Slashdot Mirror
Clearing up FreeBSD confusion
Anonymous Coward writes "Daemon News has published an article that attempts to clear up misconceptions about FreeBSD. The article is primarily aimed at Slashdot readers based on the comments made in postings. " Well written, informative, and makes things much more clear.
"But many are simply curious about why a new user
;)
:)
would choose Linux over FreeBSD, despite FreeBSD's technical superiority."
Technical superiority.... ?
How? Because you have a more mature TCP/IP stack?
How does one define "Technical Superiority" without resorting to marketing-type hyperbole?
If you want to say "*BSD has been around longer, evolved from an existing, tested code base, has had more time to have its more traditional components tweaked, and is more tightly controlled" then please say that.
But don't say "Technically superior".
I remember checking out the *BSD platforms about a year ago and I could find *no* real SMP support to speak of. It was "in the works" as it went, and when I tried it (a dev. snapshot was all I could find...) I was greeted by a locked up box.
On the other hand I've been running Linux on that same box (updated processors recently) for well over a year and while the SMP performance has certainly improved recently, I've never had stability problems.
(And your claims for server support are rather thin if you can't support something that has been in mainstream use for years in other OS's... even under the dread NT
So which one is superior? Neither. *BSD is often more stable because of stricter controls on accepted code and a longer period of time to stabilize. Linux encourages innovation and new features (whiz-bang or otherwise) and often provides support for a broader range of hardware than other UN*X/*BSD OS's. (Notice I don't say "Superior Support" ?
The bottom line (as has been put often enough before..) is that both are excellent OS's (though both still have lots of things that can be improved). While loud-mouth advocacy may help sell cars and M$ products, I've always believed that those of us who run real Operating Systems should be mature enough not to fall into that trap. Have I been misled..?
Hello,
/. or LinuxToday that I can dig up if you are
While I applaud your effort to set some facts straight I believe your article on Daemon News probably did more to inflame any anti-*BSD sentiment by Linux supporters than anything else. I personally have never run BSD, although I have a CD for FreeBSD 3.1 CD and now have sufficient bandwidth via my cable modem to download any distribution over night. However, I do know a "little bit" more than your average Linux bigot about the development of *BSD and Linux. I'm not going to attack you or make you feel uncomfortable, rather I'd like to explain how some of your words are inflamitory to the Linux community. Hopefully this will help you in communicating with the Linux community in the future in a less inflamitory way (you honestly sound like a "BSD bigot" even to someone who is rather open-minded about alternative operating systems).
"And while calls of fragmentation of the BSD community run rampant throughout the Linux world,Linux is a far more fragmented operating system. With over a hundred distributions, Linux is often unable to provide compatibility with itself even though "there is only one Linux kernel." I was bitten by this several years ago when I wrote a set of scripts to manage the startup rc files on a Slackware system. When the system was moved to Red Hat, the scripts broke. Thousands of others had problems when the move to glibc was rushed by Red Hat while other distributions remained cautious."
This is common misconception that many people, including Linux users, believe and causes problems. Unlike the FreeBSD "current" code tree there is no globally accepted recommendation to upgrade to the latest kernel or distribution. It's also a mistake to consider the various Linux distributions as the same operating system. Much like the various BSD OSs, the various distributions of Linux are designed for specific purposes. Caldera is squarely aimed at the corporate market. RedHat is the "average Joe" version. Turbo Linux is aimed at the asian market (with specific Japanese and Chinese versions). Slackware is for the Unix "guru" who has no problem with recompiling programs by hand. Failure to understand this is just that, a failure to understand the Linux marketplace. Similarly, distributions don't generally follow the "latest" releases, even when if comes to the standard releases. For example, RedHat 6.0 is still on the 2.2.5 kernel release while the official kernel is at version 2.2.12. While this is not "tying," as anyone with enough knowledge can run any software on any system and upgrade the their hearts content (I'm running 2.2.12 right now on a RedHat 6.0 box with PPTP patches, XFree86 3.3.4, gcc 2.7.2.3 (to compile the kernel) and gcc 2.95.1) it reinforces the concept that each distribution can generally be considered a "different" operating system much like the various *BSDs are.
One would have to ask though, why did you upgrade between Slackware and RedHat without investigating the differences? That's not a "normal" administrative move. It's actually more like reinstalling an operating system than upgrading. The differences in package methods should have given you a huge clue that the OSs were not the "same." That the rc scripts are different in various OSs is well know to any administrator. Solaris, HP-UX, *BSD, Unixware, and practically everyone else does it slightly different. Linux gives you both common methods of doing it. It's much like the perl programming language, "there's more than one way to do it." I really don't mean to imply that you're not a competant administrator, but I'm really surprised that you had this issue. Perhaps it wasn't your choice to make the "leap" between Slackware and RedHat. If that's the case, then the administrator who made the choice, especially on a production box, is at fault and probably should have had disceplenary action.
Lastly, I think it's fairly unreasonable to imply that there are "hundreds" of viable Linux distributions. A vast majority of those are only used by a very small number of users and for specific reasons. They can almost be thought of in the same vein as embedded systems (such as the "distribution" meant to use Linux as the OS for a router). There are actually only a few "large" distributions that the vast majority of the Linux community use. The number, while slightly larger than the number of *BSD "distributions" comes no where near "hundreds."
"Many others have mentioned that FreeBSD should quit complaining about the GPL while using GCC and other GNU tools. While it is true that a number of GNU utilities are used in FreeBSD, they comprise fewer than 8% of the utilities and 15% of the libraries. "
My personal view, which I suspect is shared by the vast majority of GNU fans, is that if BSD continues to use the gcc compiler then the percentage of other GNU tools that comprise a BSD system is relatively irrelavent. Most people I know view the C compiler in a Unix system to be one of the most important aspects of the system, much like the kernel. It would kind of, but not exactly, be like BSD using a HURD or Linux kernel and still insisting that it's BSD. This is in contrast to the GNU camp, which I think believes that the tools define the name of the OS and not necessarily the kernel (hence the GNU/Linux debate). It could still be argued that it shouldn't be called GNU/Linux in this case, but that's a side issue. I personally consider the kernel, C compiler, and C library as the main determining factors as far as what "kind" of system an OS is. Correct me if I'm wrong, but it's definately possible to grab the source for FreeBSD and recomile everything on a RedHat Linux system right? If so, does that transform the system into a BSD system, with the C library, compiler, and kernel based on RedHat Linux? I didn't think so.
"The Linux mindset can often be characterized as "code exists, throw it into the distribution.""
This is mildly offensive. I'm assuming that you are specifically talking about specific distributions instead of the kernel itself. If you're talking about the kernel then you are sadly mistaken. First, Linus has the ultimate say in what goes in the kernel. Second, there is a distinct development and stable kernel. The development kernel is NOT for your casual users and is expected to fail on a regular basis. Yes, code is often thrown into the development kernel so that it can be tested on "live" systems of the developers. This code is either tested to the point that it is considered stable or yanked out of the kernel before it is moved over to a stable version. To imply that untested code is "thrown into the distribution" is the offensive part. For example, I've been trying to get the new RAID code into the stable kernel but it was yanked at the last moment, even though it has been available and in test kernels for months, because a developer found what could have been a bug. While this was disappointing on my part, it shows that code is not "thrown into the distribution."
If you're talking about non-kernel parts of the distribution then I think you are mistaken also. As I mentioned, there are only a few different distributions that are used by the vast majority of the community. One of the reasons for the "fringe" distributions is the fact that the main distributions don't always update their code often enough for some on the leading (not bleeding) edge. That's why, for example, SGI Linux is basically RedHat with some updates. I belive another popular "fringe" distribution may be Mandrake, which is also based on RedHat with the latest updates for various packages. So, to claim that main distributions are just "throwing [code] into the distribution" is patently false. The fact that most Linux distributions make it extreamly easy for any old user with some experience to create a custom distribution is probably more of an example of the openness of Linux development than "fragmentation" or "code instability."
"Linux was developed by an undergraduate student at the University of Helsinki to correct the flaws of Minix. However, FreeBSD is based on the 4.4BSD distribution of Unix from the University of California at Berkeley released in 1994."
Here you imply that FreeBSD is somehow better because it was released by the high and mighty "University of California at Berkeley" while Linux development was started by a poor "undergraduate student" at the lowly University of Helsinki. While I don't believe that was your intention, at least I hope not, that's the way it comes off. I could care less about the University of California at Berkeley, much like I could care less about the University of Helsinki. They are both "institutions of higher learning." I don't think it's fair or logical to rate an OS on what University it blosomed out of. One can always ask the questions "what have you done for me lately!" In that regard, I suspect that the University of Helsinki gets my "vote" for the most innovative.
"The first widely used TCP/IP stack was included in 4.2BSD and was reused in dozens of other operating systems."
Here you imply that the Berkeley TCP/IP stack is used in Linux. Although I'm no expert on the matter, I believe this has been discussed ad infinitum and shown that the Linux stack is a "new" implementation. A comment I would also like to make is that you seem to imply that, since BSD has a long heritage that it is somehow "better." I, for one, don't believe that just because something has been around for so long that it's necessarily better. I'm not saying that everything must be constantly rearchitected, but has been shown that it is often necessary to rewrite from scratch from time to time instead of holding on to old code for the comfort factor. I believe there was an article referenced on
interested...
"There are No Applications for FreeBSD"
I completely agree with you here as that statement makes as much sense as saying that there are no applications for Linux. I do think that the comment about performance improvements by running Linux applications under FreeBSD was unnecessary, even if true. I thought your article was supposed to be about correcting misinformation about FreeBSD rather than comparing FreeBSD and Linux and showing that FreeBSD was somehow "better."
"FreeBSD is a Dead End"
I liked this part, it was informative and didn't include any jabs at Linux users. I didn't know about PicoBSD and would be interested in learning more... (This is the tone that your whole article should have taken).
"The majority of FreeBSD is owned by the Regents of the University of California, where it was originally developed. Removing the existing license without the permission of the Regents would be no different that releasing a version of GCC with a BSD copyright in place of the GPL."
I think everyone knows this. So why not petition the Regents to release the original code under the GPL? That would seem to solve all problems, would it not. You CAN release code under multiple licenses. I don't see why, with the release of the license by the Regents, that FreeBSD could not be released under both the GPL and the BSD license, do you? Besides, I would think that much if not the vast majority of code that was released by the Regents has been modified to the extent of being all but replaced, no? (And please don't take this opportunity to say "no, because the BSD code has such a strong history that there was no need to modify the original code." Everyone know that there are bugs in almost all software, including FreeBSD and Linux. While there are probably areas of FreeBSD that have not been touched since the original release I'd find it hard to believe that it's in the majority. If so, I'd have to ask "what have you guys been doing since 1994?").
FreeBSD SHOULD GPL itself if it wants to use GPL code. I though that that's just obeying the law. If you're telling me that FreeBSD uses GPL code and does not release under the GPL then I think that's a surprise to a LOT of people. I don't think "we" are trying to convert "you" to GPL bigots (well some are but...), however, I don't think that the various factions (BSD/OS, Free, Open, Net) petitioning the Regents to release their code as GPL SO THAT YOU CAN INCLUDE other GPL CODE into your distributions would be viewed as a "victory" by the GNU camp. Sure, we would "get" the Regents code under the GPL, but you could STILL release FreeBSD under a non-GPL "Regents" BSD license as long as you didn't include any GPL code.
This would not split the code tree, as the only differences would be in the license attached to any distributions. The base code would be available in both BSD and GPL licenses. Any additions to the code would have to be released under the GPL license if it included GPL code. But, you stated that the reason you don't do this is because of the license from the Regents. You imply that you would have no problem releasing under a GPL license if only you could. If that's NOT the case, and you simply don't agree with the GPL license, then don't use our code. You have no "right" to use our code, and complaining about it will do you no good. It's like us wanting to use BSD code but making the decision to not because of the license. Some code, like X, is under a non-GPL license, but we, like everyone else, have the right to pick and choose what we decide to use. I don't see any inconsistancy of GNU folk insisting that you release under the GPL if you use GPL copyrighted works.
I really don't understand what the issue is. I think it's kind of like an ego trip for the Regents and "advertisement" by the University of California at Berkeley. That's honestly the only reason why I can think that they don't release under the GPL. I mean they don't get any money for derrived works, right? They only get the free advertising.
"Some FreeBSD users may indeed be jealous. But many are simply curious about why a new user would choose Linux over FreeBSD, despite FreeBSD's technical superiority."
Why the reason for this insult? A new user would choose Linux because it's more in the news. A new user would choose Linux if they wanted an OS that was under blazingly fast development. May be a hacker who wanted to check out some esoteric hardware. Or, may be a "average" user who would probably stick with a major distribution and not compile applications or the kernel themselves (instead waiting for bug fixes and enhancements from the distribution vendor, much like the BSD twice yearly release cycle with security fixes in between). Or may be because "official" support from a variety of vendors (such as database vendors) has been announced for Linux but not BSD (much like user who wanted to run a particular database would most likely choose RedHat instead of a "niche" distribution because "official" support is pretty much RedHat's domain right now) [Side Note: I may be wrong here, but I certainly am not aware of any press releases by any major database vendor, for instance, about official FreeBSD support. To say that the application could run under FreeBSD may be true, but then why worry about official support. Might as well run some custom kernel and compile all programs yourself if that's what you want.]
"In many ways, this is how Linux proponents view Windows users. Others do not care."
Oh, no, more insults. I honestly can't believe you made this comment. After explaining above that both BSD and Linux have as a common goal to provide a free Unix environment to users they you imply that Linux is SO far behind BSD that it's more like Windows!?! How, then can BSD be so much better if it's striving for the same goal? No, Windows has a fundamentally different goal in mind and I belive the vast majority of Linux and BSD supporters agree that it's a "BAD" thing. To take fundamental control away from even system administrators and try to "dumb down" the administrative skills required is just a bad thing to do. It puts too much responsibility on the correctness of OS code that can have devastating results. Given the industry track record, not even Microsoft's specifically, on software bugs I'd say this is a humongous mistake (just as integrating the GUI into the kernel was). I really don't know what to say here other than I hope you didn't mean what was so plainly implied.
"Unlike Linux advocates, FreeBSD advocates do not believe FreeBSD should be running on every microchip. Most FreeBSD advocates merely wish it see it perform best where it does best: Internet servers and high end workstations."
No, but you have OpenBSD, NetBSD, BSD/OS, and at least PicoBSD, which are designed to run in various situations. I think this statement comes from the misinformation you have about Linux and the assumption that there is only "one" Linux operating system. There are multiple. The main distributions, like RedHat, only run on "popular" "internet server and high end workstation" architectures such as PPC, Alpha, SPARC, HP-PA and Intel (RedHat may not even run on all of those). Smaller, niche distributions run on the Palm Pilot, and other embedded devices. Hey, that's kinda like the embedded PicoBSD, right?
Conclusion:
May be I'm reading way too much into the statements in your article and if I am then I appologize in advance. However, I obviously don't think this is the case. Hopefully I have provided some constructive criticism that you can use to hone your ability to "deal" with us wild Linux bigots. It would have been much more useful, IMHO, to do as your article was stated was it's goal and "clear up many of the misconceptions" instead of trying to pit Linux against BSD. If it was only focused on providing information to the Linux community about *BSD then that would have accomplished the goals set forth. I realize that Linux would have to be mentioned in some of the topics you chose to address. However, I don't think the confrontational tone was constructive at all. More information on the new features coming out in *BSD, such as PicoBSD, would have been helpful. Less "hauty" attitude about the origins of Linux and *BSD would have helped (like I said, most people just don't care that *BSD came out of Berkeley).
I would be interested in a reply to let me know how you took my email. As stated numberous times, the goal of this email was to inform you about how your statements could have been more inflamitory than you intended. Hopefully I accomplished my goal.
Sincerely,
fwr
I was looking forward to a piece that cleared up myths about FreeBSD. I've had extensive experience and know better than to buy into most of the false statements directed at the OS.
While the author does a good job of outlining what those myths are, he does a terrible job of debunking them. If you want people to stop thinking false things, you don't go and try to convince them by telling them a bunch of other false things.
It's also not very smart to write an article aimed at misguided Linux users and spend about as much time trying to clear up some misconceptions about FreeBSD as you do exposing your own cluelessness about Linux.
eg: "FreeBSD is not a Linux clone" - This is quite true.
"Linux is a FreeBSD clone" - Please. Not even a little bit.
Neither OS is a clone of the other, but both operating systems have strived to become better, sometimes with enhancements inspired by the other OS. This happens every bit as regularly in BOTH camps.
Anyway...
--- Tao
Primarily aimed at slashdot readers. Hey! We're a demographic!
I have read some, but not all of the works of Marx. I know full well the Cold War propaganda has nothing to do with communism. I also know that at least 90% of the people who use the "GNU/FSF/RMS is communist" line are trying to slander the target by associating them with the propaganda.
Here are some of the key points of theoretical, Marxist communism, and how they fit or don't fit with the GNU Philosophy (note, some are quotes of a translation of the Communist Manifesto, others are paraphrases of concepts):
the history of all hitherto existing society is the history of class struggles
The GNU Philosophy does not divide things along class lines. With GNU, the starving shareware author is as much in the wrong as the Microsoft Millionare, and the rich hardware manufacturer with Free specifications and driver source code is as much in the right as the poor amateur Free Software developer. The people too poor to be able to access computers are not even addressed in the GNU Philosophy. This is not an oversight on GNU's part, it's just Communism is a Social, Economic and Political system, GNU is a information and development system.
The bourgeoisie have developed substantial means of production, economic exchange and wealth, dependant on the proletariat. The proletariat, once it has gotten strong enough, will use its power over these tools to overthrow the bourguoisie.
The proprietary programmers wealth is dependant on rampant consumption, it's not dependant on the GNU developers. The tools the GNU developers are using to overthrow the proprietary system are not the proprietary programmers works, but works made by GNU for GNU, or by non-GNU Free software developers.
The proletariat are being oppressed and kept propertyless by the bourguoise. The proletariat will fight back.
The GNU developers are claiming the proprietary vendors are oppressing the consumers, not GNU. GNU is offering a way out, not a conflict.
The communists want to put total political power in the hands of the proletariat
The GNU philosophy wants enough political power to reform IP law. More might be desired by some individuals, but it is completely outside the scope of the project or the philosophy.
Private property, and its corrolary, personal wealth are oppressive evils and should be abolished.
The GNU project considers intellectual property to be a flawed concept, but physical property is perfectly valid. As for personal wealth, the GNU project encourages developers to charge as much for their services as they can get away with. It creates a system where people can focus on paying for tangible services rendered, rather than intangible IP licenses.
In bourgeois society, living labor is but a means to increase accumulated labor [and therefore create capital]. In communist society, accumulated labor is but a means to widen, to enrich, to promote the existence of the laborer.
So people working creates Capital (or code in the case of the GNU project). The bad society keeps the capital for the people at the top, the good society uses it to enrich the people making it. I'd say that's a match.
Communism deprives no man of the power to appropriate the products of society; all that it does is to deprive him of the power to subjugate the labor of others by means of such appropriations.
I'd say that's a match there.
To create the communist state, you have to tranform the capitalist state. Among the many changes suggested to get the ball rolling are state centralized communications and trade
The GNU project stresses dispersed, decentralized communications and "trade".
Equal obligation of all to work.
The GNU project does not obligate anyone to help it. People do so because they want to, not because they have to obligation.
To each according to his needs, from each according to his abilities.
The GNU project pays no attention to your needs (avoiding the "Tyranny of Need" as Ayn Rand puts it) you use what you want to use. It's software, there's always more. You put in what you want to put in. Nobody is holding a gun to your head and saying "contribute more, we know you can". Thus the GNU project might go slower sometimes, and faster others, but it always continues, because someone always wants to contribute something.
Basically, it's a very incomplete match. I'd say that the GNU Philosophy shares a few scattered ideals with the Communist Movement, disagrees on others (irreconcilably so when it comes to work and property), and shares none of its ideas on how to acheive the ideal society. The GNU Philosophy is no more Communist than the Libertarian movement is. They are all idealistic philosophies, that doesn't make them the same.
----
----
Open mind, insert foot.
The most important question is not which kernel is superior, but whether all the Unix-like systems will run the same applications with no change other than a possible recompilation.
I use Linux, but I want to be able to share code with any and all Unix-like systems, free or commercial. I want all of them to be completely compatible so that I can choose whichever one gives me the best price, performance, or reliability for a particular application.
This article seems to contain just as much prejudice and ill-judgement as the rantings it criticises. To quote:
When the system was moved to Red Hat, the scripts broke. Thousands of others had problems when the move to glibc was rushed by Red Hat while other distributions remained cautious.
In many ways, Linux is a FreeBSD clone
But many are simply curious about why a new user would choose Linux over FreeBSD, despite FreeBSD's technical superiority.
Unlike Linux advocates, FreeBSD advocates do not believe FreeBSD should be running on every microchip.
Most negative talk about FreeBSD is baseless and intended to destroy, scare, or subvert potential users
All of these statements say or imply things about Linux or Linux users that are not generally true. Most ranting by Linux zealots about FreeBSD is based on prejudice and misinformation. Repeating the same mistake in the other direction doesn't help.
The BSDs and Linux are all excellent, and very similar, operating systems. The only way to compare them is with strict technical or legal comparisons. Rhetoric doesn't help anyone to make decisions.
Meanwhile, my friend scoured the net for the appropriate RPM's. Then found out he had to find RPM's for the libraries, etc.
not all distros were created equal. Those of us who use debian feel sorry for our little brothers who can't just type apt-get install and have apt d/l the app and all dependancies and set it up.
seriously, though, I've never actually used the FreeBSD package manager, but it is hard to imagine anything nicer than debian's apt/dpkg system.
When in danger or in doubt, run in circles scream and shout.
I use Linux all day, and all night. I have used it exclusively for years.
The main resons I switched to linux:
-I got bored with OS/2
-I wanted a more stable and robust OS
-I wasn't learning anything new in OS/2
-Windows was not an option
-The Linux 'headspace' seemed a better academic endeavour, and I looked forward to expanding my horizons.
One other main reson why I joined the linux bandwagon was becasue The crowd was technically good, reason and logical thinking seemed to be the aura of the community. I'm not so sure that this is the case anymore.
The surge in linux popularity is bitter-sweet for me. While attention is always nice, I would rather ATTRACT users by exaple of a solid implementation, and sound thinking than PROMOTE the OS into the mainstream.
Lately I find myself saddened by the sheer ignorance of some Linux advocates, and I feel dirty by assiciation.
It may be time for me to move on to BSD for this very reason. It seems serene over there, and I get the impression that their development model may be superior to Linux's.
Linus and Alan maintain tight control over kernel development. The best code floats to the top and get implemented. This allows for the kernel to be as robust and as solid as it has proven itself to be.
BSD's approach encompases the entire OS, not just the kernel. I would have to assume that having tight control over the entire distrubution ensures that the tools are solid, and most bugs have been worked out before something is added, as the author points out. Surely this is where BSD's legendary stability must stem from.
Another point is the package management. RedHat and Debian both have good efforts under way, and I'm sure many days could be spent debating which is superior. I think they're probably both equally crippled because both distributions depend on the development cycles of developers that are not necessarily affiliated with one distribution or an other. Control is lost. The Linux distributions are placed at the mercy of seperate development teams, with different goals.
Correct me if I'm wrong, but by FreeBSD having tight control over the inclusion of core tools into the distribution, they can build a secure and stable distribution much more easily than can be done for Linux. Also, the issues of package dependancies, upgrades etc become exponentially more simple to handle. In this regard, I doubt FreeBSD can be touched by any Linux distro.
Obviously contributed software or applications is a different matter. I'm speaking of the core tools that form a distribution.
Hopefully the LSB can and will solve this problem. I really hope it's sooner than later.
I like Linux. More and more I dislike the self-devouring and ignorant behaviour of the ill-informed Linux advocate that spews propaganda which they detest in the 'other faction'.
Try as many Linux distros as you can, and try FreeBSD as well, and decide for yourself which you prefer. I'll be trying BSD real soon now, then I'll make my own decisions.