Good Thing.. It seems to me that BSD is more similar to Linux in this respect than most BSDers would like to admit.
If there was _one_ "free" BSD I would know which one to try. I don't want to install three different ones. How about we take the best features from each (open's security, free's performance, net's portability, etc) and create one great OS?
That's because you're a narrow-minded geek. NEXT (and now Apple) did contribute code to the *BSD community. AFAIK NeXTStep, SunOS and MacOS X Server have not hurt the various free BSD flavours. They're still alive and kicking.
In a nutshell, I think FreeBSD/OpenBSD actually profit from their liberal license policy.
The BSD license is more free than the GPL, but the price of that freedom is proprietarization, code forking and yet another round of incompatible embrace-and-extend corporate wrangling around.
If proprietarization has a price, how much is it? When has a BSD licensed author ever lost money by using a BSD license? When has the code ever stopped being free? XFree86 was ready to proceed without The Open Group when they changed the license. The code you and I were using was still free and open.
The GPL will not save you from code forking. EGCS and GCC make a good example. The GPL does not have a clause saying forks are not allowed.
incompatible embrace-and-extend corporate wrangling around.
Microsoft does this without the code. Embrace-and-extend usually applies to the protocols not the code.
BSD appears to be willing to concede the desktop to Windows
Where does this idea come from? I have never seen mention of that.
I'm happy to leave the integration to the distribution makers, but I like the lack of central control.
I got the impression that Linux has more centralized control than FreeBSD does. FreeBSD has more people at the top controlling it. ( He re) I have no idea how many people actually control Linux. Less or more than 150?
BTW, I currently use Linux and have discovered network bugs in the TCP/IP stack as part of my service to the software community. However, when the K7 comes out, I will trying out FreeBSD on it or my older computer. I have become increasingly tired of the GPL religion since I program for fun and learning and not to belong to a church.
Sean Farley
BSD is still better kernel in terms of raw performanc than Linux? Urban Myth! It went away with the introduction of kernel 2.0.x and certainly with 2.2.x.
Please stop spreading this FUD or provide data to backup your 'facts'.
Personally, I'd be a little offended if I were working on a project and ended up having my code used by a big corporation who out-marketed my product and left me in the dust.
;)
I see this attitude among a lot of GPL supporters. I can see someone wanting payment for the software or they want the code to be free for others to use. Why do so many GPL advocates want both?
It's also why you aren't going to see 'Microsoft Linux'
I doubt you are going to see 'Microsoft *BSD' anytime soon. I don't think any of the UNIX systems have the number of lines of code that Windows 2000 (NT v5) requires to run.
Sean Farley
Uh yeah lets look at cdrom.com:
It pushes static content. No big deal there.
Uh yeah lets look at yahoo.com:
More flat nearly static content. No big deal there.
Oh and lets not forget that Yahoo is not running your out-of-the-cd FreeBSD install. They have done major tuning to get it running the way it is today.
When they started out, FreeBSD *was* more stable than Linux was at the time. It was also most likely a personal preference which is the case a lot when people choose Linux or FreeBSD or any other OS for that matter.
If you think Linux doesn't have any big-name success stories, look no further than Dejanews, Freshmeat, and Slashdot. These are some of the most impressive pieces of work I can think of
off the top of my head.
The line between FreeBSD and Linux stability is mostly bullshit now. FreeBSD advocactes just use it now as an excuse to not use Linux.
Thank you for turning this into an OS grudge match.
If you think that yahoo is static content, you are a bigger fool than I can imagine. All Yahoo pages are backended off of SQL databases, the pages are generated dynamically each time.
Have you made reliability measurements, or performance measurements? I have. FreeBSD uptimes are measured in months, Linux in days or weeks, and Solaris in years. This was between Solaris 2.5, FreeBSD 3.1 and Linux 2.2.4/2.2.5
I briefly searched through natd man page but yelded little result...
I know it can do port forward but ipautofw is a different thing. It is necessary for things like icq or iphone-ish programs to work behind a firewall.
Thanks for being more polite than the other anonymous coward in this thread, I rarely see a polite response from a *BSD'er who doesn't understand my POV...
:)
;)
...I don't see why anyone should help those who only help themselves.
My pleasure. I can say the same about you. Most times I get classified as a troll. BTW, I run Linux.
What they really want is the ability to share the code without losing their other freedoms.
What freedoms could you lose from the BSD license?
I just don't like to see people getting their operating systems stolen.
I understand the initial feeling. Remember though that parts of Linux came from FreeBSD. I may be wrong, but I believe that the Adaptec 2[7-9]40 driver was originally of FreeBSD origin. Since then, a GPL version has been written which replaces the original. Linux was able to "steal" the driver due to the BSD license. It would not be fair to only share with one and not another. If they don't share back, they are the ones with the poor morals.
I'd be happier if Microsoft *could* steal *BSD, because then maybe that would mean that they could write/sell a better OS for once.
Lay off the LSD; you must be hallucinating.
If you're not afraid of sharing code, why not make it official?
I ask a similar question often about the GPL: if you are sharing, why do you have to enforce it? Forced sharing sounds like a term the government might make up.
Sean Farley
I'm dearly tired of discovering subtle differences in every UNIX, even tho their heritage has so much in common.
I agree with this. Even though FreeBSD and Linux are both under free licenses, they do have their differences (not all subtle). Most differences are not intended to hurt the competitors as much as they believe they have a better way to do it.
Sean Farley
Am I imagining things, or are we seeing too many Linux-specific programs now? Is this a good thing? I prefer working with easy, cross-platform open source tools that will compile on FreeBSD or Linux, not requiring me to "port" from one to the other. When I see XYZ Program for Linux, I'll keep looking for something similar that's built for platform independence.
It's as bad as some of the Java apps that say they will only run on NT or Solaris. WHY?
Your thoughts?
2038 or something after that -- After all the dust has gone down, MICROS~1 is screwed, all commercial vendors are screwed, emacs is integrated with vi, and GNOME has destroyed KDE, we will have the Ultimate Holy War - the longest and biggest of all - since both sides are not evil and there are no advantages on both sides.
Of course, well before that, we will be running multi-OS emulation software microkernel that will make us be able to use all the good ones at the same time - *BSD, Linux, and Windows 3.1 (the latest good version of that crappy shell). Then this whole stuff becomes totally meaningless.
Source code for FreeBSD are available and about 75% of FreeBSD and Linux code base are very similiar. If you have to do a diff on the two source code base you'll notice a very similiar structural formation code tree.
I happen to like Linux and happen to have over 3 months of uptime. Seems like you downplay Linux uptime.
The Complete FreeBSD by Greg Lehy. You can get
it with the four CD-ROM set from Walnut Creek.
If you are getting started with FreeBSD this is
a no-brainer.
Having 20 year old code is not necessarily a feature.
Undoubtably, some of that 20 year old code was Done Right in the first place - and still more is 20 year old cruft with workarounds for 20 years of compiler developent, processor advances, and stuff that was Done Wrong in the first place. Stuff that is so deeply ingrained into the code that it will NEVER be fixed.
Sometimes, you gotta be ready to throw away code.
And as for your commit privs and the VM system - Linux is a meritocracy. If you write code that effectively replaces the Linux VM system with the BSD system, and you can demonstrate that it's cleaner/faster/better, then I have no doubt that Linus will pull it into the tree.
Free Software would be much better off if some of the BSD folks got down off their high horses, grew out of their snits, and brought all that talent and expertise to Linux.
"Not Invented Here" still rules the Ancient Unixen it seems
> I disagree. Each of their goals are different
> enough to justify 3 projects. For instance, to
> get something secure as OpenBSD is pretty much
> impossible unless you're really focused on
> securing every facet of the system.
Why can't you be really focused on securing every facet of the system (OpenBSD) and still have good performance (FreeBSD) as well as portability (NetBSD)?
> To have its crypto, it can't be developed in the
> US.
Thats right. Develop and distribute the crypto code outside of the US. The rest can be developed and distributed anywhere.
> I personally use Debian Linux, FreeBSD and
> OpenBSD, and have never thought "Hmm.. it sure
> would be better if this were all standardized"
I personally use Slackware Linux (and rarely RH or Debian), Solaris, and Irix. I'm always thinking it would be better standardized. _Especially_ the Linux distros (I'd probably feel the same way about the BSDs if I was a BSD person-- which I might be someday if I can ever decide which BSD to use)
-Anonymous Coward
I'm a 15-year-old FreeBSD user.
Demographics are meaningless.
You might've had a good point, had your argument not been weakened by your poor English skills. You're 33 and write this way? What a *perfect* example of a typical Linux user.
sounds like the boot disk used a single 'do-it-all' utility to emulate a bunch of command line programs and act according to argv[0].
:)
debian rescue disk does similar thing with program called "busybox". check freshmeat.net for a standalone copy.
I assure you that is not the standard config once you install the system
secret: Linux kernel is configred in the exact same way (with text config file)
make menuconfig/ make xconfig are just scripts which automatically generate this text file
if it's open source and it only runs on Linux, it's because YOU HAVEN'T PORTED IT YET :)
the fact is, there are more people developing on Linux than BSD, and they just can't do the testing/compiling. (I am one of them but I try and write stuff as portably as possible nonetheless)
Does porting to BSD present a problem if the program uses SysV signals and stuff like that? I know Linux tries to support both BSD and SysV APIs where possible, and I know the BSDs have SysV shared memory, but do they lack some other SysV stuff?
wherever it may be. Even in someone's commercial application. It just means that the end user gets a more stable product. Through the advertising clause, the user will always know about the open source version should they want that.
BSD coders understand that if you release source code there is the potential for evil. No amount of license posturing or flag waving can change that. An unscrupulous company can snarf up GPL code just as easily as BSD or MPL code. The only remedy for these type of companies is to NOT RELEASE SOURCE IN THE FIRST PLACE.
So the license isn't the issue. The issue is can you stand having your code misused. (It WILL wind up in a close app sooner or later, regardless of license). The BSD crowd know you really can't control source once its open.
In the end, regardless of what happens to the code, the open version always remains.
Linux is beating FreeBSD. It has been for a long time. Read the article again.
We run an assortment of BSD boxes and one Linux box. Guess which one regular uptime in the 100s of days, only to be rebooted for good reasons? Linux. Guess which ones reboot spontaneously ever so often? ... Guess which box is going to be replaced with BSD? ...
Yeah, lets look at ftp.cdrom.com:
ame (ftp.cdrom.com:alexlh): ftp
331 Guest login ok, send your email address as password.
Password:
230-Welcome to wcarchive - home FTP site for Walnut Creek CDROM.
230-There are currently 3594 users out of 3600 possible.
230-
This is on a single dual PPro-200. It saturates it's 100mbit connection.
> Being on Linux-kernel makes me feel like a
> kernel developer, even though I have never
> contributed a line of code, but the idea that
> I could has it charms me.
No, you can't - not unless you get Linus's
authorization to do so! All the people out there
working on Linux kernel have to go through Linus
first before they see their code in the next
release of the kernel - Linus decides what to
commit and what to discard. This is hardly
efficient, especially considering the amount of
code he has to inspect.
Having said that, I think FreeBSD has a better
development model, since the committers can make
changes to the source tree without requiring the
permission of some high authority. Well - they
are the authority - the very fact they have
commit privileges means they are trusted by the
core team. This doesn't mean these people are
someone's protégés, or special in some way,
it just means that they have done their fair
amount of work and contribution to the development
of FreeBSD. If you know your kernel/UNIX
system internals and think you could do that too,
then do it - submit some code, fix reported bugs,
write a new device driver or a program, port an
existing one from Linux - there's a lot to be
done! You can contribute to FreeBSD just as you
can do so with Linux. And if you're willing to
do so on a regular basis, I am sure you will be
most welcome to join the team.
Alex
Still, I admire the BSDs and would have no qualms using it. For now, however, I stick to my Linux box. It just seems like a lot more is going on with Linux, applications and otherwise.
They upgraded to net/2 and eventually 4.4BSD Lite when it came out, but much of that code has been redone since then. Since it traces it's history (ie actual code came from) back to AT&T's original UNIX, it is far older. Also, Linux's Minix style may be older than 386BSD, but the code itself had many many less years to be refined.
Hey you missed a couple:
www.imdb.com (how's that for dynamic update). imbd.com sees more traffic that slashdot, freshmeat, and userfriendly combined on the average day I would venture.
FreeBSD 3.1 has also been out the same time as the linux 2.2.x kernels, so the comparisons would be accurate. Heres a sample from our site:
fbsd1 6days
fbsd2 55days
fbsd3 55days
fbsd4 54days
fbsd5 12days
fbsd6 55days
fbsd7 55days
fbsd8 55days
fbsd9 55days
fbsd10 12days
fbsd11 20days
fbsd12 9days
Average: 36 days.
Linux1 7days
Linux2 12days
Linux3 5days
Linux4 8days
Linux5 12days
Linux6 13days
Linux7 3days
Linux8 12days
Linux9 4days
Linux10
Linux10 5days
Average: 8days.
And the linux servers take a fraction of the load of the FreeBSD boxes. We used to be a linux shop, but it is now clear to us where we need to go.
I urge people to run tests on identical hardware of both systems, we show FreeBSD topping Linux over and over. 10% on file access, 20% on network access, 20% on swap access.
That and we are tired of dealing with the Linux distros. The FreeBSD dist. has some problems, but at least it is a unified center.
Obviously it's not totally meaningless to you.
> I never said that they didn't, but I'll be
> happy to argue it now. Boy, the quality of
> discussion always goes down when we talk about
> FreeBSD. Wow, we got a few measly patches so
> that someone else could port our operating
> system without contributing back the important
> changes. I would consider a good contribution
> from Apple, say, Carbon, or some windowing
> code, or something to help us in the UNIX
> "quest for the stupid user interface". But no,
> they take your code, and release it under a
> more restrictive license, without any of the
> higher-level tools, and say that they're 'Open
> Source' on one hand. On the other hand, they
> make the rest available for a stiff fee, as a
> proprietary, closed-source product that's
> mostly just BSD where all the new features are.
> And you say "I think [we] actually profit from
> [this]". How meek you've become
Most of the code theft I have seen is done, hands down, by the linux comunity. I have lost count of the number of times that some linux "programmer" has taken code from Net/Open/FreeBSD, slapped a more restrictive license on it [GNU], "contributed" it to a linux project; then in doing all of this making it incompatible with the original. Plus the modified form cannot go back since it has been adultrated by the GPL. That sounds alot like linux not supporting the community that helped it, the same crime you accuse Apple and others of doing. Perhaps you should remove the plank from your own eye.
If someone can make a million dollars off of my code and give me nothing for it, then shame on me for not taking advantage of it first.
People also look at the GPL as this magic bullet that forces everything to be open. Not true, the Copyright holder can rewrite the license at any time, however they see fit. FSF could grant exclusive rights to all future versions of GCC to microsoft if they so wished (they have the copyright), and there is 0 you could do about it. Heck they could give *everything* to M$, and M$ wouldn't have to release the code at all if the FSF desired, and again, there would be 0 you could do about it.
Magical? You must mean "mythical".
Ran Minix & later Linux on a 386sx/25 with 1M ram.
Put NT on something like that? No way.
With load averages like that, even NT would stay up for years.
In stark contrast there is Linux. Linux is the success story. Linux is supported by the biggest names in the computer industry: IBM, Compaq, Dell, Intel, Corel, Oracle, Sybase, and Informix to name a few. Linux is the only OS beside NT to actually gain market share each of the last three years. FreeBSD, always anemic, is losing what little marketshare it had, as Linux soars to the top. But then again, FreeBSD is more of a hobby OS. Linux is an enterprise OS, backed by enterprise savvy corporations.
You don't have to be a rocket scientist or a Wall Street insider to know that the future belongs to Linux. It is just a fact of life.
It is likely that you have bad memory. Linux and
OpenBSD have different memory allocation patterns,
so crash frequency will vary.
Our "NET2" is not your "NET2". That is the obvious
name for version 2 network code.
BSD and Linux do not generally share IP bugs.
Remember the land attack?
Linux code is designed for a large-memory system.
BSD code is designed for a damn VAX.
Oh no! It's Mrs. Trouducul from grade school! I'm sorry I forgot that comma splice. I'll never do it again; I promise!
Ok, seriously now. Do you realize how immature YOU come accross by attacking Linux users in general based on someone's spelling and sentence structure? For all you know, english may not be his first language. Or maybe he was tired/distracted.
If YOU represent the FreeBSD community, then I hope to never have anything to do with any of them.
With Linux, there is one undisputed leader.
Great hackers and raw newbies alike must send
patches to Linus.
With FreeBSD, commit access is a status symbol.
This causes jealosy, envy, arrogance, etc.
Hear hear. The last thing we need is this kind of argument.
This is the Unix world. Here we use the right tool for the right job. Servers like ftp.cdrom.com wouldn't use Linux because it's the most loaded ftp server on the 'net and Linux can't handle that kind of load. On the other hand, who'd want FreeBSD on their desktop?
Let's all be happy that we have a number of quality free operating systems to choose from and not argue over taste.
Linux is an enterprise OS and FreeBSD is a
hobbyist OS? Thats an extreme view of either
operating systems and true of neither.
Some Linux distributions may become Enterprise
Ready however Linux is still as much of a hobbyist
OS as FreeBSD.
Infer from that last what you will but just
because you clearly advocate Linux does not
mean that it is actually better than any other
OS out there. The one thing Linux has is a vibrant
hobbyist ( yes HOBBYIST ) user base that has the ability to contribute to the overall evolution of the code base.
The advantages and disadvantages of this can
be raged about for years. I wonder how many
people turned to Linux merely to rage against
Microsoft. I wonder how many of these actually
contribute useful code base changes.
Jeez sometimes it looks like a school playground
with people going into gangs and drawing battle
lines just so they can mudsling at each other.
Much of what I have read here on either side
of the many fences is posted with no technical
knowledge or facts to back up arguments.
Very fine line between FUD and BULL.
If you look at the commits done one the 3 BSD :-))
projects you'll see many syncs between the trees,
IMHO forking happen only because of divergence of
opinion between developers (easily happening when one of them is RMS or Theo de Raadt
BSD and GPL permit forking equally, you missed the point.
Using a 2.1.x kernel is always going to be prone to crashes. Any kernel version where the second number is odd is a development version. If you stick to 2.0.x or 2.2.x (once that x gets past 3 or 4), your machine will be much more stable. But we know not to compare stability using development kernels, right?
Before the bickering starts, what exactly do you mean (original poster) by "raw performance"? NFS serving? BSDs probably still wins. SMP? 2.2 Linux kernel probably wins. Task switching overhead? Hell if I know. Memory management? Probably depends on specific setups, etc...
My point being (since I offered no supporting evidence), is that these blanket statements about perfomance should be taken w/ a grain of salt (and probably shouldn't be made with qualification, specification, or supportive evidence).
I use both (BSD and Linux based OSes) and like them both.
*sarcasm on* I doubt the Matrix special effects team was concerned about web serving performance. *sarcasm off*
It probably just fit the bill from a configure-and-run standpoint. Nice performance, stability, and inter-operability, etc.
If you refer to 386BSD, it sort of died due to a number of things, including the legal aspects (no one could sure that their work on a BSD derived kernel might not be nullified for legal reasons).
When Linux was released, it was viewed as a completely unencumbered (w/ regard to AT&T lawyers) Unix-like OS, and that alone probably got it a lot of attention and interest from developers. It took a couple more years for the BSD source divorce from Unix Labs to be complete.
FWIW
Could you name arrogant or jealous people please ?
Commit access is really easy to obtain if you contributed a few good things. I suppose that in the same way Linus dot not check patches from known and thrustfull people...
you are wrong.. the first poster right.
linux is backed by dozens of multi billion dollar corporations while fleabsd has what, walnut creek and no more?
and even that is in trouble. walnut creek is losing money on freebsd.. just a matter of time before they kill the project once and for all.
'No, you can't - not unless you get Linus's authorization to do so! All the people out there working on Linux kernel have to go through Linus...'
what a load of horse crap. *you* are the one who obviously needs to read up on your competitor.
as one of those "people working on the linux kernel", i can tell you that you're full of it. i submit my patches to alan cox, who puts them in linux-ac*. if the *community* approves of them after testing, they go into linus' branch.
what a load of FUD.
walnut creek can not "kill" freebsd because they don't own it.. as long as people are working on it and using it it will be alive, and that doesn't seem likely to change in the near future.
This posting is a sad glimpse into the tiny brain of a Linux zealot. It's people like you who keep us away from Linux.
I'm not sure that anyone is saying that FreeBSD is great because it's just one distribution; IMO FreeBSD is great because it's a complete OS, moreso than any one Linux distribution could be considered a complete OS (Debian comes closest in this regard, IMHO- Red Hat has little if any policy that I can tell (not that that's automatically a bad thing)).
What I don't understand, and this is not directed at the poster I'm responding to, is why so many Linux advocates feel the need to attack BSD (among many other things) in the public forums. Do they feel threatened? Are they just being jerks? Do they have nothing better to do? I don't know. I used to use Linux, as a hobbyist, for a combination of technical and nontechnical reasons. The nontechnical reasons were all of the warm fuzzy things that probably drew a lot of people to Linux to begin with. But at some point, the Linux community flipped poles: Linux advocacy stopped being the promotion of Linux, and mutated into attacking everything that wasn't Linux. At that point, all of my nontechnical reasons for wanting to use Linux just flew out the window. I know that a lot of other people feel the same way, although probably not anyone on Slashdot.
FreeBSD also has the ee editor. It's a full screen editor with a onscreen keyboard commands and pop up windows. Very cool.
From Linus' mouth, there is only maybe a dozen or so people who are a part of the core group. These guys review everything that goes into the kernel. (For good reason.) It's their final word. Now that's just for the kernel.
The rest of Linux is a little more relaxed.
FreeBSD decends from 386BSD which was an intel version of Net/2, which was a public distribution of 4.3BSD-Reno. FreeBSD 2.0 is based on 4.4BSD Lite 1, though. Later, 4.4 BSD Lite 2 changes were imported into FreeBSD.
4.3BSD Reno decends from 1 BSD (77), which was based on Sixth Edition, which decends from the original version of Unix. There you have.
As for Linux, Linux kernel exists because Linus didn't like Minix kernel (with good reason). It was not based on Minix.
At first glance, I read this as
Pretty much sums it up.
I now interrupt your comment for a dose of reality.
You're wrong, my boy. Linus never has nor will allude to a "core development group" because there is no such thing. There are people who are associated with the different architectures and different parts of the kernel, but that does not a core development team make.
You also seem to forget that because Linux is GPL, you can make your own custom patches available on the net and nobody can tell you to stop.
From the man page:
The uptime utility displays the current time, the length of time the system has been up, the number of users, and the load average of the system over the last 1, 5, and 15 minutes.
The machine could have been at load average 60 just twenty minutes before.
yeah, but say joe random company takes the *BSD source and their prodigy wize kid makes it out perform everything out there, their not going to release their modifications back into the *BSD community, whereas if the same thing happened with linux, we would all benefit and it would strengthen the OS and the community.
FreeBSD never required installation on a primary partition. You can install it anywhere you want.
linux is both an enterprise OS and a hobbist OS. Thats the nice thing and if IBM does things to make linux better, we all get to benefit from it.
I understand your reply was given because of the uncalled insults to your person, but let me explain the main difference in philosophy.
You want to free software.
We want better software.
If Microsoft released Microsoft BSD, at least I wouldn't have to run a crap operating system just to get my games.
Please note that while you can release BSD source under a more restrictive license, you cannot release it under GPL, because GPL excludes any other licensing restrictions.
It is possible to license code under both licenses, especially licensing BSD code as GPL code. Unfortunately, doing it the other way around again defeats the purpose of the GPL.
That is not true. GPL states that no other restriction must be imposed. BSD states that if you advertise the use of that code as a feature, the name of the copyright holder must be mentioned. It also says that the BSD license notice must be preserved, which I think it's a restriction also, but it's not usually mentioned in explaining why BSD can't be GPLed.
Likewise, GPL can't be BSDed.
In fact, think what would happen if TCP/IP wasn't created under BSD... We would have many different slightly incompatible TCP/IP stacks (yes, that company comes to my mind). Thus, we would never be able to achieve the critical mass that was needed to turn Internet into what it is today.
But when was the last time Sun or HP did a sync against FreeBSD? Or Microsoft?
The difference is that BSD allows proprietary code forking. And that is a whole different game than the technical disagreements.
20 years old code base is 20 years of fine tuning. Sure, sometimes you have to throw the code away because it wasn't well done in first place, and we do that.
Notice, though, that I was only answering tongue-in-cheek, half-serious/half-not, to the absurd suggestion that we ought to drop what we are doing and help Linux. If you really want to know why, consider Linux community dropping what they are doing and go help one of the BSDs. Linux has more numbers? Well, then go back in time to the time that was not so.
Sheesh, some people...
if you think IPv6 or SMP don't count as `major' changes, well, I don't know what will...
;-)
Newbus. So, when will NetBSD drop newconfig and go with newbus?
just kidding, just kidding...
Linux code is designed for a large-memory system.
;->
BSD code is designed for a damn VAX.
Hey, is that why Linux can only use 960 Mb, and FreeBSD goes much higher?
Sure it is FUD. Just to balance a little the FUD that the BSD's development model is closed.
FUD. They used FreeBSD because it was their preference to use it. Linux is no more a desktop OS than FreeBSD is. You use what you PREFER. Stop spreading this FUD that Linux cannot be used as a server. Its just LIES.
Forget about the SysV stuff. It's usually the include files, that give most trouble. Linux include files have a lot of namespace pollution, so people tend not to include the correct include files (as dictated by the standards).
That, and Linux-only calls.
Nice theory. Of course, we have only three BSD distributions, X Free and Apache don't look like splitting any time soon, and so on.
On the other hand, there are multiple Linux distributions, many GPL projects have splitted, even in things as conservative as compilers.
So, the facts do not match your theory. Obviously, because it is flawed. You compare apples to oranges. You say BSD results in splits because companies release code under non-BSD license. If you don't like the split caused by code released under non-BSD license, don't use the non-BSD license, pure and simple. It's a non-BSD license.
Err... thank you for proving everything you claim we say about the Linux community. :-)
Why Linux users are so inclined to make trolls?
First poster? You mean, you.
We usually don't go out of our way to tell people that FreeBSD is more reliable or faster. It's just what we answer when people ask us why FreeBSD.
But you probably did not read carefully enough all these messages. We like BSD because of the BSD license. We like BSD because we prefer a unified source control. We like BSD because it is BSDish.
Most people use FreeBSD because of the above reasons.
People who can actually make informed claims about stability are usually people like Yahoo, who tested both under heavy utilization. And I suppose many of the sites Linux community use as example have tested FreeBSD and found out that Linux was better. There are studies showing FreeBSD performing better (Gartner Group comes to my mind), just like there are studies showing Linux performing better. But the bottom line is that the average user, and the above average user, and even the reasonably heavy user won't notice any difference, so this is pointless.
There you have.
I'm a scientist who runs lots of CPU-bound simulations under either Linux and FreeBSD. If the binaries are compiled with the same version of gcc, they will run at EXACTLY THE SAME SPEED. The applications will spend all their time doing fp math, not calling kernel functions.
Graphics rendering certainly falls into this category, although there might be other issues related to swapping or memory paging. But if the program is well written and run on an adequate machine (not constantly swapping to disk), it will still spend 99% of its time on calculations, so any differences in the kernels will be insignificantly small. Provided the same compiler is used, you should get the same speed on Solaris, NextStep, or even the Hurd alpha (provided these could handle > 1G memory without crashing).
Software _CAN_ be faster under emulation. This is clearly the case for SCO binaries under FreeBSD. This was what motivated a number of South American banks to run their SCO DB's on FreeBSD a few years ago. The improvedn I/O performance gave them nearly an order of magnatude of throughput improvement ifmy memory serves me correctly.
SCO developers also report that there is minor but consistent performance improvement for Linux binaries under their emulation environment.
Lastly recall that when SUN was doing its initial WABI demos back in 93-94, they demonstrated that Windows apps ran as fast if not faster under WABI than they did under Windows (3.11 back then. yetch!).
Absolutely not true!
The only question is which hit the i386 platform first. 386BSD was based on the 4.3BSD Net/2 release which had been posted on the net for ages before Bill Jolitz started his series of articles in Dr. Dobb's about porting UNIX to the PC. I just checked and the first installment was publisehed in January 1991.
The claim of which was relaeased first, 386BSD or Linux is the only thing that can be argued, but who cares. They are only separated by by a matter of weeks one way or the other. After 7 years, do a few weeks really matter.
There was one striking difference in the funcionality between Linux and 386BSD at the time. If you were happy using kermit and a modem as your only form of networking, you could use Linux. I, on the other hand, was able to slap in a 3c503 card and put a 0.0 386BSD on our university network. It may have had its bugs, but at least it had networking. (Yes, it was really the 0.0 release.)
Lastly to address the BSDi/Berkely Regents v. USL/ATT legal questions, I can say that at no time were any of the 4.3BSD Net/2 derived OS's ever considered illegal. At the time this was a frustrating and dirty piece of FUD that Linux freshmen liked to propogate. Only once a settlement where each side acknowledged that the other had done more thant they were credited for and 6 files were removed from the 4.4BSD-Lite release, did the older didtributions cease to be distributible. Part of the settlement was that everyone had to move to the 4.4BSD-lite sources and that no questions or challenges by USL could ever be brought about usage of these sources.
BSD uses mbufs for the networking code.
The mbuf design saves a little chunk of memory and occasionally avoids a header copy in return for lots of boundary checks and pointer dereferences.
Boundary checks and pointer dereferences were inexpensive on the VAX, while copying headers was expensive.
We designed the Linux networking code around linear buffers. They are much faster on modern machines, but do waste a now-insignificant amount of memoryand require a copy of headers, which are usually a cache-line to cache-line transfer.
The point is that VAX-generation machines were word and pointer oriented, while current machines are very cache and cache-line oriented. Of course the Linux networking code was designed in the early '90s, while the BSD code was from a decade earlier, but that's exactly the point, isn't it?
DJB
I dual boot freeBSD and Linux 2.2.* on debian slink. What I see and feel is. freebsd is slower at disk performance but then again its probably not tuned as it should be. but then again all the bsd users refuse to help you do anything :P. :), ease of use Linux, nicer to use DEBIAN :>. Well if there was Debian freeBSD that might be good too. I dont like using freeBSD basically because I dont like the attitude on those bsd users.
:D
And I believe it does SMP faster then Linux 2.2.*. From what I see on my box. Networking ill give to Linux
Peace and Calm, happy and helpfull
Just like a Linux user
The widespread use of Linux has nothing to do
with multi-billion corporations - Linux has been
developed by a group of volunteers - and as the
number of volunteers continued to grow, big
corporations recognised this fact and decided
to join in.
As far as FreeBSD is concerned, I don't think
it is going to die anytime soon, since a large
number of big companies, including ISP's, are
currently using FreeBSD as the server OS of
their choice. Where server performance is
required, people will go with what time has
proven to be good. Linux's networking has for
the most part caught up with that of BSD, however
Linux wasn't always as stable and efficient as
it is now, and some ISP's would rather switch to
Solaris than Linux.
As for the desktop side of things, the people
currently using FreeBSD will continue to do so
in the future. They can't really lose - after
all, FreeBSD can run Linux binaries! Most open
source programs written for Linux will compile
on FreeBSD with little modification; new drivers
that are being written for Linux can be ported
to FreeBSD real quick. You can't lose with
FreeBSD. You can't lose with Linux. As more
people switch to either Linux or FreeBSD (or
OpenBSD/NetBSD), our chances at defeating
Microsoft will get better. All open source
systems are good for each other - at this stage
they are not competitors - not until Microsoft
is out of the scene.
Uhm... no FUD. Of course Linux can be used as a server. I've done it myself a few times now.
I personally doubt that Linux, as it currently stands, could handle the pressure that ftp.cdrom.com gets at the moment, but the invitation is open for someone to please prove me wrong. I won't get hurt. Promise.
Linux's kernel has sooo much hardware support.
In FreeBSD, you had to have a SB16 or nothing.
(Unless things have changed recently) I've
never seen a joystick, quickcam, TVcard, etc
driver as an option.
I do hate the companies can take the code, upgrade
it and then not give anything back.
The license isn't such a big deal to me. I think
programmers are free to choose the license. It's a right you get as the coder. If FreeBSD had the
hardware support of Linux, I'd be all over it.
Nothing's changed. If it had, we'd see Linux beating FreeBSD. I'd like to see that.
Brian Fundakowski Feldman
It's nice to see FreeBSD get some good press. If I was asked to choose one free Unix to standardize on, I'd choose FreeBSD.
I'm confused. I thought FreeBSD was based on 386BSD, which is dated June 1992 while Linux was based on Minix which was dated 1987.
386BSD of course was based on bits and peices of Berkeley Net/2, but it wasn't a direct relative of it.
Wouldn't this technically make Linux's kernel older than FreeBSD's? (as far as origins are concerned).
--
The world is neither black nor white nor good nor evil, only many shades of CowboyNeal.
Too bad most apps are made kinda linux specific so you need to wait for a port until you can download and compile... the ports system is wonderful though...
Yepp, still don't like the fact that applications doesn't compile because some strange functions that only exists in Linux...
A minor example is that pthreads are linked with -lpthread in Linux and -pthread under FreeBSD...
My current problem is x11amp, it seems that you need OSS to use it. I use the FreeBSD sb driver.
Genreally with linux emulation there is no slowdown (or negligent amounts). Some things have been foudn to run faster under BSD, even though they're linux emulated.
It's not so much of 'emulating', but just understanding the other format and having the libraries ready.
Just pick one you like for the reasons you like it and call it good enough.
FreeBSD's superior stability and speed probably made up for the slowdown due to Linux emulation.
After all, Windows apps run faster in win95/98, but that doesn't stop people from running them in WINE.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
BSD is not centralized. You have a choice of BSDi (commercial), or FreeBSD, NetBSD, or OpenBSD, all of which are Free Software. The advantage here is that you not only have a choice of distributions (as you do with Linux) but of tweaked kernels, since they all use slightly modified kernels (something you can't do with Linux's centralized kernel development).
So it would appear that Linux is actually more centralized than BSD is.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Our web server has been up for 167 days without a reboot. Our EMAIL server has "only" been up for 122 days (somebody kicked the cord loose one day, argh!).
But, of course, that's all just stories. Actually measuring reliability is a difficult task. How do you tell how reliable an OS is? Can't run it in a laboratory setting, it'll never crash there, it's in the real world that systems crash. Yet it's in the real world, doing real work, that it's hardest to monitor enough systems doing enough work to get statistically significant numbers.
-- Eric
Send mail here if you want to reach me.
Posted by DonR:
:)
I'm a Solaris and AIX sysadmin and have run Linux and Solaris at home as well. FreeBSD is my main OS now. It "fits" me.
I am a Solaris sysadmin as well. I feel the exact same way as you do. Linux is great and all, but to me, FreeBSD just somehow "feels" more right to me. Funny, this unix thing.
---
Donald Roeber
Posted by DiegoGuy:
I have used FreeBSD in the past, and what impressed me the most was the ports system.
Linux definitely needs something like this.
The way I see it, the people who need quality and speed and stability know where to look for solutions to a server enviroment ( and desktop ). FreeBSD, or OpenBSD Or NetBSD. I don't think BSD needs to be published like linux. We see examples of this "People who know, use it." philosophy on the net, Yahoo.com? ftp.cdrom.com? And Matrix ( i guess that's what the pres said )
erik!umenhofer!firebelly.net
Some of us can't afford nice lines for web serving....And I can't afford to be picky as to what IS my vhost runs. Jeeez.
Read the message you replied to one more time. He didn't say there's only one free BSD, he said there's only one FreeBSD(TM) distribution. AFAIK there is only one FreeBSD, even though there were some people talking about making a Debian distribution of (Free?)BSD.
Please alter my pants as fashion dictates.
ps -eaf | grep httpd
and get the right output. :->
I've had to work with FreeBSD a bit lately, and while I don't hate it, I can't help but feel that some things about it are just a bit old fashioned. I had just said farewell to the last of the machines running SunOS 4.1.4, at long last, and so FreeBSD makes me feel like SunOS is back, like a zombie, refusing to go to its final resting place peaceably... :-|
It *is* stable, though not terribly moreso than Linux, from what I have observed. I have Linux machines run for months with nary a hiccup on a regular basis. But, heck, you gotta install that new kernel sometime...
From what I can tell from many of the posts here, Free/Open/NetBSD's raison d'etre for some people seems to be opposition to the GPL. I like the GPL just fine myself, so that isn't a drawback to Linux for me.
That's really what it's all about. If you want your code to be free to use in any project, proprietary or otherwise, the BSD license is good for that.
But if you are primarily concerned with your code being *used* by the largest possible number of users, the GPL is the way to go.
I'm 21, and I've been using FreeBSD for 3 and a half years now (I first heard about it at Virginia Tech in 1995).
-lee
Nice article. I run Linux mostly, and I like the Sys V ish stuff. But I also work with a dog-eared SunOS 4.1.4 box that's just a peach in terms of stability. I am planning to try out FreeBSD as soon as I get my taxes paid off (damn the gov't! What tax reform?) and can actually save some money for a spare box. I'd like to have more than one type of system to play with at home as well.
"shop smart:shop s-mart" ash
What the hell...I'm just about finished w/ the minimal install to an old 486. I use Trinux on this puppy, so I'll be able to use both. So far, the ftp install - no cdrom is working fine. I kinda like the partitioning scheme. It's remaking all devices now...so I'll see what all I can find to do with it - it's only a little machine w/ 120megs and 16mb RAM. Hope I have fun!
"shop smart:shop s-mart" ash
Ah, but if someone does swipe GPL code, and we do find out, we either get a new GPL'ed app, or we sue them. It's a different mindset.
:) I just wouldn't want a BSD-style license on an entire operating system.
Personally, I'd be a little offended if I were working on a project and ended up having my code used by a big corporation who out-marketed my product and left me in the dust. That's what the BSD-style licens allows for that I don't like.
(hence, MacOS X. As far as I'm concerned, if Steve Jobs wants a UNIX, he can try to write it himself, but that's not the BSD opinion... It's also why you aren't going to see 'Microsoft Linux')
Meanwhile, if people do contribute code back to a BSD project, that's great. Just because it isn't GPL'ed, doesn't mean it's bad, I'm a big fan of WINE.
pb Reply or e-mail; don't vaguely moderate.
> > I just wouldn't want a BSD-style license on an entire operating system.
:)
... wouldn't want [this]". You can't deny that. If you want it, then I pity you, but that is of course your choice.
> That's because you're a narrow-minded geek. NEXT (and now Apple) did contribute code to the *BSD community. AFAIK
> NeXTStep, SunOS and MacOS X Server have not hurt the various free BSD flavours. They're still alive and kicking.
Thanks, you're pretty polite yourself. If I were narrow-minded, or for that matter, less cordial, I wouldn't reply to this. I am a geek, however, and for you to deny that label as well would be pretty silly by now...
> In a nutshell, I think FreeBSD/OpenBSD actually profit from their liberal license policy.
I never said that they didn't, but I'll be happy to argue it now. Boy, the quality of discussion always goes down when we talk about FreeBSD. Wow, we got a few measly patches so that someone else could port our operating system without contributing back the important changes. I would consider a good contribution from Apple, say, Carbon, or some windowing code, or something to help us in the UNIX "quest for the stupid user interface". But no, they take your code, and release it under a more restrictive license, without any of the higher-level tools, and say that they're 'Open Source' on one hand. On the other hand, they make the rest available for a stiff fee, as a proprietary, closed-source product that's mostly just BSD where all the new features are. And you say "I think [we] actually profit from [this]". How meek you've become.
I also specified that this was my opinion, as in "I
pb Reply or e-mail; don't vaguely moderate.
Thanks for being more polite than the other anonymous coward in this thread, I rarely see a polite response from a *BSD'er who doesn't understand my POV...
:) If you're not afraid of sharing code, why not make it official? Then people might *respect* you (and your code) enough to buy your product. Also, there would be no question of fairness in OSes. Does the webserver give more priority to people using IE? Submit a patch to make it even, and use the new version... :)
If GPL advocates just wanted payment for software, they would make it closed-source, and invalidate their entire philosophy. What they really want is the ability to share the code without losing their other freedoms. Anyone else can also sell the product, or distribute it free. Money doesn't have to be an issue, but one way people can show their support is by buying a product. Again, it's a different philosophy, and it doesn't always apply to all software. I just don't like to see people getting their operating systems stolen.
Actually, though, you have a good point there. I'd be happier if Microsoft *could* steal *BSD, because then maybe that would mean that they could write/sell a better OS for once. It's just that the OS is such a commodity, and such a battleground these days, that I don't see why anyone should help those who only help themselves.
I also don't see why a proprietary software company who also contributes back code shouldn't just use the GPL.
pb Reply or e-mail; don't vaguely moderate.
It is possible to license code under both licenses, especially licensing BSD code as GPL code. Unfortunately, doing it the other way around again defeats the purpose of the GPL.
:)
:)
:)
However, let's make your example theoretical. (I'm sure this has happened before, but being specific in this case only leads to bickering over facts that don't appy to the discussion
Let's say that a device driver foo was originally written under a BSD style license. Some random GPL fanatic comes along, takes this driver, adds to it, and releases *his* driver under a GPL style license. The options the original BSD fanatic has are: (1) continue to hack his own driver. (2) hack the other driver under the GPL.
If this particular BSD fanatic shares anyhow, what does he have to lose by contributing to the GPL'ed driver? If he doesn't... well, we know he has already, because of the BSD style license.
The only reason the 'enforced sharing' clause is there (I know, it sounds like something from the cold war) in the GPL is because at one point in time people *stopped* sharing, and started turning free applications into proprietary ones. I don't want to take sides here, but I'm sure it started somewhere around RMS writing Emacs and other companies adding to it without contributing back. Since they violated an unwritten tradition, (around the BSD-style licenses) the GPL was born.
If people always shared their code, there would be no GPL, and there would be much rejoicing. They don't, so the people who don't want to see their code used by other people without the benefits of the additions to that code use the GPL. Those who are trusting of human nature, or want to improve other people's code without necessarily improving their own use a BSD license.
I guess it comes down to if you want to help others, or if you want everyone to help each other. Therefore, it's a matter of opinion.
Aww man, you run Linux too? So you mean that I *still* haven't had an intelligent conversation with a BSD user?
Actually, I tried a boot disk with FreeBSD on it, and I liked the kernel configuration, that was slick. However, I miss all the friendly options from the GNU utils. And I thought that both the device layout and the way all the system utilities pointed to one big executable were very strange. I like to know how much space ls takes up. I hope this isn't a standard configuration, but rather something done for this boot disk. However, it was odd.
pb Reply or e-mail; don't vaguely moderate.
That same person could very well contribute back the same changes to the *BSD community, however this behavior is allowed by the *BSD licenses, which is what i was saying I didn't like in the first place.
And yes, if you were concerned about someone making a million dollars off of your code, and givving you nothing for it, then you would have licensed it under the GPL, not under a BSD license.
The GPL does force everything to be open when the copyright holders are many and disparate. This is why the linux kernel is safe. And even if Microsoft had rights to all future versions of GCC, they would never have exclusive rights to the current or previous versions, and therefore development would be unchanged. However, I really doubt the FSF would do this. (Over RMS's dead body...)
Anyhow, let's stick to licensing issues here, this is getting silly. There already is nothing to stop Microsoft from stealing all the BSD code they want (except not knowing what good code looks like), and there is something to stop them from stealing GPLed code, given the sanity of the copyright holders.
Therefore, I don't see where this argument is going, if anywhere...
pb Reply or e-mail; don't vaguely moderate.
I'm using Linux on my workstation because Linux had support for the most outlandish hardware. Linux is much more bleeding edge ( suprise) and (in my opinion) has a number of very good devicedrivers (take Donald Beckers line of NIC drivers). If there is a little tweak that will make a device go 5% faster, than it will be used. It does make the Linux kernel a bit less stable. However, if you stay away from 'new' and 'experimental' kernel options, that is usually not a problem.
I use FreeBSD as a server. The kernel is more modular (especially NetBSD/OpenBSD) and the source is easier to understand. FreeBSD is usually more a coherent mass (like the article states). A small example: glibc 2.0.7 implements writev(2) with write(2), while the linux kernel supports writev(2). Not very efficient. However, this is more of a small problem with glibc than a problem with the linux kernel. (Maybe the Hurd kernel doesn't implement writev.)
Mathijs
FreeBSD, OpenBSD, NetBSD, etc.. They are all free and seem pretty similar at first glance.
;)
That's because they are, especially NetBSD and OpenBSD. Most of the source for OpenBSD comes from NetBSD. (OpenBSD is kind of a paranoid version of NetBSD.)
FreeBSD runs on Intel and Alpha CPU's and is probably your first choice if your using a x86
machine.
NetBSD and OpenBSD run on just about any CPU, so if you've got hardware like a SGI Indy or a nice Sun, you want to get one of these.
The difference between NetBSD and OpenBSD is mostly that OpenBSD is more security minded (hope Theo doesn't see this
FreeBSD has a bit better support for x86 hardware than NetBSD/OpenBSD.
In the end, these BSD variants have more similarities than differences. It doesn't really matter which one you choose. It's more a matter of taste, I guess.
If your using x86 hardware, go with FreeBSD. If you're really 'paranoid' (are is that 'sane'?), go with OpenBSD, since it tends to be more secure.
Hope this helps.
Mathijs
And, FreeBSD's level of "Friendliness" and relatively rapid evolution probably wouldn't mix well with the breadth of architectures supported by NetBSD.
I personally use Debian Linux, FreeBSD and OpenBSD, and have never thought "Hmm.. it sure would be better if this were all standardized". That reeks too much of mob rule, force and mediocrity for my tastes.
I've always viewed Linux and FreeBSD as being somewhat equals. Linux is better in some areas (e.g., driver support) and FreeBSD is better in others (e.g., I hear FreeBSD tends to be a bit faster for server-type applications).
That said, I think the main reason I run Linux on all of my computers is because five or six years ago, Linux is what installed easier on my 386. I downloaded both Slackware and FreeBSD. I never could get FreeBSD running right on that hardware, but Linux did, so it stayed. The university I graduated from used FreeBSD on a LOT of lab machines. From my limited usage of them (just for compiling projects and what not), it behaved just like all of the other UNIXes I've used.
I wish the article would have addressed more technical issues. I want to know specifically what FreeBSD does better and what Linux does better. Of course, I'll support anyone's choice of either, I love Linux and I've heard few bad things about FreeBSD!
I'm getting a new second-hand computer this weekend (I'm playing the "how many computers can I fit in an apartment game")... I think I'll stick FreeBSD on there just give it a go.
We run a farm of over 80 FreeBSD servers pumping over 150Mb/s (the cam shows the routers).
And on the age thing... I'm a 38 year old network engineer who runs SuSE on my desktop but work for a 25 year old that runs this FreeBSD farm. Go figure.
-- I have a private email server in my basement.
You have to compare FreeBSD more to a Linux distribution, than to just the Linux kernel, although this isn't quite accurate either.
The *BSD developers hack on the kernel, the install system, the package system and a large number of the utilities/programs that are distributed with their system. They also do "ports" of packages that are outside of their control.
Compare this with, say, the Debian GNU/Linux which has 400 developers that can commit changes to their distribution. These 400 developers mostly just do ports, although some work on the install and packaging systems.
The kernel that Debian uses is based off of either the development tree or the release tree of Linux, depending on which the Debian folks thinks is the best thing to do at the time. They also add patches that they think are appropriate and select additional device drivers. The other Linux distributions do similar things with their kernels. So, while there is one central body that controls the FreeBSD kernel, Linus has a lot less direct control about what gets put into the Debian distribution.
To the best of my knowledge, FreeBSD only has one "vi", while Debian has at least three, each of which has a Debian maintainer, and the upstream developers. The Debian user can easily choose whether they want vim or nvi as their editor and the upstream developers have very little control over which they choose, or how the different vi's get packaged by Debian.
While FreeBSD has much more central control than the Linux distributions, I'm not sure that this is really A Bad Thing. It has plusses and minuses.
In the case of the Linux kernel, Linus having a very central control of it has worked well, as has FreeBSD. On the other hand, Debian's freedom to pick and choose which kernel will work best for them frees Linus up from having to worry about the release schedules of the distributions. He, and the rest of the Linux kernel crowd, can worry about developing the kernel, the same goes for the developers of glib, lilo, binutils, etc. It is up to the various distributions to make sure they all work well together.
SPF support for most open source mail servers can be found at libspf2.
...the fact that it requires a primary made it a BITCH for me to find room for it. I ended up installing it on a secondary box because my main box has no spare primary partitions at all. Overall, though, the FreeBSD installation software is pretty good. :-)
BTW, I agree about Jordon Hubbard. Very classy guy from what I've seen online. I know a number of Linux folks who could learn a thing or two from him in terms of how to advocate a given platform.
--
-Rich (OS/2, Linux, Mac, NT, Solaris, FreeBSD, BeOS, and OS2200 user in Bloomington MN)
Mainframe/UNIX Bit Twiddler and long time Windows/Linux Hobbyist.
The Theorem Theorem: If If, Then Then.
You've said that before... but you've never explained why (anywhere I've seen). Care to enlighten me now?
Ta.
--
I admire their choice of OS for the Matrix, but I'm wondering if they had good reason for their choice? If they were going to choose a flavor of UNIX, why did they choose FreeBSD over other Unices? If it was a matter of the free availability, why did they choose it over Linux?
I would be curious to find out these answers. Does anyone know?
"In true sound..." -Agents of Good Root
From this, it looks like he was only comparing it to Irix. Perhaps Irix and FreeBSD were all he wanted to choose from. I can't imagine Renderman performing better under "Linux emulation" on FreeBSD than natively on Linux. I take it there isn't a FreeBSD port of Renderman (too bad).
"In true sound..." -Agents of Good Root
I would agree with you, but that was before I tried Debian. I've been pretty impressed with the Debian packages system. There are some minor tradeoffs between the two implementations, but they are both very similar in terms of power.
--
Jake
... that there's only one "distribution" of FreeBSD compared to a lot for Linux (RH, SUSE, etc), also FreeBSD is a BSD standard, program written following the BSD rules always compiled fine under FreeBSD. Some years ago i remember at university we installed a network of FreeBSD to have a standard, and also because we could find tons of books about BSD standard and nothing about linux (not true today :o). Anyway i used a lot HPUX from 8 to 10 and it's a mix about all standard...a -morron" :-( :o)
what i regret also about linux is that linux users sometimes are "LiNuX-is-better-than-your-fucking-OS-so-you-are-
let's live in a free os community
i don't want to start a war between FreeBSD and Linux, use what you prefer! you? MacOS, great! you? BeOS, great! etc
--
"Science will win because it works." - Stephen Hawking
Of course, rather a lot of pre-5.x SunOS (and probably rather a lot of stuff in other UNIXes people might consider to be "BSD") came neither from BSD nor System V, but from the vendor of the OS (the SunOS 4.x VM system and dynamic linking system, for example, were designed and implemented at Sun).
I really wish people would either stop commenting on performance/reliability of Linux vs Freebsd.
Or post some extensive test results to support their claims.. I have yet to see anything that even comes close to this. (Closest I have seen was in the freebsd zine, which compared redhat 5.0 running 2.0.34 to freebsd 2.9?(8?) which showed them both to be about equal.. with freebsd possibly ahead by a nose)
I think it shows that people rely to much on "what they think" based on personal experience.. This is fine in of itself, but when people make generalizations like "freebsd is more stable then linux" or "I have found that freebsd is 20% faster then linux overall" It would be nice to see supporting facts.
I use/have used both for a number of years and I can't in all good conscious say one is faster or more stable over the other.. My personal experience of stability with linux or freebsd is GOING to be different then someone else. I most likely have different hardware and have used different distributions / versions of them both. So for me to make a general statement would be pointless.
If this is so important to some people I would suggest there be a test procedure drawn up on how to test these things.. perferably using a current configuration of each. Having a "expert" from both camps on hand wouldn't be a bad idea either.
-- You can be a geeklord too
The development methods for the two are so radically different that I don't think you can compare the two by the amount of code freezes.
Really, There might be more bugs in a linux kernel releases but there are more drivers, and more features. More people also use the code, and thus bugs becomes more apparent.
As you stated you're "personal" opinion is based on an experince in 1994. I think we both remember the state of linux in 94. Freebsd on the other hand took less time to mature because it was based on an existing code base.
I believe times have changed since then.
Just to give you an idea of the servers we have around here and the uptimes:
15 alphas (osf 4.0) ranging between 4000s, DS20s, and 8400s. (with ranging uptimes of 30 days to a year)
8 linux boxes (all redhat 5.2) (3 are servers with uptimes exceed 150 days, 2 are routers with uptimes exceed a year, rest are workstations)
3 BSDI boxes (uptimes ranging from 30 days to 100 days)
12 suns from sparc20s to ultra enterprise 5000. (varing degrees of uptime)
5 AIX RS6000s (30 days to 6 monhths) (they are new)
misc other servers..
Reguarding uptime it has been my experince that the uptime of all the above operating systems are directly related to the following:
a) hardware used
b) software used
c) strain
Again my opinion is if its important we need someone who is willing to put there money where their mouth is and show some statistics.
If you run linux on a packard bell its not going to get high uptimes. If you buy a server from var research I think the odds are you won't run into situations where you need to reboot often.
-- You can be a geeklord too
I have found FreeBSD's installation tool to be very simple. The first installation I did with it years ago was much easier than my first RedHat installation just months back.
So, I don't think installation effort is much of a claim against FreeBSD at least. I haven't played with other *BSD systems.
scottwimer
-- Beer. It's what's for breakfast.
I think the reason Linux is so much more popular than *BSD is that Linux was free first. If I remember correctly Linux was about 6 months old when the announcement was made that BSD source code was going to be released. Then there were some legal problems between Berkley and AT&T which took a while to resolve. Being first, even by a tiny amount, has tremendous market implications.
Actually, this isn't the case. There were articles about the development of a freely available BSD system in Dr. Dobbs Journal in 1990. Linux wasn't available until late 1991.
Although it's not FreeBSD specific, I highly recommend the 'Unix System Administration Handbook' (http://www.amazon.com/exec/obidos/ASIN/0131510517 /distributednet - This link will help distributed.net out btw). Very good book, a must for any unix admin. (Read a slightly larger review at http://www.distributed.net/research/).
Moo!
dB!
http://www.distributed.net
I am 33 years old, and I love Linux, *BSD holds no interest to me.
Age does not = being conservative the thought is just silly as hell.
I rather like the idea of using GNU tools first, *BSD does not, to my understanding.
Being on Linux-kernel makes me feel like a kernel developer, even though I have never contributed a line of code, but the idea that I could has it charms me.
Systems that are controled social chaos (meaning Linux) tend to be more stable, over the long run then, forced ridgedy (*BSD) are in the long term.
Last Linux gives more choices, if you like infinte control chose Slack, Ease of use chose SUSE, stablity Debian, Or Out and out performance Stampede.
There all fun, slightly chaotic, and the right choice for you.
"Think of it as evolution in action."
I've been told that I'm genetically clue inhibited but...
I have Red Hat pre-6, Debian 2 and Slackware "4" running on various boxen at home and plan on trying to pick up a copy of FreeBSD 3 at Comdex on Monday.
There's so much that's good and so much to learn from all of the work going on that I feel like a kid in the world's biggest candy store (kinda like a Virtual Frye's for software) and when I grow up (I'm 42) I'm gonna be like Linus and Alan and Jordan H., etc.
I guess I just don't get the O/S Holy Wars. I'd ask for an explanation but my flameproof Fruit of the Looms are in the wash right now.
Destined to be clueless for life I guess... (*sigh*)
non-newbies also need a choice too. Please don't turn all OS's into newbie OS's. I think newbie's need an alternative to Windows but that doesn't mean we should dumb down unix -hurting those who love unix the way it is now- just to give newbies that choice.
---
I was wondering, given that FreeBSD is sorta
centrally controlled, could it not be that it
does nor engender the same sense of community?
Do we not feel that Linux is "my" OS regardless
of whether we are developers or consumers. I'd
say that a lot of this is due to the lack of centralization and some of it is due to the license, the GPL, which provides a sense of
protected ownership. The GPL makes us feel that
Linux is OUR intellectual property, it belongs
to all of us. Perhaps knowing that their work
will not be ripped off (read tcp stack) does
motivate some developers.
The lack of centralization makes us feel that we
can all make a difference. Its hard to see a
tightly controlled source base giving people that notion. While a lot of this is just our notions
(surely one can compile KDE for linux without ports, or whatever), our notions and emotions do dictate what we do.
In fact, centralization makes some of us uncomfortable, which is why, for example, some people do not want companies to standardize on RedHat. Finally I do believe that the fanaticism and evangelism, dangerous in its extreme form, is a direct product of this sense of community.
The Inscrutable Gargoyle
The joystick programming is up to the end program to poke at /dev/joy0. As for TV tuners, anything with the bt848 or bt878 chip is supported. I use my Hauppauge WinTVpci card all the time to watch TV!
You can't steal what somebody gives away for free. The UCB/BSD-style license gives the code away for use in whatever way you feel like. Please tell me how this is less free than GPL.
Like I've said a dozen times before, BSD-style licensed code is NOT "ripped off". You can't rip off something that is given away free to be used however the user sees fit. Why do you GPL advocates insist on twisting reality like that?
Ah, more FUD. Perhaps the patch was sloppy or wasn't well bug-tested? Ever think of that? Grow up. The core developers are there to prevent sloppy crap coding from getting into the FreeBSD kernel, and I'm glad they are there.
NetBSD's kernel isn't split along port lines, only architecture specific code. NetBSD/sparc and NetBSD/mac68k, for instance, both are built from the same source tree.
Wow, I see that the old "Linux or nothing" attitude is still rampant in the community. Some of us just don't like the Linux development model. Some of us think that *BSD just has it right in the first place. Why don't you Linux users get off your high horses and bring your talent and expertise to *BSD? Just leave your rabid GPL cheerleaders behind, thank you. :)
Nice, resorting to childish architecture attacks? BSD orginially ran on the PDP and VAX series computers, yes, but now it runs on more different types of computers than Linux does, and does a good job of it.
Hi, how is this any different from how FreeBSD works? If you can directly commit, go ahead. If you can't, use send-pr. If you're unwilling to do so, then put out a patch. Nobody will tell you to stop. You can even go one step farther; make your own version of BSD based on FreeBSD's code! Nobody will stop you there, either.
If somebody writes some code themselves, what right do YOU have to it? The BSD-style license gives the code away for free, no strings attached, no morality or ethics imposed.
Wow, you're badly informed. FreeBSD certainly has plenty of hardware support. I have an SB AWE64 which works fine with a joystick connected (I don't use it for anything, but is there as /dev/joy0). Quickcam has been supported for quite some time (qcam driver). I have a Hauppauge WinTV card that is very well supported; even made some MPEG movies using it. As for companies "stealing" code, what business is it of yours? How do you know companies aren't swiping GPL'd code and not telling anyone? Several companies have used FreeBSD code in their products and have even given back code to the project (Whistle, for one). I know rabid GNU cheerleaders tend to pretend that this doesn't happy, but sorry folks, it does. You don't need a license to force people to dump code back into a project.
I'd have to say unless I drasticly change I'm a linux guy. When I program I use sysV changes that are supported by linux. I use the /opt fs for extra software on my system.
I have a FreeBSD 2.2.3 system and I think it is a great stable operating system. I off load work to it and use it to check compatability of code I write. But that is just me.
I do appreciate everything that BSD gave unix, vi, biff, and so on... (I would say etc. but...)
I remember when I had Linux (0.99.x, baby!) running on a 386 w/ 4mb - now that was a sweet setup. I even had it running a MOO, and it was plenty responsive. Now mind you, I did get it to boot and run Win NT, so it may have been some sort of magical box.
Well, running NT was a mistake - I was installing a hard drive from a "top-o-the-line" 486 and booted it by mistake. I have to say I was quite shocked - it was slow as all hell, but I could run Eudora from across the network.
The stability argument holds true because FreeBSD unlike linux has a lot more code freezes and those freezes last for a lot longer timer.
:) one 95 based ACT server.
F /...
I used to use linux a lot but after running into tons of ext2 problems and scsi issues (this was 94 mind you) the servers I was responsible for were migrated over to FreeBSD. I have stayed with FreeBSD every since... I still know linux and still use it in professional situations but if it comes down to what I feel safer *trusting* then FreeBSD gets the nod from me. Of course this is my opinion but it is based on years of experience and for me (and the people who pay me) it seems to suffice.
Just to give you an idea of the servers we have around here and the uptimes:
2 alphas (osf 4.0/3.2) 187 and 207 days uptime.
3 linux boxes (all redhat 5.1/5.2) 14/21/54 days. The 14 and 21 are from hardware upgrades.
2 FreeBSD boxes (3.1/2.2) 26/115 days.
and for comparison
Averages 2-3 days before reboots.
---
Openstep/NeXTSTEP/Solaris/FreeBSD/Linux/ultrix/OS
--- I do not moderate.
I love the BSD style of kernel creation. Maybe thats cuz I use vi for everything but it takes me no time to zoom through and change/comment what needs to be done. On a side note if there isn't anything, it would prolly take 3 minutes to hack a perl script up that would go through the LINT file asking Yes/No/Change style questions.
/usr/src... all I had to was get it outa da kernel and edit. Nice!
F /...
I am really glad that there is now the option to include the configuration file as a string inside the kernel. This actually saved me a ton of time the other day on a system when I had popped out the drive that contained
---
Openstep/NeXTSTEP/Solaris/FreeBSD/Linux/ultrix/OS
--- I do not moderate.
Uh. I am sorry but this is not clearly thought out. The main point that you raise is that "all good features eventually spread to all distros"... uhm no offense but this *does* happen in the BSD camp. It has happened since the initial BSD forks and will continue to happen. This is because the aim of the BSD camps is to have the best distro they can, so of course they implement the best features and lose the worst.
F /...
The GPL has nothing to do with this.
---
Openstep/NeXTSTEP/Solaris/FreeBSD/Linux/ultrix/OS
--- I do not moderate.
This posting is quite full of inaccurate statements. I am assuming you either mean it as total flamebait or your lack of *BSD knownledge is quite high. Luckily for you ignorance is correctable.
F /...
Your posting is exactly what the article was talking about as far as extreme loyalty to operating systems. So full of rhetoric and obvious love for linux it amazes! This is a beautiful example, that I am sure you shrewdly conducted, to aid those out there who didn't understand what all this evangilising was about! Kudus!
---
Openstep/NeXTSTEP/Solaris/FreeBSD/Linux/ultrix/OS
--- I do not moderate.
This is not a flame, but an honest question from someone with very little BSD background (other than legend.)
Why the splits between Net, Open, and Free, and god knows what else BSD? Why do people say "FreeBSD is great because it's one distribution" ignoring the existence of OpenBSD etc?
I just don't understand.
On a totally unrelated note, is there any interest in a Python script that makes automating URL submissions to a variety of web search engines very easy? I've written one recently and I'm wondering if there's any interest.
Email me in private if you like (remove the SPAM-B-GONE.)
If it was a matter of the free availability,why
did they choose it over Linux?
FreeBSD is just as "free" in terms of price, just as easy to download, and arguably freer in terms of license.
Perhaps it was the superior performance?
There was a link on Gartner Group's site (now removed) that described FreeBSD's superior web serving performance.
SunOS 4.X (and earlier) is a direct BSD descendent (given the Bill Joy connection, one might say that Sun was more BSD than BSD). But SunOS 5.X was no closer to BSD than other SysV R4 varients.
There are some old Sun hands out there who still haven't forgiven them for abandoning BSD.
I've been thinking about moving a box that I have set up as a web server/database server/ip masq box over to FreeBSD. Does anyone know if FreeBSD has something similar to ip masq?
This sig is false.
Can someone suggest good (and up to date) literature for FreeBSD? I've had lots of success with Linux, partially because I've had good friends to turn to the few times I've had major problems. I don't know anyone, however, who runs any BSD systems, so some printed help would be good.
This sig is false.
FreeBSD (most x86 optimized)
NetBSD (most portable)
OpenBSD (most secure)
http://www.bsd.org
Romans 10:9-10
Slowdown ?
Where ?
Do you guys even know how Linux emulation works on BSD machines ???
It's just a question of having a kernel with COMPAT_LINUX compiled in. Then processes are tagged from start, and use another set of stubs for system calls.
The `slowdown' is *dwarfed* by the context switches, which is the reason system calls under Unix are a bit slow (and hence, that there are not that many of it). The only inefficiency is *memory* since you need to have another set of dynamic libraries loaded.
I've been checking Linux emulation with such programs as xanim. There is *no measurable slowdown* from the native version.
The only programs that perform slower under BSDs are those that can use svgalib under Linux, but this is bound to change with GGI...
This is plain FUD.
Linux distributions usually have all the shiny knobs, but more often than not, this means more rope to hang yourself.
BSD distributions usually try to keep configuration to the bare possible minimum, and automate everything.
As a recent personal example, I can remember having trouble with redhat's CD which insisted on me having a swap partition, even though that machine didn't NEED any swap (128 Mb).
One other point where BSD is easier is man pages. At least, you have documentation for all commands and system calls. This is more compact than howto, but this means LESS to read.
Depends on which class of newbie you belong to. Newbies who are not afraid of manual pages may have a simpler time with BSDs... and they can still read the BSD 4.4 manuals, printed version.
Well, Linux drivers are highly specialized, to an absurd point sometimes.
Don't forget that the standard drivers are made by OSS, a company who makes a living selling drivers.
Considering their business, having shitloads of drivers to every card in existence makes lots of sense, but you sometimes get absurd results: this is getting so specialized that my laptop's soundcard isn't recognized by *anything* in Free OSS, except as an older SBPro !
Of course, the commercial version of OSS has truckloads more drivers, and recognizes it.
On the other side, OpenBSD generic Windows Sound System driver is not that sophisticated: it doesn't try to check every functionality of that card, it simply picks it up as yet another cs4231 clone.
Yep, you've got it, I've got a sound card with *BETTER* support under OpenBSD than I ever managed to get under Linux... which is why I'm playing quake under OpenBSD + GGI, not Linux.
But code *is* thrown away and rewritten in the BSD world.
The i386 serial driver has been completely rewritten, the ffs code has little in common with the older ufs, there is all the vfs code, and a major effort is going on to completely re-vamp the memory handling system (as known as uvm).
Apart from that, if you think IPv6 or SMP don't count as `major' changes, well, I don't know what will...
It's done for the boot disk, and it's not odd if you stop to think about it.
If you've got 15 different small programs (even if they use shared libraries) some of the code (like the C run time startup) is going to be duplicated in all the binaries.
If you can merge all the binaries together, and choose which chunk of code to run based on argv[0] then you've just saved yourself a bunch of disk space.
Take a look at the FreeBSD manual pages and look for crunchgen for more information.
And yes, this is only done for the boot disk.
N
It's worth pointing out at this point that BSD has binary compatibility with Linux...so if there's a binary available, chances are, it'll work.
-lx
The code's all there...why don't the Linux folks just take it? When something is developed for BSD, its developed for the entire software community.
-lx
NetBSD: Best for overall hardware support across different architectures.
OpenBSD: Emphasis on security-related issues.
FreeBSD: Best overall x86 support, and has the greatest number of ports.
They all work great - I would suggest starting with Free, and move on from there, especially if you're wanting a workstation OS. I tried Open for a while, and it has some great features, but the security features started to burn me out, and the lack of documentation. Haven't honestly tried Net yet
-lx
I've had bad experiences using x11amp with OSS - that was the first time I've seen BSD crash and burn. We're talking "play, play, play, skip, skip, poof! reboot."
-lx
I was the same way. I tried many of the free OS's and even ran Linux for the longest time. Then I tried FreeBSD just for a change... I've never looked back. We run an all FreeBSD shop now.
... Use the one you like.
But it all stems back to people asking "which OS should I use?"
I found this article in one of the FreeBSD advocacy pages.. It explains how to setup NAT and get natd running. A bit of it is dated, so you might want to check out the second link too, which is an ipfw setup page for FreeBSD.
http://www.computerbits.com/ar chive/9708/lan9708.htm
http://www.metronet.com/~pgilley/fre ebsd/ipfw/
If you read the comments in the Linux kernel source for 2.0, a few of the SCSI drivers have been borrowed from FreeBSD, and for the longest time, the IP stack was from NetBSD.
I haven't dug thru Linux kernel code for a while, so I'm not sure if this is true anymore....
It will have one when the question is not NT or Linux, but a question of GNU vs. OpenBSD vs. FreeBSD vs. NetBSD vs. GNU/Linux vs. [insert future OS here].
--jon. Postel is dead. May we all mourn his, and our, loss.
It's also a fair bit harder for a clueless newbie to set up; I don't recommend that a green user go out and install OpenBSD on their PC.
BSD still has a superior kernel in terms of raw forking speed.
--jon. Postel is dead. May we all mourn his, and our, loss.
Originally there was BSD/386 (which tended to use X386). BSDi sold commercial copies; when Berkeley's advanced computer lab closed down, FreeBSD formed. It forked into NetBSD and then OpenBSD because of the usual internal strife. So now there are really four major BSD distributions--BSDi, FreeBSD, NetBSD, and OpenBSD. I like OpenBSD best myself.
--jon. Postel is dead. May we all mourn his, and our, loss.
That said, BSD is going to be with us for a while longer, if only because of the ease with which GNU/Linux binaries can be run on BSD and the ease with which device support can be migrated over (I won't address licencing issues here).
I haven't seen the Matrix movie, but I did hear on Systalk that FreeBSD was central to the production of said movie.
Cheers,
Joshua "Still running OpenBSD on one PC" Rodd
--jon. Postel is dead. May we all mourn his, and our, loss.
I'm courious, what are the advantages of a BSD over a Linux. I understand that BSD is more mature, but what does this actually translate into, all I have hear was faster forking and a faster IP stack. Also,it seems as if Linux has really started closing the gaps in terms of desktop software (as opposed to server)....Word Perfect, KDE, Gnome, stuff like that. How does BSD stand up in this respect.
what sort of effect do you think it will have on BSD that apple is using its kernel as part of Mac OS X?
it's bound to have interesting effects on the whole BSD culture that a bunch of random mac users are suddenly going to have a BSD flavor installed on their machine. Even if they (the mac users) largely aren't aware of it.
I don't know whether Mac OS X will actually be able to run BSD programs normally. I think the bsd kernel is pretty much left alone. But still, that's a relatively large boost in the user base.
Of course, because of Mach, Mklinux and kernel hosting you'll be able to boot a linux kernel off the same microkernel as mac os x, but that will only be if you specifically go and install it. The BSD pieces will be installed for everyone.
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
The little BSD devil is a very nice illustration, that should prove attractive to us Mac users. Some of the Penguin art I've seen is a little...ehhh... kinda weak. Artie the Gimp rules though, he could easily carry his own comic strip.
The Mac OS brings to the table the Mac system smileys, and the QuickTime Penguin.
Any other OS Character Mascots out there?>
The music is not in the piano -Clement Mok
That *BSD has not had the same success in the comercial market as Linux has had. I would disagree. *BSD has had subsantialy more success, it's called SUN, HP-UX, Digital Unix, BSDI..... Oh, one problem though, those aren't open source operating systems. That is the main reason *BSD hasn't had the same percieved market presence that Linux does. Once a company gets a hold of the *BSD code, they hack it up, and release their crippled version under a proprietary license, so you never hear of *BSD, instead you hear of SUN, HP, BSDI ....
Linux, on the other hand, will not let this happen. It can't happen because of the GPL. the GPL ensures that Linux will always remain open to the users. *BSD's fatal flaw was that companies could take control of it, and lock up the source. The license is the reason I choose Linux over *BSD. I am assured that Linux will always be Linux. My efforts will remain in the public domain. No one will be able to take that away from me, or any other user. I give respect to *BSD, we have all benefited from it! However, I live by the GPL.
-Master Switch out
-Master Switch, one more element in the machine
I switched to OpenBSD about two and a half months ago (from Linux). During these months, the system has hung once, and rebooted "by itself" once. This is very little compared to the weekly or even daily reboots or hangs I have experienced in Linux (2.1.x and 2.2.x).
:). Linux was more or less unusable at loadavg 10-20 or something.
:-)
For example, I accidentally got the loadavg up to over 60, but OpenBSD followed along nicely (sure, there was a 10 second delay between the mouse updates, but so what?
I guess that people who don't do a lot of weird programming that I tend to do, and people who are not permanently connected to the net, don't have a reason to try BSD. But otherwise I really recommend it.
Before I changed to BSD on my main box, I was running BSD on second box (a 486) for half a year. That is really something I recommend (if you have two or more boxes). This is better (in my opinion) than the other common alternatives: 1) not try other OSes out at all, or 2) switch before you know that you'll like it, or 3) some frustration (sometimes) with multiple OSes on the same box.
Anyway, not just the OS should be free, but also the choice! It's good to see that slashdot is not "linuxdot" only
(I haven't had time to try FreeBSD or NetBSD... they are probably good too, as well as Linux)
A. Gavare
natd.
If your only using ppp and are lazy ppp -alais works great too.
I actually make more mistakes in the linux tool, but thats probably because I a more familar with FreeBSD.
A tool for FreeBSD does not exist to my knowledge, but I've often thought about writing one. The only problem is, I'm more proficent at default FreeBSD way.
Maybe I'll add that back to my list of projects.
www.firebelly.n et is running Apache/1.3.6 (Unix) on Linux
Can I get an eye poke?
Dog House Forum
Like the *BSD's, Linux's kernel's get split on ports, until they can be merged with the main kernel tree.
I hate the FreeBSD central team. Back in the 2.2.6-current days when you needed a patch to run Quake 2 under the Linux emulation (dunno if that's still true) I suggested to them that they should just integrate the patch into their OS.
They basically gave me the finger. You might still be able to find the exchange on the freebsd-current archives if thy go back a year or more.
Sorry, but FreeBSD just doesn't have a real open development model from what I can tell.
Glückwünsche, haben Sie Slashdot ermordet, indem Sie zum korporativen Druck beugten und Subskriptionen einlei
Well, gee, they should have told me that then instead of being all rude to me. Ever think of THAT?
No FUD here, just facts. If it was sloppy crap coding then they could have mentioned that as a reason, or told me that it needed more testing before being integrated into the kernel.
Glückwünsche, haben Sie Slashdot ermordet, indem Sie zum korporativen Druck beugten und Subskriptionen einlei
I'm glad to finnaly see some FreeBSD coverage here on sloashdot!
The reason i use FreeBSD as my main os is simple:
it's very straightforward with it's directory structure and kernel configuration. I've never been too hot for sysV based unix (ie- RH linux) but i do support all and any open source/free os.
Besides from my experience FreeBSd has been rock solid... and can emulate linux on the kernel level... so there is really no need to wait for true port to FreeBSD. It gives me the best of both wolrds.
On a side note kudos to whom ever said we shouldn't start an os war bestween free unix. we are all against the same enemy. microsoft.
I think that's an appropriate title for the article. Like typical *BSD fanatics, they are full of themselves and always use inaccurate stereotypes and heresay when talking about Linux.
;-)
/. serves the open source community at large and not just one particular segment thereof. They might be selfish, pompous, stuck-up prats living in a kin-delusional fantasy world, but they're better than Microsoft.
For example, Linux kernel development is not quite as dissimilar as they would have you believe. There _is_ a core group of developers. The difference is, they don't go around sprouting how cool they are because they're core developers of some OS. They just quietly get on with the task at hand. Also, because the group is more informal, others not part of it feel more inclined to contribute as the test is the quality of the code, and not how much you wine & dine the core group.
When talking about flame wars, it's interesting to note that the *BSD crowd are inferring that it never happens to them. Wahahahahahahaha! Lets mention a few famous ones, shall we? NetBSD. OpenBSD - Theo de Raadt in particular. This is how far up themselves FreeBSD folks are, they "banish" people who helped make their OS what it is and refuse to acknowledge their existence like they are some form of lower-intelligence scum.
The whole GNOME vs KDE saga has nothing on the flame fest that causes fractures in the BSD community!
They also say that *BSD users are older and more mature (where did they get those numbers from? Plucked from their tight arseholes again of course like all their other extremely general commentary)
and try to infer that this somehow reflects on they quality of their operating system over that of "late teens and 20s". If you *BSD folks were any further up yourselves, you'd turn inside out!
Age does not necessarily mean maturity, and quality depends a lot on skill. I'm not going to make the bold statement that Linux has more skilled programmers (like the BSD people would say in the reverse) - however applying basic statistics shows that due to the larger group of programmers working on Linux, there is a greater chance that there is more skilled input. Of course, people win lotteries so statistics is wrong sometimes
Finally, there's the age old "BSD outperforms Linux by x times" debate, for which I won't say anything but the fact that the benchmarks speak for themselves. Oh, sure, FreeBSD might have a lower process switching latency on an 16Mhz 386, but who cares (apart from them)? Benchmarks of 15-year-old technology are useless in the real world today.
On a different note, however, I'd like to say that I would enjoy more BSD related stories because
Matt
Really? I actually run MacOS, so who is the person with the post full of inaccurate statements now? If you really think that my analysis of the article is that inaccurate, how come you do not point out any flaws in it? If it is that full, surely you can come up with at least one item that you believe is wrong, and present an argument as to why you believe so backed up with some hard data. All I, and everyone else, can see are two quite general statements with no solid foundations in anything.
I reply to this post because it scores above the threshold - I can only guess that the rest are BSD fanatics screaming blue murder at me for daring to point out flaws in their precious evangelical article. They then wonder why they don't get much press, when they treat critics with an IQ above their uptime, of the little press they do get with such scorn and the typical BSD fanatic's baseless condescending commentary.
If you are not interested in intelligent debate and discussion, then do not bother replying in the first place.
Back to the article, I was interested to read that the film "The Matrix" used BSD systems. It's great to see SFX companies doing wonderful stuff with Unix systems, breaking the traditional NT/Lightwave/3DS mindset that was becoming all too popular earlier this decade.
Matt
Free software will have won when the newbies no longer choose between Windows and Mac, but choose between Linux and FreeBSD.
A Government Is a Body of People, Usually Notably Ungoverned
The article stated that FreeBSD people tend to be in their 30's whereas Linux people tend to be in their teens and 20's.
I'm a 37 year old Linux nerd. (Was a SysV nerd for many years before Linux existed.)
Does this make me immature or progressive?
;-)
That's funny, I have a joystick, quickcam, tvcard, and
a non soundblaster soundcard on my freebsd box.
And a sb64, quickcam, and tv card on my BSDi box.
Granted BSDi hardware support is pretty behind the times,
but saying *BSD* doesn't have good hardware support is nothing but FUD.
- Just because you have to build a kernel doesn't mean it's
not supported.
I know some linux nuts who burn a kernel like it's a daily ritual.
This is my signature. There are many signatures like it but this one is mine..
Linux rocks. I have used it for over four years. But I love FreeBSD, which I started using only a couple of months ago.
Install and setup is not difficult. Adding packages and the Ports system is great and badly needed by Linux.
The user PPP on FreeBSD is not only easy and straightforward it has many capabilities for displaying info on a conection.
ALMOST everything you do on Linux can be done on FreeBSD. Linux does have the most development effort behind it app wise.
So which to use? Try both and see which one "fits" you the best.
I'm a Solaris and AIX sysadmin and have run Linux and Solaris at home as well. FreeBSD is my main OS now. It "fits" me.
Oh and FREE OSes HAVE won. Bill Gates doesn't have to admit defeat for us to claim victory.
"Linux is something for Windows haters, BSD something for Unix lovers" (Heike S., Febr. 98)
At least FreeBSD (I don't have lots of experience with Net or Open) has ports for all of the above. I've tried WordPerfect under FreeBSD's Linux emulation and it works fine - I use LaTeX for most everything so I haven't used it much though. KDE, Gnome and over 2100 other ports are available for install by simply typing "make install" in the port directory. I find it much easier than using RPMs in my experience.
I don't recall him saying that it was faster under FreeBSD although there is some evidence that some software does run faster under emulation than natively on Linux.
You can now use x11amp without OSS on FreeBSD - you need to enable Luigi's sound drivers in the kernel.
There's also a splash screen available from:
splash screens
Free software will have won when the newbies no longer choose between Windows and Mac, but choose between Linux and FreeBSD.
I beg to differ: Free software will have won when the newbies can choose between all of the above, and do choose from among the free.
It seemed mostly like a fluff piece. But I agree with the premise that the *BSD folks haven't done too well in terms of evangelism, and that the rabid Linux advocacy movement is primarily driven by younger folks now.
BSD people rave about the quality and cleanliness of theor chosen flavor. Linux people rave more often about its social aspects.
I wonder just how much of the usage gap is publicity-driven, and just how much of it really is free choice... It's an question for which I can't even begin to posit an answer.
The BSD's used direct 386BSD code and have since integrated later CSRG releases.
Linus used Minix to get the original Linux kernel started but it certainly isn't based on any Minix code.
But come on, what is wrong ?, if you don't correct it how are the truth suppose to get heard of?
I agree with the "Down the toilet" thing because nobody shows any statistics or any benchmarks and yet BSD is said to be better, but I don't beleve it just because someone (especielly the maintainsers of some BSD version) sais so.
Prove me wrong if you can, please do!
De lyckliga slavarna är frihetens bittraste fiender, legalisera!!!
Everybody sais I think this and My Personal Experiense is that but haven't relly anyone got any facts at all?. No benchmarks no Statistics no nothing???. I can't say I agree with any thing you are saying. For example in many post one can read thet Linux people flame BSD because they are inexperienced. What I tend to notice is that its actully the opposite, every BSD article I've read got lost of bad opinions about linux but Never every do I read an article by a linux user that sais so much bad stuff without showing any facts at all.
;)
But hey, maybe linux people are silly fools who just follow the stream because all these post are lame and BSD biased (almoast) and generally looks about the same. So maybe we linux users just follow the BSD streem and continue to say, BSD is better but i use linux cuz its more of a workstation and got a cute penguine instead of that daemon
Is relly BSD so good that everyone that uses it falls in love or are just the average linux user not intressted in BSD because every "fact" about linux and BSD tend to be written by a BSD fan who relly thinks BSD emulates linux faster than linux ever run by itself and goes on like that, maybe saing some other non-existant-fact until he has bored you insane.
/prove me wrong - with facts please!!!
De lyckliga slavarna är frihetens bittraste fiender, legalisera!!!
It sounds fair, still no facts though but if all you are saying are true, none are needed.
;-)
I just hope the next BSD article won't feature the same bad-attitude-against-linux--bsd-people and instead has a more open mind(like your post) so we won't need another one of these discussions
De lyckliga slavarna är frihetens bittraste fiender, legalisera!!!
of course it hasnt been up for years!
-- your knees hurt, don't they?
Yep. natd.
"That's Tron. He fights for the Users."
There's the FreeBSD Handbook, published by Walnut Creek. It's an excellent reference and covers many topics. It's also available at www.freebsd.org/handbook. In addition, there's #freebsd on efnet irc...just be sure that if you ask a question in there, you've first read the handbook section on the subject and any related man pages, or you're likely to be kicked quickly.
"That's Tron. He fights for the Users."
On my desk at home, I have systems running FreeBSD, OpenBSD, and Linux. For quite a while, I was a big fan of Linux, until the day when an employer of mine asked me about FreeBSD. At the time, I knew nothing more than that it existed. So I looked at it. I installed it. I *liked* it. A lot. I loved having unified source distributions for the whole OS (kernel and all standard programs)...made installing patches and upgrading to newer versions much smoother in my opinion. I still like Linux, still think it's a :)
great OS. I'll even concede that for a desktop system it kicks FreeBSD's butt in a lot of ways. But when it comes to making one hell of a server, I'm FreeBSD all the way. Just ask www.yahoo.com and ftp.cdrom.com...they'll tell you why
"That's Tron. He fights for the Users."
I believe the differences can be analogous to the process of evolution. I'll keep this short and simple (Note: distro = distribution):
********
BSD = code forking.
Reason: Company A develops *BSD distro with proprietary feature X. Company B develops *BSD distro with proprietary feature Y. Companies are not obligated to share source code with each other (i.e. no genes are shared).
Result: Companies continue to develop products that evolve separately. Features begin to vary wildly and eventually they become separate species altogether.
********
GPL = no code forks.
Reason: Company A develops Linux distro with feature X. Company B develops Linux distro with feature Y. Under GPL, both companies must freely share code with each other (i.e. genes are shared).
Result: All the good features (survival genes) eventually spread to all distros, and all the bad features (non-survival genes) eventually die out through atrophy. Although there may be minor genetic differences, all distros are of the same species.
********
Thus, it is fairly simple why GPL products (including but not limited to Linux) will always win out eventually over competing non-GPL products.
This is the simplest explanation I can think of as to why *BSD, although a more mature product, is not growing as fast as Linux.
However, to borrow from Star Trek, we should always strive for Infinite Diversity in Infinite Combinations. That is, we *need* a variety of products so that we are always striving to improve through competition.
-- UOZaphod
"The unicode stuff in the latest version is working fabulously well. My russian mafia friends are ecstatic."
Driver support for this type of hardware is likely to lag on FreeBSD... it is intendended primarily as a server OS after all. If you seriously want to use consumer (not server) peripherals like tuners and joysticks, you are probably better off with Linux.
That said, by all means install FreeBSD and tinker with it. It never hurts to broaden your skills. I run a headless FreeBSD box at home and use Linux to telnet and X into it. You could always pick up an old 486 for next to nothing, or go the dual boot route. Good luck.
BTW, I'll let you in on a secret. One reason I run so many operating systems at home, each on their own hardware? Every system has a win95 partition so I can occasionaly reboot 'em all and host Quake/BattleZone frag fests. :) I currently have 5 systems on the house network... I think I've got parts lying around to make a sixth.
Thad
(hoping I can still claim my *cross platform testing lab* as a tax writeoff) >:)
The Bolachek Journals
Well, at least three of the systems were purchased with Win95 already on them, so those are legal. One system only has a DOS partition (it is too slow to play BattleZone anyway). I'm not sure about the last system. It has a legal copy of NT but I honestly can't remember where the Win95 came from. It was probably purchased with Win95 but that may not be the copy that was reinstalled when I repartitioned and put NT on it.
Yes, it is steep, and that is why I am a big advocate of games for Linux (I'm even writing one of my own).
Thad
The Bolachek Journals
And contrary to some comments I've seen saying otherwise, FreeBSD's install is really slick. Redhat and other Linux distros have only recently caught up to where FreeBSD has been for some time in ease of install. Furthermore, the Ports system rocks! Linux needs something like this.
BTW, Jordan Hubbard seems like a rather nice guy. He provided me some very useful feedback on a project I'm working on. Comparing him to Linus, I'd be hard pressed to say which one is more cool. ;-)
Thad
The Bolachek Journals
all i have to say is freebsd rocks and here is one reason why...
:) )
rup xx.xxxxxx.com
xx.xxxxxx.com 8:45pm up 319 days, 23:34, load average: 0.00 0.00 0.00
the box is rock stable, esp for its age and hardware in it, even survived moving buildings ( carried by hand with ups attached
-cry0gen
PS we used to use linux untill we tried fbsd and we have been in love ever since
why dont we go the other way around?
:/
Doesnt matter anyway, it will never happen
-cry0gen
never had a box that wasnt used all the time?
:)
plus ive never had a linux box go past 100 day uptime w/o rebooting itself
NT wouldnt last that long
-cry0gen
this wasnt meant to be a flame war :)
i like linux as a workstation runs pretty bad ass as one, just saying something about freebsd
shesh
cant say anything w/o starting something
cant all unixes get along?
-cry0gen
When has the code ever stopped being free? XFree86 was ready to proceed without The Open Group when they changed the license. The code you and I were using was still free and open.
The GPL will not save you from code forking. EGCS and GCC make a good example. The GPL does not have a clause saying forks are not allowed.
Well, the original UNIX forking is one example, as is X (display postscript, for example). The cost of the damage caused is difficult to calculate, but the forking has caused little but incompatiblity for programmers and problems for system administrators. As you may be tired of the GPL religion, I'm dearly tired of discovering subtle differences in every UNIX, even tho their heritage has so much in common. If the GPL will do through force what enlightened self interest didnt do for compatibility, then I'll take that.
The GPL does, indeed, not forbid forking of the code, but it does take away a lot of the incentives to fork (IE, making a proprietary version, including a new feature and competing on 'Buy our stuff. We have FOO. They dont.'). Since that leaves only serious technical disagreements as a reasons to fork, it happens more rarely (and never in a proprietary fashion, allowing backports and discouraging incompatibility).
Embrace-and-extend becomes a lot easier when you can just take the code supporting the protocol and add your own stuff, rather than having to reimplement it from scratch.
As regards the conceding of the desktop, there is was a recent interview with Jordan Hubbard at http://www. internetworld.com/print/current/webdev/19990412-fr eebsd.html. I also remember reading similar sentiments on usenet several times. Mr Hubbard and some other fairly outspoken BSD advocates may not represent the BSD community, but in such a case the other developers should speak out too, because this idea appears to be fairly prevalent. I most certainly hope the BSD crowd would join in in the quest for the desktop, because together we'd be stronger.
And, as far as the tightness of control, since it appears people have differing opinions on this, I'd guess it's just a matter of how involved you are and what you mean with 'control' :). 'Linux' probably has a lot more than 150 people 'in control' because the separate packages tend to be maintained by separate and unorganized people. The different packages in themselves have a lot fewer of course (some control the kernel, some glibc, some gcc, etc). And in the end, the distribution creators have control over what happens in the integrated end result. The same is, of course, true of BSD in some fashion, because the BSD distributions include external software too. I guess it's just a matter of opinion in the end :).
Oh, and I forgot to mention, good luck with your free FreeBSD installation. It's a great choice of operating system. I'd definitely try it out at home myself if I hadn't had the chance already at work :).
Well, on the desk next to me...
9:08am up 485 days, 10:17, 12 users, load average: 0.13, 0.05, 0.01
That's linux 2.0.30...
Well, there are no serious technical reasons. Both will experience minor temporary advantages in one direction in some parts of the system at different times. But due to the high exchange between systems I dont think any such difference will last, and most claims in either direction are usually just advocacy.
I've had both keel over at high loads. But I expect any system running at 60 loadavg to keel over eventually, and in any case you have a serious sizing problem (not to mention response time problem) if you have a sustained loadavg around 60.
So in the end it comes down to a few practical, political and personal preference issues.
SysV. I do not maintain just BSD based systems. Since I have to maintain AIX, Solaris and HP-UX which are all more or less SysV style, any time I get to a BSD based system it's an annoyance. This is an annoyance with Linux too, at times, but at least it's a little more SysVish (we can argue the merits about that...).
GPL. Some BSD supporters argue about the extra freedom of the BSD license, but in my opinion, if BSD should make major inroads and raise corporate interest then we'd just get another... SunOS, HP-UX 9, etc. The BSD license is more free than the GPL, but the price of that freedom is proprietarization, code forking and yet another round of incompatible embrace-and-extend corporate wrangling around. No Thank You. We Have Done That Already. BSDI has a tendency to play nice, but the others dont.
A realistic view on the market. BSD appears to be willing to concede the desktop to Windows, and be content with being a very good server platform. While that may be a possibly realistic view, it isnt in my opinion an acceptable one. Because Microsoft will not tolerate either BSD or Linux as a server platform, and they'll do everything they can to make sure that the Windows clients wont work with anything but NT, or that there are major proprietary advantages of using NT. Giving up the client market means, IMO, giving up any chance at existence at all. It wont matter how good you are if Microsoft has total control of the clients. And most of the major advances in Linux have come as far as they have because the people behind them were not realistic.
And finally, for various reasons,I actually prefer the more componentized and anarchistic development of parts of the Linux systems. I'm happy to leave the integration to the distribution makers, but I like the lack of central control.
Yep. And as a system administrator supporting HP-UX, Solaris, AIX, IRIX, BSDI, SCO and some things I forget, and a sometimes cross-platform unix programmer in my free time, I'm finding at times that I'd be happy to sacrifice some of the 'better' for more of the 'same'. It usually just results in having to use 'the worst common denominator' and/or writing dirty kludges anyway.
"...for "gotta have it now" nat on a dynamic IP ppp connection."
The kernel mode natd/pppd combination also does wonderfully well with danymic addresses.
Some Linux users hate that Penguin, too. I figured a much better animal to associate with the OS (based upon it's lineage) would be the platypus.
yes, FreeBSD has IP Masq -- it's called natd. They also have a packet filtering interface called ipfw. The stuff in Linux 2.0.* is quite similar to this with (I think) ipfwadmim.
I used to run FreeBSD at home with a private network using NAT to get out. I liked it a lot and found it very stable and easy to maintain. I switched to Linux so I could run FreeSWAN.
>Hear hear. The last thing we need is this kind of argument.
;o) but most often when I talk to Linux people about it, they flame me groundlessly. Kinda bothersome, especially considering the fact that very few of these people that I've come across (and that's lots *sigh*) have ever used FreeBSD, and certainly not for any length of time or to any depth... This isn't an invitation for lots of people to go "I've used both!! And I still like Linux!!" btw, I'm just quoting my experience in the hope that some fellow sufferers will recognise it. ;o)
True. I'd consider myself a FreeBSD evangelist
Cian
On the basis of a (subjective) recent trawl around the net for info the capabilities of both Linux and FreeBSD over the past few months I would suggest that Linux is walking away with desktop stuff but FreeBSD is still the OS of choice for server end stuff (where hardware is usually more carefully spec'ed and the device support issue disappears). I've used it, or seen it run, for mail, dns, web and ftp servers in a number of places with minimal effort but it has been a bitch to get going as a desktop. Usually a FreeBSD server means a long time UNIX hacker has put it in as a cheap fix and it has stayed on a "ain't broke, don't fix it" basis.
I work for A.N ISP: we were recently used in some Intel ad campaign because at the time we were pretty much using Intel boxes everywhere ( we hid the Suns while they were round :)). Of course we weren't running NT.. Actually way back at the start the 2 machines were running NT, but that didn't last. The company's progression has been Sunos4->NetBSD&BSDi->FreeBSD. Why don't we have any Linux boxes? back when the company was first expanding, Linux simply wasn't good enough. BSD 4.4 based systems are far more mature than Linux: we simply could not have run the systems we did on the Linux of the time. I don't think we've seen parity of performance/maturity until the 2.2 kernels arrived, and now we've got so many BSD systems ( and an increasing number of Solaris systems ) we simply don't want to learn a new OS. I'd imagine many BSD shops wanting to move OS might feel the same.
We have had our share of OS bigotry too.. one character of the UK ISP scene is quite vocal in his opposition to Linux, although he left some time ago. I find this sort of prejudice intolerable, especially in a commercial environment: pick the tool that is going to suit the job best for you, be it Solaris, NT, Linux, *BSD, AIX, whatever. Each of them has their strengths and weaknesses ( in our case, Linux is an outsider, and so gets dropped. If the company had started two years later, maybe BSD systems would be the outsiders ).
GNU tools have made the user interface on free unices pretty much homogenous, so the argument generally comes down to the kernel. Personally I find FreeBSD has a coherence lacking in Linux, which for single-purpose systems is a major plus: for workstation or general purpose use it's a limitation, as FreeBSD lacks most of Linux's bells and whistles. The BSD kernel is also very well-documented, another major plus when you're hacking it.. this isn't to say the Linux kernel's not hackable, it just takes more time if you're not familiar with it.
Performance wise, if you ignore the strange anomaly that is FreeBSD 3.1 ( how many bugs? ) FreeBSD is absolutely solid. Web servers, mail servers, routers, proxy servers, whatever, all have continually proved themselves under extreme loads, which makes a change from the old days of 3am taxi rides to push reset buttons.. This isn't to say Linux can't do the same, however I don't think it'd have coped quite as well a year ago. Today I think the preference comes down to personal choice as much as anything else ( if you factor out support ). We choose FreeBSD because it fits the way we work.. but we can choose.
The future is something else: we've reached scalability limits for Intel-based systems for quite a few services.. our core network uses Cisco routers, and we're going to Solaris/SPARC for a lot of other stuff. Bit of a shame, and a bit of a pain when you're trying to do something different, but it gets us nice big blue Suns to play with at least:)
Oh.. and I'll add that with one exception the technical staff are all under 30.. :) and none of us have beards.