“The BSD phenomenon is quite remarkable: a community of incredibly talented and experienced developers, administrators, and users, joined by diverse technical interests and a common desire to build and use the best operating systems in the world.” Read the rest of the editorial here.
Daemon’s Advocate: How to be an Advocate for BSD
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
that supports most CURRENT hardware.. I am looking to get into the BSD channel. are the guis any good, or are they very minimal.. I don’t mind, and how much like SkyOS is the average BSD build.. (if at all) I haven’t tried either, but am looking.. any suggestions?
“FreeBSD is about operating systems, not about advocacy–advocacy is for Linux users. Which, if you think about it, is a somewhat refreshing perspective. BSD is, after all, about a commitment to technical excellence, and not about making an operating system platform into a religion. ”
this is just trolling. every product needs advocacy to promote it.
if you talking about freebsd talk about its merits instead of finding fault with linux.
the linux advocacy howto makes very good points about this. replace linux with freebsd and all of it would still apply.
I liked the article except how he perceives linux users. What a superior attitude. I like BSD, but I must say I find the FBSD install wonderfully easy and the system fairly easy to use as well. I don’t see linux as an “easy” first step or a fad. And I don’t see BSD as the ultimate end all of a kernel.
Other than that, I think he’s right. If I got his intent right it is to be an advocate without being in people’s face. Just be a nice person and point out your kernel of choice as a solution to problems people discuss. Ya don’t need to hand out flyers on campus!
I didn’t find the tone of the article condescending or arrogant at all. The author was just trying to convey the different ideals surrounding each OS, and wasn’t even trying to say that one was better than the other (just that they are different).
I liked the article, although it didn’t really say much other than giving some ideas on how to promote *BSDs without being perceived as a zealot.
“I liked the article, although it didn’t really say much other than giving some ideas on how to promote *BSDs without being perceived as a zealot.”
Doh, I just reread the title of the article. Nevermind 🙂
“I didn’t find the tone of the article condescending or arrogant at all”
check this
“FreeBSD is about operating systems, not about advocacy–advocacy is for Linux users. Which, if you think about it, is a somewhat refreshing perspective. BSD is, after all, about a commitment to technical excellence, and not about making an operating system platform into a religion.”
what does this suggest?
linux is a religion?
freebsd doesnt need advocacy?
linux isnt committed towards technical excellence?
“linux is a religion?
freebsd doesnt need advocacy?
linux isnt committed towards technical excellence?”
That’s pretty much the impression I got over past the 6 years. Though I think the fanaticism has become less rabid, or perhaps it’s just that I don’t read comp.os.linux.advocacy anymore.
“That’s pretty much the impression I got over past the 6 years. Though I think the fanaticism has become less rabid, or perhaps it’s just that I don’t read comp.os.linux.advocacy anymore.”
i can say the same about *bsd too. that obsolutely means nothing
Linux combined with gnu tools and stuff is clearly an operating system. calling it religion is just trolling
“freebsd doesnt need advocacy?
”
if operating systems can just excel by being technically superior there would be no need for advocacy. that fortunately or unfortunately is a myth
if freebsd doesnt need advocacy what the hell is the article trying to to do.
the article mentions linux more times than freebsd. what kind of advocacy is that eh?
linux is a religion?
No, GNU is a license advocating a phylosophy of fairness that could be called a religion for a few zealots like me.
Obviously the world will still exist if people are greedy, but some of us believe it would be a better place if we were not. Some zealots, like me, also believe it would be a better place if we did not have any form of currency and promoted science as our primary religion. But what do I know. Absolutely nothing! I’m just a computer scientist.
In this age of rampant fundamentalist religions believing in the unprovable existence of a God, science and technology can also be called a religion, like aethism. I group BSD in with science in the sense that it is focused purely on the techinical rather than the economic and philisophical consequences of its existence. GNU, on the other hand, has more concerns about the social consequences of IP, NDAs and other tools capitalists choose to inflict on their employees.
I am also deeply worried about this religious war that exists between capitalists, fundamentalist religious fanatics and scientists. I fear their clashes could one day cause serious problems for our precious economy. But my perspective is different than the average, so I must be wrong. Right?
…that supports most CURRENT hardware.,
I’m not quite sure what you mean by that question, especially since you segue directly into asking about GUIs. I suppose you are asking about support for 3D graphics hardware, etc…?
FreeBSD might not support the complete range of hardware that the various distributions of Linux support, but it supports virtually all modern *mainstream* hardware, with a few exceptions. In some areas, FreeBSD has actually been ahead of Linux for hardware support. However, since the thrust of FreeBSD has been primarily as a server OS, rather than as a desktop, the hardware focus is usually on such areas as SCSI/IDE/SATA, RAID, high-end network devices, etc…
Really, FreeBSD has access to pretty much the same GUIs as Linux, since it uses the X window system. When it comes to graphical applications, hardware-accelerated graphics, etc…, that falls under the capabilities of the X servers themselves, which are cross-platform for Unix systems.
I have no idea how or why someone would compare BSD to SkyOS, but maybe someone who has tried both could comment.
“the article mentions linux more times than freebsd”
The article is about BSD not just FreeBSD. BSD, *BSD, or FreeBSD occur 41 times v. 8 for Linux.
Obviously both BSD and Linux need advocacy, that is, active support from their users. Advocacy doesn’t mean bashing other operating systems, it means actively pointing out the merits of the OS of your choice while respecting the different choices other people have made.
Saying something like “we let the product speak for itself” is just plain stupid. Every product, be it commercial or non-commercial, needs some advertising and the best advertisement any product can get are happy users, because they’ll be likely to tell others too how pleased they are for discovering this excellent product. Grumpy, ill-tempered users are bad advertisement for any product.
I am talking about a Dual Opteron 250 setup, on the IWill DK8X, DVR-A08 pioneer. ATI 9550 PRO.. and I think an LSI 4 channel PCIX SCSI U320 card..
how about that..
I would like to avoid dual booting if possible.. just a simple Router/file server, webserver, etc.. I dunno right now..
linux is a religion?
I’ve seen a couple cars around the Atlanta area with that fish religious symbol with the word Linux in the middle.
I’ve seen geeks at the local CompUSA handing out Knoppix CDs like the religious hand out their pamphlets.
I know Linux geeks that can quote the GPL like scripture, telling you the paragraph that states the answer any questions you may have. Much like a religious person would point you to John 3:15 to spread the word.
Have never seen any of this in relation to any other OS. May have seen an apple decal on a Mac fans car, but never a religious symbol.
“I’ve seen a couple cars around the Atlanta area with that fish religious symbol with the word Linux in the middle.”
That’s called being a geek.
“I’ve seen geeks at the local CompUSA handing out Knoppix CDs like the religious hand out their pamphlets.”
That’s called socializing, people tend to do this. You see, they find a hobby or activity. Then, do to a social nature, they seek to find others to bring into the activity.
“I know Linux geeks that can quote the GPL like scripture, telling you the paragraph that states the answer any questions you may have. Much like a religious person would point you to John 3:15 to spread the word.”
They have issues and far too much free time. I wonder if the author even has it memorized.
>”I didn’t find the tone of the article condescending or >arrogant at all”
>
>check this
>”FreeBSD is about operating systems, not about >advocacy–advocacy is for Linux users. Which, if you think >about it, is a somewhat refreshing perspective. BSD is, after >all, about a commitment to technical excellence, and not >about making an operating system platform into a religion.”
s
it suggest that linux users don’t accept critics, and you prove so…
>It would be unfortunate for FreeBSD, because at the moment, >Linux is basically undisputidly technically superior to >FreeBSD.
That’s a troll…
Or you don’t know anything on OS, so get a OS book and talk after you have read it.
t would be unfortunate for FreeBSD, because at the moment, Linux is basically undisputidly technically superior to FreeBSD.
Right-o. Now that’s how to advocate! Bravo!
I’m going to go wipe all my boxes and put Mandrake on them now.
Some readers appear to have missed that that was a quite
from someone at CMU, and not a direct reflection of my
own opinion. Given that the point of the article is that
advocacy should be important to BSD, if approached
correctly, this perhaps should be clear… 🙂
>>It would be unfortunate for FreeBSD, because at the moment,
>>Linux is basically undisputidly technically superior to
>>FreeBSD.
>
>Or you don’t know anything on OS, so get a OS book and talk
>after you have read it.
>
I know something about operating systems. Where do you want to start?
Portability? The Linux kernel supports a lot more CPU architectures than even the NetBSD kernel. It is ported to architectures without memory management units, which even NetBSD can’t do. FreeBSD supports a handlful.
Memory management? Linux runs on an SGI system (single system, not cluster) at NASA with 4-8TB of memory in a NUMA configuration with 256 distinct memory nodes, it can run x86 “highmem” systems of up to 32GB memory reasonably well, and there is a patch for the 2.6 kernel which allows removal of unneeded features that allows it to boot in 2MB of memory with a shell. FreeBSD recently got support for x86 PAE (highmem), but it is not clear how effective it is or how much memory it can handle.
Process management? Linux has a runqueue-per-CPU scheduler which supports SMT, SMP, and NUMA, it outperforms FreeBSD in fundamental operations like context switching, system call overhead, fork/exit paths, etc.
Scalability? The aforementioned NASA system has 512 CPUs, and it is using a (modified) 2.4 kernel. I think they are hoping to get 2.6 to 1024 CPUs without too much modification. The scalability oriented FreeBSD 5 development branch I would estimate would run into trouble at 8 or even 4 CPUs.
Filesystems? Linux has a host of well tested and stable journalling filesystems. FreeBSD has softupdates with their filesystem, but it is very complex, it can’t offer as much protection as “data journalling”, and recovery is still slower than a journalling filesystem, even with background fsck.
Disk IO system? Linux is run on systems with THOUSANDS of disks with individual disk device nodes.
Networking? Linux can process more packets per second from a single interface, while also being more parallel. As for higher level protocols like TCP, this is an area I’m not so sure about. Both FreeBSD and Linux share common things developed by researchers… congestion control algorithms for example. Linux does support something called SACK, which is essential for high performance long distance links, while FreeBSD does not.
Device drivers? Linux basically supports quite a bit more hardware than FreeBSD… I perhaps not qualify this as an area of *technical* superiority, but it is important.
Tell me, what is one area where FreeBSD is technically superior to Linux?
This article is nothing more: Linux is bad, Linux is an religion, Linux community sucks. Too mutch Linux in *BSD article, don’t you think? And if *BSD don’t care about adv. then why author brother to write it?
>The Linux kernel supports a lot more CPU architectures than even the NetBSD kernel
You discredit yourself, and i haven’t to read after that, the beginning is already to much, bye!
ok…so A. I am a FreeBSD user..but I like Linux and still use it….and I DO NOT like this article..but I have some questions about your post
“Process management? Linux has a runqueue-per-CPU scheduler which supports SMT, SMP, and NUMA, it outperforms FreeBSD in fundamental operations like context switching, system call overhead, fork/exit paths, etc.”
The ULE and Linux scheduler are very similar…where did you get this information?
“Scalability? The aforementioned NASA system has 512 CPUs, and it is using a (modified) 2.4 kernel. I think they are hoping to get 2.6 to 1024 CPUs without too much modification. The scalability oriented FreeBSD 5 development branch I would estimate would run into trouble at 8 or even 4 CPUs. ”
More power to linux…but again..where did you get this info about FBSD crapping out after 4/8 CPUs?
“Filesystems? Linux has a host of well tested and stable journalling filesystems. FreeBSD has softupdates with their filesystem, but it is very complex, it can’t offer as much protection as “data journalling”, and recovery is still slower than a journalling filesystem, even with background fsck. ”
mmm…debatable..highly debatable..but there is really no problem here..both OS’s have a very capable way of handling data and you could easily make the claim that SoftUpdates is the better technology…again where did you get the info about journaling protecting data better?
“Disk IO system? Linux is run on systems with THOUSANDS of disks with individual disk device nodes. ”
…….and?
”
Networking? Linux can process more packets per second from a single interface, while also being more parallel. As for higher level protocols like TCP, this is an area I’m not so sure about. Both FreeBSD and Linux share common things developed by researchers… congestion control algorithms for example. Linux does support something called SACK, which is essential for high performance long distance links, while FreeBSD does not. ”
…well..I have always been under the impression that FBSD’s networking was quite a bit faster then Linux…although I think they closed that gap in 2.6..but I could be wrong..again…where is this data
>>The Linux kernel supports a lot more CPU architectures than
>>even the NetBSD kernel
>
>You discredit yourself, and i haven’t to read after that,
>the beginning is already to much, bye!
I discredit myself??
Let’s see, NetBSD runs on 15.
alpha, arm, hppa, i386, m68010, m68k, mips, ns32k, powerpc, sh3, sh5, sparc, sparc64, vax, x86_64
Linux runs on 18.
alpha, arm, cris, h8300, hppa, i386, ia64, m68k, mips, powerpc, ppc64, s390, sh3, sh4, sparc, sparc64, v850, x86_64
NetBSD counts its ports by “platform” which is a lot different to Linux, which counts it by CPU architecture. This is why you hear of NetBSD running on 40 platforms.
The bulk of the code is really in the CPU ISA. The platform consists of a few specific things you need to do at boot, and for other chipset access, maybe a specific driver or two.
It seems YOU “don’t know anything on [sic] OS, so get a [sic] OS book and talk after you have read it.”
You obviously couldn’t even attempt to address any of my other points. Too bad.
“Process management? Linux has a runqueue-per-CPU scheduler which supports SMT, SMP, and NUMA, it outperforms FreeBSD in fundamental operations like context switching, system call overhead, fork/exit paths, etc.”
The ULE and Linux scheduler are very similar…where did you get this information?
They are similar yes. As far as I know, ULE lacks NUMA awareness, but I think it can handle SMT alright. ULE seems to be not very stable at the moment though.
As for context switching, system call overhead, fork/exit, etc… Linux has often been ahead of FreeBSD in these areas for a long time and still is AFAIK. I don’t actually have a running BSD system here, but we could run lmbench and trade results
More power to linux…but again..where did you get this info about FBSD crapping out after 4/8 CPUs?
Just various places from the net I guess. Not first hand information… on loads where one never enters the kernel, FreeBSD of course would scale to more than 8 CPUs probably… But that isn’t interesting, and it isn’t *FreeBSD* that is doing the scaling
But I’m not sure where they’re up to with their development branch. The 4/8 CPUs was an estimate. The stable branch has the same sort of scalability as Linux 2.2, which is to say, very basic: 4 CPUs is often a problem there.
mmm…debatable..highly debatable..but there is really no problem here..both OS’s have a very capable way of handling data and you could easily make the claim that SoftUpdates is the better technology…again where did you get the info about journaling protecting data better?
Well metadata journalling apparently provides the same protection as softupdates. That is, your filesystem doesn’t go into an incorrect state. Data can still be lost though because a write() can return without the data being on disk.
In practice this doesn’t really matter, because anything serious that needs to guarantee integrity uses something like fsync() to ensure the data is on disk.
“Disk IO system? Linux is run on systems with THOUSANDS of disks with individual disk device nodes. ”
…….and?
Heh. I was hoping someone else could fill me in. As far as I know, FreeBSD’s IO system just wouldn’t be able to keep that many disks busy. I guess in part due to SMP scalability and interrupt reentrancy.
Also, I don’t know if FreeBSD’s block layer would even cope with that many device nodes? Can it?
…well..I have always been under the impression that FBSD’s networking was quite a bit faster then Linux…although I think they closed that gap in 2.6..but I could be wrong..again…where is this data
Common misconception. No doubt there are workloads where you can show either one to have an edge. In the low level packet handling, I think Linux is ahead.
http://lists.freebsd.org/pipermail/cvs-src/2003-April/002501.html
This post is a bit old, but it is comparing 2.4… 2.6 has further improved on 2.4… and FreeBSD 5 is still worse than 4
As far as higher level protocol handling goes – as I said I’m not sure, but Linux has SACK for example, so it is definitely untrue to say FreeBSD is faster/better/whatever than Linux at networking.
Note, I didn’t say Linux was better than FreeBSD at networking either.
“They are similar yes. As far as I know, ULE lacks NUMA awareness, but I think it can handle SMT alright. ULE seems to be not very stable at the moment though. ”
ya, I think it lacks NUMA as well..but it has been pretty stable for awhile and is quite stable now.
“As for context switching, system call overhead, fork/exit, etc… Linux has often been ahead of FreeBSD in these areas for a long time and still is AFAIK. I don’t actually have a running BSD system here, but we could run lmbench and trade results
”
lol…well..I would be willing to do that..I am actually interested in the results…I have a linux install on this comp as well..would be interesting to run it on both.
“Well metadata journalling apparently provides the same protection as softupdates. That is, your filesystem doesn’t go into an incorrect state. Data can still be lost though because a write() can return without the data being on disk.”
True..but a rarity…there are very few crashes where a sync does not get called, but it could happen.
thats all I really have to comment on at the moment since I am tired as hell
….I added my email address to this post…(think it appears as making my name a link)..if you want to cooperate on the benchmarking I would be interested
_exit(0);
>NetBSD counts its ports by “platform” which is a lot different to Linux, which counts it by CPU architecture. This is why you hear of NetBSD running on 40 platforms.
Run is different to be stable…
Ok about your other point:
TCP/IP piles on *BSD are the best, their is not to prove it,
BSD have more than sixteen years of experience.
Firewall of openBSD is a stateful firewall, he can rewrite bad TCP sequence if needed
Where is the jails function in linux, humm, i don’t see it?
Filesystem of FreeBSD is almost good that any FS of Linux.
Some linux programs emulated by freeBSD run faster than on linux.
OpenBSD is the Safest OS in the world, ->Blowfish 448 bits,
-> all code is audited
About stability, just see the longest uptime on netcraft, what?! where is linux?
DiskIO, i don’t know why *BSD will not run “THOUSANDS OF DISKS”
SACK is supported in *BSD.
Device driver -> support more device don’t say that it support well. It is better to support less devices but better.
i think you need to differences performances and stability, because if linux “as you say” is fastest than *BSD, if it crash before BSD, where is the advantage?
BSD are the reference in stability.
OK, since this is turning into a flame war, I’ll try to be calm.
Can someone suggest (in a nice clear, and intelligent manner) why I should convert from Debian to *BSD?
I run workstations, servers, and basic to medium firewall/routers. The servers are all Debian/Stable, and the workstations are combinations of stable+backports, or unstable, or unstable+experimental. All machines are all x86, and only a couple are SMP. Most are Intel or VIA chipsets, with the occasionaly donated SiS (ew) based one. Graphics is nVidia, or non-existant (servers). Disks are all IDE, with a couple of 3ware raid cards. Looking at SCSI…
Common apps/tasks include GNOME, KDE, OpenOffice.Org, remote X, apache, exim, MySQL, PostGRESQL, bind, and a lot of ssh’ing. Not much in the way of custom binary apps.
Now – what is the advantage in either moving to *BSD, or making my next box on BSD?
OK.. where do I start?
TCP/IP piles on *BSD are the best, their is not to prove it,
BSD have more than sixteen years of experience.
Touche. They’re called stacks, by the way.
Firewall of openBSD is a stateful firewall, he can rewrite bad TCP sequence if needed
Linux has stateful firewalling too.
Where is the jails function in linux, humm, i don’t see it?
Usermode Linux provides equivalent functionality.
Filesystem of FreeBSD is almost good that any FS of Linux.
What do you mean by good?
Some linux programs emulated by freeBSD run faster than on linux.
Which ones?
OpenBSD is the Safest OS in the world, ->Blowfish 448 bits,
-> all code is audited
Are you trying to tell me that Linux doesn’t have blowfish encryption, or something?
Also, “all code is audited” is just OpenBSD marketing. What does it mean? Audited by whom? To what standard?
All Linux code is audited too, for some definitions of audited.
About stability, just see the longest uptime on netcraft, what?! where is linux?
Linux isn’t on there because their uptime counter wraps at 500 or so days. Uptime isn’t really a measure of stability though. A big system nowadays could do more work in a week than one of those has done in its entire lifetime.
DiskIO, i don’t know why *BSD will not run “THOUSANDS OF DISKS”
Well I guess that is because you don’t know *BSD’s device and block layers. What I want is an answer from someone who does.
SACK is supported in *BSD.w
OpenBSD supports SACK. FreeBSD doesn’t. I’m pretty sure NetBSD doesn’t either. So no, SACK isn’t supported in *BSD.
Device driver -> support more device don’t say that it support well. It is better to support less devices but better.
And I suppose you’re going to tell me that FreeBSD supports devices better somehow, right?
i think you need to differences performances and stability, because if linux “as you say” is fastest than *BSD, if it crash before BSD, where is the advantage?
BSD are the reference in stability.
They aren’t *my* reference in stability. FreeBSD 4.9 introduced a bad SMP stability problem, softupdates crash with large files, detaching a USB mass storage device freezes the system, 4.10 has introduced MySQL problems… these are a couple of things from the freebsd-stable mailing list. Jeez, it isn’t like FreeBSDs are immune to bugs now.
Linux isn’t either. I didn’t claim Linux was more stable than FreeBSD or any other BSD because I don’t have the evidence. Something you don’t have either.
Why is this turning into a Linux vs. BSD war? Look at the benchmarks and get it over with.
FreeBSD is more scalable and reliable than any linux machine. It runs applications way faster. Have any of you bashers even run FreeBSD? It’s a solid platform. Remarkable at much work they do considering they are smaller than the linux people–but much more mature as BSD has been around since the 70s. linux is the ‘new’ ‘hot thing’
that’s all i gotta say about that.
“This article is nothing more: Linux is bad, Linux is an religion, Linux community sucks. Too mutch Linux in *BSD article, don’t you think? And if *BSD dont care about adv. then why author brother to write it?”
This is unbelievable. This just proves the point of linux & religion, even though this wasn’t exactly his point. The article is about FreeBSD advocacy, not about linux bashing. But of course, I expected _some_ linux users (and yes, I am one too) to read it, pick out one sentence that looks very bad for linux users, quote it without saying anything its context (that Watson speaks about his experience in 1995! -“I recall first arriving at Carnegie Mellon University in mid-1995…”), so you can switch on the indignant mode and shout: our community is attacked! Look at all those BSD folks, they talk down on us!
Past tense folks! Of course, you can just ignore the fact that a paragraph begins with “There was a time when…” – but why? It seems some people just can’t leave without feeling ‘attacked’ by these bad bad daemon folks, without the urge to defend their rel… I mean favourite OS, even if it wasn’t attacked.
This is sad.
>OpenBSD supports SACK. FreeBSD doesn’t. I’m pretty sure NetBSD doesn’t either. So no, SACK isn’t supported in *BSD.
So why there is bench on the net about FreeBSD and NetBSD with Stack support ?
>Uptime isn’t really a measure of stability though
yes it is
>Usermode Linux provides equivalent functionality.
Similary but not equivalent
>They aren’t *my* reference in stability
there is not only SMP in life …
OpenBSD is considered the safest OS by professionnal
Only one remote hole in the default install, in more than 8 years!
what about linux ?
A lot of security hole that are dicovered in linux are already resolv in OpenBSD…
Where come from OpenSSH you use in linux?
>Are you trying to tell me that Linux doesn’t have blowfish encryption, or something?
there is a difference between having it and having it by deault.
“Now – what is the advantage in either moving to *BSD, or making my next box on BSD?”
To satisfy your curiosity? If you are 100% satisfied with your current setting, and you are not curious about FreeBSD, I don’t see how or why would you try it. I like FreeBSD more than linux. My personal reasons are:
– It is easy: it is a great learning platform, because it is consistent and clean (see man hier) – no operarc or mplayer.conf in /etc/ like in gentoo – and documentation is superb (even man pages are better layed out than gnu man pages).
– It is very stable – comparable to debian-stable (I used deb for a while too) with similar package management. Basically you can choose to use packages only, and pkg_* tools work exactly like apt-get. pkg_add -r kdebase = apt-get install kdebase. And there is ports of course, with up to date packages (it is very current, similar to portage, but it different in a few respects, for instance, it has real reverse dependency checking, a web interface via freshports.org, etc.). You can check how up-to-date ports is on http://www.freshports.org
– well, this is related to my first point, but FreeBSD configuration is a lot more easier than linux was for me. A case in point is ipfw (FreeBSD’s packet filter) almost english syntax), the simple and clean text config files like /etc/rc.conf (and to help you: /usr/share/examples/whatever). For instance, I remember Debian having a separate config files for eth devices (was it in /etc/networks?)… mandrake had it too in a different place (/etc/sys/whatever?), while in gentoo, there is yet another way to configure your eth. In BSD, almost everything is configured via rc.conf, and your eth device is simply configured with one line: ifconfig_rl0=”inet 172.17.141.160 netmask 255.255.252.0″ This example reflects FreeBSD philosophy I think… to keep things simple and easy for the user.
Well, if you liked what I have written above, give it a try. Our priorities might differ of course. I don’t have any formal computer education, I began using linux 3 years ago out of curiousity and frustration with windows. I kept switching from distro to distro (distrowatch was my most frequently visited page) – until I tried FreeBSD: I felt at home for the first time. And being a *nix noob and all, I have learned more (and more easily) using FreeBSD for half a year than with linux in two years.
>OpenBSD supports SACK. FreeBSD doesn’t. I’m pretty sure NetBSD doesn’t either. So no, SACK isn’t supported in *BSD.
So why there is bench on the net about FreeBSD and NetBSD with Stack support ?
That is SACK support. Please refrain from commenting if you don’t know what you are talking about.
>Uptime isn’t really a measure of stability though
yes it is
You like an operating system with a nice, stable idle loop then, do you? Well, now I know how you define stable that explains a few things.
>Usermode Linux provides equivalent functionality.
Similary but not equivalent
Yes equivalent. It provides a way to partition the system.
>They aren’t *my* reference in stability
there is not only SMP in life …
OpenBSD is considered the safest OS by professionnal
Only one remote hole in the default install, in more than 8 years!
what about linux ?
A lot of security hole that are dicovered in linux are already resolv in OpenBSD…
Where come from OpenSSH you use in linux?
See, I don’t buy into the OpenBSD marketing machine. They have funny ways of accounting these things. I don’t doubt that it is very secure by default. Like stability, I don’t have good evidence or even know how to measure security, so I don’t make any unsubstantiated claims about it. Unlike you.
>Are you trying to tell me that Linux doesn’t have blowfish encryption, or something?
there is a difference between having it and having it by deault.
Err yes there is. I don’t know what you are trying to say. Is it simply that you must have the last word about everything? Or do you think you have formulated a really succinct debate winner there?
NetBSD, DragonflyBSD & FreeBSD are more scalable and reliable according to benchmarks and statistics from reliable sources like netcraft. so ..yeah.. .. they just have less rpms it seems.. and less corporate backing in the beginning…
Cheers for that.
I’m never 100% satisfied, there’s always something that I can tweak or fix…
It sounds like the principle thing that FreeBSD has over most Linux distros is that things are designed intelligently, rather than slapped in at random. By the sounds of it, Debian has caught up to FreeBSD in terms of keeping things clean, and apt-get is very similar to ports in functionality (the obvious difference of source versus binary not withstanding) and usability.
I guess if I started with FreeBSD a few years ago, rather than Debian, I’d still be using it. If I had a choice… given that my first linux experiance was Debian 2.2 with a 2.2 (or was it 2.0?) kernel, FreeBSD seems more likely.
Anyway, thanks for the non-zealoterous write up.
Saying NetBSD is less portable than Linux is just laughable. All NetBSD kernels are compiled from the single source tree, so supposedly machine-independnet Kernel codes are identical across architectures. Dare you say the same thing about Linux. Oh, don’t forget NetBSD’s super-excellent cross compile support.
On device drivers, NetBSD device driver model is simply much better than Linux. Don’t we all know piping is a great concept of UNIX. Device driver talks to bus interface, and then to CPU, and you can insert other “filters” inbetween. Try that with Linux having zillions of drivers for a single chip.
Ok, I will shut up now.
http://en.wikipedia.org/wiki/Operating_system_advocacy#Pro_7
Usermode Linux provides equivalent functionality.
Better in some ways even.
Also, “all code is audited” is just OpenBSD marketing. What does it mean? Audited by whom?
Auditted by the OpenBSD developers.
To what standard?
Are there standards for this type of thing?
They have removed all (almost all?) unsafe strings from the code (strcpy, sprintf, and strcat). glib still doesn’t support the alternatives (portable OpenSSH has to bring them along).
When a bug is found, the entire code tree is searched for similar bugs.
Before something is committed, more than 1 developer looks at and tests the code, but I’m sure this happens in just about every project. Not sure if all projects have a “paper trail” of who looked at and tested it though. I don’t pay attention to the development of most OSes.
Linux isn’t on there because their uptime counter wraps at 500 or so days.
Sounds like a bug to me, but I agree with most of your statement.
Did Linux ever get the SCSI subsystems rewritten? I could probably find the answer if I really tried, but I’m lazy. I just remember hearing that this was needed a couple of times (2.3 days probably).
http://bulk.fefe.de/scalability/
as the vernacular goes … “nuff said”.
>Yes equivalent. It provides a way to partition the system.
not equivalent Jail is directly manage by the kernel
>Is it simply that you must have the last word about everything? Or do you think you have formulated a really succinct debate winner there?
I can say the same about you
>>OpenBSD supports SACK. FreeBSD doesn’t. I’m pretty sure NetBSD doesn’t either. So no, SACK isn’t supported in *BSD.
So why there is bench on the net about FreeBSD and NetBSD with Stack support ?
That is SACK support. Please refrain from commenting if you don’t know what you are talking about.
————————————————————–
I have badly tapped the name and so ?
http://irl.cs.ucla.edu/sack.psc.f.html
So how they do if SACK is not implemented ?
“It sounds like the principle thing that FreeBSD has over most Linux distros is that things are designed”
The main problem with linux vs *bsd flamefests is that they are pointless as long as one speaks of linux as such. (well they are pointless anyway, but a comparison of a particular linux distro and a particular bsd distribution, listing pros and cons can be educative). One can compare only a linux _distro_ to a particular *bsd – and it is a particular distribution that seems to be slapped together, not linux (the kernel). This fact is always forgotten – we almost always hear comparisons of linux with Free|Open|NetBSD.
I have seen many times that some people on osnews or slashdot tries to imply that the bsd user community looks down on linux – nothing can be farther than the truth! The best way to evaluate a community is to look at a community board. If you check http://www.bsdforums.org, you’ll see that it is a very helpful community (or check freebsd-questions mailing list, and you’ll see that the tolerance for newbies and ‘stupid’ questions is at least as good as on any ‘newbie’ linux distro forum). No RTFMs there – before FreeBSD (and before debian) I used mandrake, and one very positive aspect of mandrake was the community, both on mandrakusers.org and pclinuxonline. What I experienced on bsdforums is comparable to these boards.
bsdforums.org is the central place for *BSD users, and it is more representative of the user community than any random post here or ./. There isn’t such ill will towards linux there as some people would like us to believe, in fact, I only read good things about various linux distroes on bsdforums (of all places!). For instance, this thread is a veritable linux ad:
http://www.bsdforums.org/forums/showthread.php?s=&threadid=20137
Threads like this (your favorite linux distro) are not uncom
http://www.bsdforums.org/forums/showthread.php?s=&threadid=74
This is why I am very much disturbed with the kind of zealotry I described in my previous post. It does a great injustice to the community I am part of, and most people would first hear about the BSDs on boards like this and ./ – not the forums proper where _most_ bsd users gather. This would give a false impression of BSD users. The most important thing: FreeBSD is not rocket science! In fact, anyone who is not antagonist towards the command line (basically any debian or slackware user I guess) would find BSD extremely easy. I picked up the basics in a week or so (had a working system on the first day!) after my first installation! Of course, reading their handbook didn’t hurt either – for it is one of the best layed out documentation, and keeping it up to date is a priority for the project. Oh yes, this is another ++ for FreeBSD (at least for me). Developers usually write a new tool parallel to writing the man page for it. The handbook and related stuff is in a separate doc tree (you can download – I think sgml – all the sources with one command and rebuild all the documentation on your system as soon as they are updated online! (I have knewsticker scrolling headlines on my KDE panel, and both the ports and the src tree has rss feed – so I immediately see updates to various ports and the src tree – including documentation – right on my panel! you don’t have to configure it either, its part of KNewsTicker panel applet by default) Of course, gnu man pages are excellent, and the difference is at first subtle enough, but it will be visible once one is accustomed to bsd man pages. Basically there is a man page for everything (config files as well) – not to mention sample config files in /usr/share/examples.
I only write this because there is no place for “I’m cool cause I use BSD” type of users, simply because the easiness of the system. One wouldn’t earn geek points by using a system that is easy to learn and understand!
That benchmark page posted by “tech_user” is bogus, telling people to leave openBSD just because of just performance? thats absurd! there are alot of other reasons to use an OS outside of performance….(security comes to mind….useability, support,etc) Though it was interesting to see FreeBSD keep up on the benches.
>Yes equivalent. It provides a way to partition the system.
not equivalent Jail is directly manage by the kernel
Garrh. I am talking about equivalent functionality, not identical implementations.
>>OpenBSD supports SACK. FreeBSD doesn’t. I’m pretty sure NetBSD doesn’t either. So no, SACK isn’t supported in *BSD.
So why there is bench on the net about FreeBSD and NetBSD with Stack support ?
Umm I don’t know, maybe because it is a RESEARCH project? That was done in 1997, they wrote the code and it hasn’t been incorporated into either FreeBSD or NetBSD either because they didn’t want to contribute it or it wasn’t a good implementation.
I am a FreeBSD user and I think it is a great system, but my primary OS is Linux. I think that the almost religious fervor of the Linux community is a very persuasive force in the expansion of Linux. When I talk to people about Linux they notice my enthusiasm for the system and they often question me about it later or ask for help getting a system up running. I usually provide free support to any Linux users I know, but ask me about Windows or OS X and I am on the clock. I make my living providing technical support to Windows users. This enthusiasm is the primary reason that Linux use has expanded beyond FreeBSD use. Neither platform is substantially superior to the other, but Linux gets more press do to this advocacy. The Linux community is on the right right path by promoting it as a philosophy and a code of honor as well as an operating system.
Can’t we all just get along?? I don’t mind a good OSS/CSS battle, but OSS vs. OSS is ridiculous :S
I didn’t find the article offending at all; I’m a Linux user (huge MDK fan) and I think he spoke quite fairly. I’ve never gotten into a “fight” with a *BSD user before, but I did have my share of battles with “yer average Linux-fanatic” (I promised to drop the z****t word).
Okay, okay, there is a tiny flair of arrogance throughout the article, but hey, I don’t mind a little bit of arrogance, I’m not all too humble myself either (being Dutch implies arrogance, as far as I’m concerned
).
The linux fish – shaped like a shark – is a joke, of course, not a religious symbol.
I’ll only be happy when all these OS’s are easily swappable. All your apps and setting will keep working even if you swap out the kernel from Linux to BSD to MACH or whatever. I think the day’s coming soon.
“that supports most CURRENT hardware.. I am looking to get into the BSD channel. are the guis any good, or are they very minimal.. I don’t mind, and how much like SkyOS is the average BSD build.. (if at all) I haven’t tried either, but am looking.. any suggestions?”
If you mean desktops by this, Yes … FreeBSD 4.9-STABLE and 5.2.1-CURRENT come with Gnome 2.4 and KDE 3.1. The upcoming 4.10-STABLE release will have Gnome 2.6 and KDE 3.2 and should be released fairly soon. It is at the RC2 stage right now if I am correct.
My preferred idealistic social-economic system that has never been properly implemented on this planet (and never will be) is better than yours. Nyah nyah nyah nyah.
Fun fact of the day: “utopia” literally means “no place.”
As I understand it, any code put into BSD can be forked / hijacked by a commercial company (M$ ready for Unix?) and used for proprietary products that can be sold by them without
providing the source code to the purchaser.
Why work for free unless you are contributing to FREEDOM?
Why promote something that others can make proprietary?
Please correct me if I am mistaken or haven’t understood the license correctly.
Sigh…
When anyone makes claims such as “undisputedly superior” (pro-Linux), or “FreeBSD is more scalable and reliable than any linux machine. It runs applications way faster…” We must learn to tune them out. Period.
There is no easy answer to whether Linux or FreeBSD is better. Each has advantages and disadvantages, and involve as much personal preference as technical rationale. In all these discussions, the best thing to do is learn the pros and cons, and make an informed choice based on your needs/preferences.
I have spent plenty of time with both Linux and FreeBSD, both as a server OS, and as a desktop, dedicated kiosk, etc… I have found many areas where Linux had an advantage, especially with regard to (untweaked) performance. In general, I found that it takes more tweaking on FreeBSD to get the performance you need, BUT that overall management of FreeBSD, (packages, kernel configuration, updates, etc…) is the most consistent thing I have seen in ANY operating system. Personally, I think that is FreeBSD’s strongest point, rather than endless arguments over benchmarks or extreme cases (how many of us are going to need a 500-CPU machine?). Linux’s strong points are orthogonal to FreeBSD’s strong points. Linux development is a little less predictable, and often more fun to watch than FreeBSD. It can result in more variety of focus, with some amazing things happening in different “pockets” of development.
FreeBSD is a system. Linux is a kernel which is then wrapped up in many systems, depending on the distributor. Arguments over stability of the kernel itself are pointless, unless you get VERY specific, because stability depends on so many things with regard to configuration, etc…, but the FreeBSD system overall changes less often than the many Linux distributions, which is simply because FreeBSD has centralized management, while Linux does not. Each approach has pros and cons. Consider your goals and needs, then choose the one you like.
On the purely technical level, there are so many arguments for and against different approaches taken by FreeBSD vs Linux that there is no easy way to sum it up. Even the discussion of the latest kernel schedulers is a case in point. Linux kernel 2.6 has NPTL, while FreeBSD 5.2+ has KSE, implementing an M:N threading module. There are theoretical arguments why the FreeBSD model will ultimately be better than the Linux one, while there are practical arguments for the Linux model, but I am willing to bet that for most peoples’ usage the differences won’t mean anything. Benchmarks will show this, benchmarks will show that, but in the end, it comes down to your own testing for your own environment.
For my own needs, I finally decided on FreeBSD (only marginally) over Slackware Linux, and the decision was mainly because of the easier package management; performance and stability of each OS was… great. I still enjoy Linux. I still enjoy FreeBSD. All the stuff most of us actually use works equally well on both, (except for Java, but that is unfortunately Sun’s fault for not allowing FreeBSD developers the same access as Linux.)
In my opinion, weighing the pros and cons is the only way to have decent advocacy for any OS. That includes Windows. I get asked for advice on OS choice all the time. I sometimes tell people they should stick with Windows, sometimes I tell them Mac OSx, or FreeBSD, and sometimes I tell them Linux. Depends on where they are and what they need.
As I understand it, any code put into BSD can be forked / hijacked by a commercial company (M$ ready for Unix?) and used for proprietary products that can be sold by them without
providing the source code to the purchaser.
Why work for free unless you are contributing to FREEDOM?
Why promote something that others can make proprietary?
I’d heard a quote once that I liked
“The GPL aims to liberate the code, BSD aims to liberate the developer.”
The GPL and BSD strive for different types of Freedom.
BSD strives for “Freedom of Use” – you can do anything you like with the software and nobody can stop you. This *includes* putting it into proprietary software and hiding enhancements etc.
GPL strives for “Freedom of the software” – it protects the existence and the development of the software by making certain that all enhancements/changes/developments (within the limits of the GPL) have to be made public as well as enabling *everybody* to modify/enhance/develop said software.
It’s just different philosophies, personally, in the true sense of “freedom” I see the BSD liscense being more “free” than the GPL. I can take BSD code and do whatever I want to do with it; with the GPL I have to share what I have done. But that’s just my view.
Err, FreeBSD jails are not even close to UML in performance or features. They are two completely different beasts.
jails, as supported in FreeBSD, provide a way to partition the system’s processes, hardware, and networking under a single kernel with a single network stack (the multiple stack patch set only works on 4.x, and from what my friend showed me, not very well).
UML is a completely new kernel running in user space on top of the Linux kernel that uses special hooks to get access to the hardware. This allows for, even requires, independent network stacks for each UML instance. In addition to this, UML requires a full install of the Linux kernel and libaries in each instance, eating disk space each time. The implementation of UML also yields large performance penalties.
jail isn’t perfect though. While it’s implementation doesn’t automatically give a performance hit, it will yield the same performance after UML takes the hit as UML due to both needing to work from the a disk shared with the host (well, not necessarily, and if they aren’t shared then you won’t take any performance hit here either, so jail would be almost free). jail also only requires the binary and required libraries, not the entire bootup components that UML requires, as jails aren’t necessarily providing a full “root” experience.
If you want something to compare to FreeBSD’s jail, try the grsecurity patches. They provide, an IMO enhanced, jail facility for Linux that is more suitable to hosting, a common application of UML in my experience.
People, please check your facts before you compare things, so you don’t look like a jackass.
“Jail”
Patches for Linux provide this among better chroot protection. There’s also chroot and a few other utilities. Moot point, if any at all.
“That benchmark page posted by “tech_user” is bogus, telling people to leave openBSD just because of just performance? thats absurd! there are alot of other reasons to use an OS outside of performance….(security comes to mind….useability, support,etc)”
Just because there are other reasons than X or Y to chose between a A, B and C does not mean a research on Z is bogus. It just means like, as you already put it after your 1st sentence, other arguments are (also) (possibly) important to chose between A, B and C.
“This is unbelievable. This just proves the point of linux & religion, even though this wasn’t exactly his point.”
What if his point contains a flamebait or troll?
“NetBSD, DragonflyBSD & FreeBSD are more scalable and reliable according to benchmarks and statistics from reliable sources like netcraft.”
Uptime is not an accurate meassure; critical downtime is*. Uptime does not say anything about reliability, scalability either and Netcraft is not able to research either reliability or scalability. Your logic is flawed.
For example nobody except you cares when that desktop computer of yours running <fill on OS here> crashes every day. However when a power plant running <that same OS> goes down and it results in problems people DO care. When the intranet server goes down on 3:00 after work was done on the database to upgrade <fill on OS here> nobody cares while the uptime went to 0 again. However when the intranet server does not go down running <that same OS> which has a security vulnerability the uptime can stay up while the security got breached by a co-worker the other day. You prefer the high uptime over a security fix? Fine, i don’t.
“so ..yeah.. .. they just have less rpms it seems.. and less corporate backing in the beginning…”
[Beavis]Huh-huh. He said BSD’s have less RPMs than “Linux”. Huh-huh.[/]
[Homer]Doh![/]
Just what does that have to do with the whole discussion? “Linux” doesn’t have a kernel, RPM (the system) works on *BSD and most Linux distributions i’m aware of however Debian ain’t using these. What does your statement say about quality? Nada.
Reflekt: The BSD’s (BSD licensed software) can be commercial just like GPL software can be. If you want commercial support, consult the homepage of the BSD you’d like to run. There are various links over there to companies providing commercial support.
”
“The GPL aims to liberate the code, BSD aims to liberate the developer.”
”
gpl aims to free people from using non free code
“There is no easy answer to whether Linux or FreeBSD is better. Each has advantages and disadvantages, and involve as much personal preference as technical rationale. In all these discussions, the best thing to do is learn the pros and cons, and make an informed choice based on your needs/preferences.”
Good post. Too bad more of us can’t be so objective…
Thanks z1xq.
Well Eugenia (or whoever moderated it down) I guess freedom ends where your beliefs begin eh?
“Every election is a sort of advance auction sale of stolen goods.” -HL Mencken
“Every election is a sort of advance auction sale of stolen goods.” -HL Mencken
“Every election is a sort of Idols: let dumb people elect an even dumber contestant.” -T. A. Holwerda
Let’s not get political here, in the end we’re discussing endless rows of 1s and 0s for cryin’ out loud.
“The GPL aims to liberate the code, BSD aims to liberate the developer.”
The GPL and BSD strive for different types of Freedom.
I think this is basically correct. However, when you say:
It’s just different philosophies, personally, in the true sense of “freedom” I see the BSD liscense being more “free” than the GPL. I can take BSD code and do whatever I want to do with it; with the GPL I have to share what I have done. But that’s just my view.
Which are you? An end user, or a developer? If you are saying that you think the BSD license is more free because you can, say, take the code, enhance, and then sell it without disclosing the changes, then you are considered a developer. Users don’t care about such rights.
I don’t find it surprising at all that someone who considers himself – or will one day become – a developer would consider the BSD license to be “more free”. Obvisously, for people whose biggest interest is to make a buck off of software, its going to be the BSD license. I would also find it equally unsurprising that an end user, who has absolutely no intention of selling software (and doesn’t know someone who does and is therefore not sympathetic), would prefer the GPL.
So, I think the real question to answer to get to the bottom of the BSD vs GPL debate is this: “Which is most valuable to society as a whole: Developer/Business freedom, or Code(End-User) freedom?”
“I can take BSD code and do whatever I want to do with it; with the GPL I have to share what I have done”
Is untrue, not accurate at best. You only have to share the source containing your changes when you share the binary containing your changes (either gratis or for a fee, which is totally unrelated).
Developer/Business freedom
Users cannot create software, while people of all types need jobs/healthcare/homes etc. The most hilarious accusation of the GPL crowd is that companies that take released source code and sell it are ‘stealing’. How you steal something given to you is beyond me.
Quite frankly if what they have added to the software is not worth the price they charge, then people will not buy it. If the additions are worth it they will. (If they spend their money poorly they pay the ‘bad judgement tax’ – we have all paid it at one time or another). With the money spent the evil company who released it provides jobs,healthcare,etc.
Its quite a myopic view of society to think someone can create money in a vaccuum and not benefit their neighbors in any way.
The real problem of the RMS crowd is that somewhere someone seized a financial oppurtunity that they did not and made money from it.
I hope this is read somewhat before it is modded down.
“:
Users cannot create software, while people of all types need jobs/healthcare/homes etc. The most hilarious accusation of the GPL crowd is that companies that take released source code and sell it are ‘stealing’. How you steal something given to you is beyond me. ”
there is no such accusation. in fight right inside gpl it is clearly said that you can charge for it.
“The real problem of the RMS crowd is that somewhere someone seized a financial oppurtunity that they did not and made money from it. ”
rms said no such things.
How can anyone call GPL “FREE”, when you are forced to do something when writing software under the license? Doesn’t sound “FREE” to me.
“How can anyone call GPL “FREE”, when you are forced to do something when writing software under the license? Doesn’t sound “FREE” to me.”
how do call something a free country when you cant walk in the middle of the road?
same thing.
free doesnt mean free for all. thats called public domain. gpl is a free license. not public domain software.
even bsd has restrictions
you cannot remove the copyright notice from the source code is one such restriction.
“Let’s not get political here, in the end we’re discussing endless rows of 1s and 0s for cryin’ out loud.”
Technology controls the flow of information. This directly impacts your freedom of speech. Politics and these little 0s and 1s are hopelessly intertwined.
Sigh again…
Can’t the proponents of each license get it into their heads that the “other guys” have different aims? I personally prefer the BSD license: simple, straightforward, open, but I understand completely why others might prefer the GPL, even for business reasons. It comes down to your needs and preferences again, and I think its great that there is room for both.
Once again, as with choosing your OS, what you need to do is make an informed choice based on pros and cons. If you work with GPLed code, and agree fully with the demands involved, then you should keep that commitment, and have every right to expec users of your GPLed code to keep that commitment also. If you do not agree, then you should not work with GPLed code. Ditto with BSD-licensed code.
Even large companies might have valid reasons for making some of their code GPL, or supporting GPL code. However, I do think that the GPL is a little more “prickly” to deal with, if you are not careful. (See a more recent OSNews piece about Usenix: http://searchenterpriselinux.techtarget.com/qna/0,289202,sid39_gci9… )
But, enough of the political parallels: you can only infringe freedom if you arbitrarily violate or restrict someone (read John Stuart Mills “non-initiation of force” precept). That applies to GPL, BSD, or closed source. My response on each point of argument:
1. Closed source: If a seller and a buyer agree on a closed source model, and each enters into that agreement freely (yes freely!), then who are you to tell them their freedom is being restricted? As long as the software itself doesn’t do something sneaky such as install unwanted software, etc…, then there is no violation of liberty.
2. GPL: No one is forced to use GPLed software. The developers and the users freely enter into that agreement. Don’t like it? Don’t use it. (However, I suspect that RMS himself would love to actually *legislate* the GPL as a requirement for all publically released software, which would be a clear violation of civil liberty for developers.)
3. BSD: if the developers of BSD-licensed code freely choose to release their code, and they understand the terms of the agreement, then whose business is it to complain if someone incorporates that code into a closed-source product? BSD-oriented developers encourage all usage as a positive thing.
In all cases, as long as all parties involved agree as to the requirements, and those requirements don’t actively infringe on someone else’s life, liberty, or property, then WHAT ON EARTH IS THE PROBLEM? (sorry for the rant…)
sorry if i was harsh regarding the GPL. I think its good in some cases like promoting standards. I’m not against it.. don’t worry. I release my dead software or things i dont think will make it under the GPL.
when people say ‘software communism’ that usually look toward harsh governments that kill their people.. thats not what i ment… sorry if that providen some offense I was fired up.
i could.. create my own software and have the rights and add some bsd code to make it work better or i can integrate it with gpl code and not keep any closing rights.. then people build onto that and expand and your not successful with it except selling support services which millions of people already do.. especially overseas which do it at like $5/hour…… sorry if i provided any offense to the linux community. i use it to ya know 😉
theres just so many people that trash the bsds and their licenses and their alot larger due to big backing early on because it was new and written from scratch.
hard to be a bsd advocate ‘des days
Why do I get the feeling that these linux advocates’ machines are just being used as desktop playthings and not production servers ? Most Linux distros to me still feel like thrown together components rather than a cohesive os (notable exception : slack)
I don’t care about all this feauture-itis when I set up a production server, I do care weather the machine is stable, reliable and easy to update. That’s why I choose FreeBSD.
Also speaking as an AIX admin, if learned more about unix from BSD than Linux. Linux teaches you Linux (and a particular brand of Linux most of the time) and gnu’s not unix.
FreeBSD does the job and it does it well.
(Disclaimer: I like Linux, love freebsd and aix pays my bills – if you do not agree with me you are most probably wrong 😉 )
“Why do I get the feeling that these linux advocates’ machines are just being used as desktop playthings and not production servers ? Most Linux distros to me still feel like thrown together components rather than a cohesive os (notable exception : slack) ”
This is so very true. Every time, I install a particular software on Linux, I’m left scratching my head where the installation files went. Also, for some strange reason, regardless of what type of program it is, they will end up somewhere in the /etc directory. There is no logic to where they will end up, you just have to search through the thousands of files and directories for the thing you want. Also, for some strange reasons a lot of companies *cough* RedHat,SuSe *cough* have decided to create configuration files for everything you can possible do with your system. So rather than use a simple script that sets up your IP address, gateway, etc., you have to go looking through the /etc/sysconfig directory to do the things you want. The worst part of it is that they’ve added so much complexity to these files in the directory, and we get nothing out of the complexity. If you want to do advanced routing, firewalling, bandwidth management, you still have to create custom scripts. So in essence, you get a jumbled mess of configuration files with nothing to show for it.
“Also speaking as an AIX admin, if learned more about unix from BSD than Linux. Linux teaches you Linux (and a particular brand of Linux most of the time) and gnu’s not unix. ”
I personally have no problem with Linux being Linux and not being Unix. As long as it is able to get the job done, it can be whatever it wants to be *ahem* Windows. My only beef with Linux is that they make it so complex for no reason.
Linux is used for servers just as much as FreeBSD, and while it does not have quite the stability record that FreeBSD has I would put a Debian Stable server up against a FreeBSD release server anyday. I love FreeBSD, but it is a fact that most of the apps on top of FreeBSD are GPL ports of Linux applications. So, how stable your server is, is dependant on how stable the apps are and the GPLed Apache Server is quite stable on Linux or FreeBSD.
“I don’t care about all this feauture-itis when I set up a production server, I do care weather the machine is stable, reliable and easy to update.”
So does Debian GNU/Linux Woody.
“That’s why I choose FreeBSD.”
It’s why i chose Debian GNU/Linux Woody ;p
Debian GNU/Linux Sarge is rather used on desktops. It contains newer software, especially regarding the desktop.
“I don’t care about all this feauture-itis”
Don’t kid yourself. You do care about that. You do want certain features on your rock stable production server. What you do not want is some certain features not available on your current production server while they’re available on other competitors; that’s slightly different than you put it.
Don’t believe me? Ok. Then you do not need a kernel as feature of your OS. Please remove it, and reboot that production server. Let’s see how productive it remains…
A more interesting question which you did not address is “which features”.
“Also, for some strange reason, regardless of what type of program it is, they will end up somewhere in the /etc directory.”
What are you talking about?
Gentoo never updates anything in /etc without your permission. You’ll have to run etc-update in order to do so, providing you a variety of choices. Debian asks you too, and dpkg -L package shows you exactly where certain files from a apckage reside or were installed to while apt-file search filename shows you to which package it belongs.
IMO there’s not much wrong with package management in these 2 distributions, therefore i don’t understand what all this general rumble about some kernel is about. I cannot speak for RPM-based distributions, i do not care much for RPM-based distributions, but i do care for general statements regarding Linux distribution(s) while they’re strictly about RPM or do not apply on their package management system(s).
/etc/sysconfig doesn’t apply on these 2 distributions either. Neither on Slackware.
“GPLed Apache Server”
Apache ain’t GPLed. It is distributed under the Apache License. See http://www.apache.org/licenses/
I stand corrected on that point. This license, while seeming mostly in compliance with GPL, is certainly not a FreeBSD copy center license. So my main point stands.
“Mr. Newbie couldn’t take the abuse and RTFMming of the FreeBSD community, so installed Gentoo instead.”
As if no Linux users ever tell you to RTFM, right?
All of that comes mostly with 2.6 which is less than a year old in terms of actually being usable in production systems. Prior to this FreeBSD was far superior to Linux in a number of areas.
Not to mention most of what you describe was donated to Linux by corporations, hardly an accomplishment of the Linux community. It’s called marketing momentum, Linux has it and BSD does not, hence Linux get’s the donations. Regardless of what Linux get’s technically, I and many others will still not trust it nor like it due to the Linux communities fanatical “linux must be everywhere” crap.
By the way, you might get a lot more respect if you didn’t puff a whole lot of hot air with your facts. You’ve a number of unsubstantiated claims, such as about softupdates not being as safe as journaling.
“Not to mention most of what you describe was donated to Linux by corporations, hardly an accomplishment of the Linux community”
hell no. its called business not donation and the community had smp and stuff before ibm or caldera was involved though it certainly helped having commerical full time participation from a number of complains. that explains the lead in 2.6.
Although an older benchmark
Linux systems, solaris systems, windows systems, and un- tuned freebsd systems.
Where Linux systems having a significantly higher performance level, under network stress.
I thought the old myth of freebsd being a better networking operating system died with this benchmark.
And with the integration of TCP vegas technology in 2.6.6… the rest is probably history?
http://www.samag.com/documents/s=1148/sam0107a/0107a.htm
I know most of you already know this but here it goes:
http://www.linuxworld.com/story/44799.htm
http://www.llnl.gov/linux/thunder/
Some more recent benchmarks that at least make some attempt at a rigorous approach: http://bulk.fefe.de/scalability/
In the end, if you read carefully, you will see that there are (trumpet flourish…) pros and cons to Linux 2.6 and FreeBSD 5-CURRENT, but that each has some impressive scaleability gains. It seems that in the end (after some kernel fixes) NetBSD outperformed them all on some networking benchmarks, but overall, the results don’t show any clear winner.
Again, remember that like all benchmarks, these are artificial, and actual performance should be judged by your actual tests for your actual application.
“(Disclaimer: I like Linux, love freebsd and aix pays my bills – if you do not agree with me you are most probably wrong 😉 )”
Or we speak decent English.
You’re right, Linux isn’t Unix. Linux and GNU do not try to be Unix: That’s one thing I really like about Linux.
FBSD is a beautifully put together product, I will most definitely give you that one. The installer is wonderful in that it only installs and turns on services you tell it you will use. It asks questions, something many installers need to learn to do.
lol..some are using the arguement of more or less apps for freebsd.
the idea that most of freebsd’s ports are gnu.. its completely baseless and untrue. alot of gnu’s ports run on freebsd and alot of people use them because of the popularity of linux.
what? comes with alot of packages? have some of you not compiled your own software? your using that as an arguement? hmmmmmmmmm…
to my understanding, freebsd runs bsd and linux.. maybe it can run more apps.. but more apps dont make it a great OS.
Running linux on a computer on an extremely fast server dosent necessarily make it the most powerful. Irix & Unicos are campions…I’m sure freebsd can be ported to a system with as many CPUs as linux too.
Err, FreeBSD jails are not even close to UML in performance or features. They are two completely different beasts.
So for all your rambling, you could only find a few implementation and minor differences?
jails, as supported in FreeBSD, provide a way to partition the system’s processes, hardware, and networking under a single kernel with a single network stack (the multiple stack patch set only works on 4.x, and from what my friend showed me, not very well).
UML is a completely new kernel running in user space on top of the Linux kernel that uses special hooks to get access to the hardware. This allows for, even requires, independent network stacks for each UML instance. In addition to this, UML requires a full install of the Linux kernel and libaries in each instance, eating disk space each time. The implementation of UML also yields large performance penalties.
Umm that is an implementation detail. UML is actually a much cleaner implementation of the idea because it doesn’t require teaching everything in the kernel about it.
And you are really scraping the bottom of the barrel with the single/multiple network stacks thing.
Performance wise, jails actually can have a significant impact on performance. UML also does, but a feature in the new 2.6 kernel’s memory mapping API allows vast improvements in performance, making it very viable for jail type applications.
snip
People, please check your facts before you compare things, so you don’t look like a jackass.
I suggest you would be wise to to the same. The only differences you could come up with were something about network stacks, performance, and a UML instance requiring its own image. The last one is even a non issue because multiple UMLs can share readonly data.
Anyone here noticed this little announcement?
http://www.freebsd.org/releases/4.10R/announce.html
there is a community myth that Linux is technically superior to any other operating system in the known universe
Nah, the myth doesn’t go like that. It goes like this: “a community of incredibly talented and experienced developers, administrators, and users, joined by diverse technical interests and a common desire to build and use the best operating systems in the world”. I found it in the article and it’s the myth that is common to both BSD and Linux.
Besides, it’s not really called myth, it’s called advocacy.
Apparently, there’s so much in common with the *BSDs and Linux distros that the only way their users can figure out to advocate their own choice is to bash each other. There’s certainly a need for more articles that suggest other ways of advocating BSD.
“Apparently, there’s so much in common with the *BSDs and Linux distros that the only way their users can figure out to advocate their own choice is to bash each other. There’s certainly a need for more articles that suggest other ways of advocating BSD.”
I think you hit it right on the mark.
“I think you hit it right on the mark.”
Ditto. Let’s focus on the positives.