Linux’s dominance “is not so much that it is better than BSD,” Meta Group analyst Tom Murphy said. Indeed, there is a perception that in some ways, BSD is far stronger than Linux, NewsFactor reports.
Has Linux Eclipsed Open Source?
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
and even though I prefer the BSDs, I think Linux based OSes are just as good, particularly Debian, Slackware, and Gentoo. I think that it just comes down to deciding what one is most comfortable with and going with it.
…that there were so many different BSDs out there. Even though there are hundreds of Linux distros, they are based on the same kernel, so you could say the Linux effort is more focused.
because it is not “perfect” or nessisaraly the optimal solution.
seriously, when it comes down to it, comercial computing solutions are decided on what is most flexable in a business sence.
Linux is far more flexable than BSD is(not in a technical sence mind you but in a comercial sence)
in 10 years, there will be flame wars over how Linux should never had hit the big time because *BSD was a better technology etc. just wait, you will see.
“Even though there are hundreds of Linux distros, they are based on the same kernel, so you could say the Linux effort is more focused.”
Not really. Every distro I play with has their own set of kernel patches, and the kernels are no more similar than those of the BSDs. An all too common misconception, even one that I myself fell for at one point in time.
Betamax baby
why is it hard to find concrete statistics on performance and stability difference between the BSD’s and the main linux’s?
things like the performance of the (g)libc maths routines, the dist I/O, networking I/O, latency times to respond to connection floods, stability to RAM corruption (or don’t they do that in these OSes?, i remember some critical ROM code which put in lots of jumps to “safe” code in the event of corruption), graphics performance using the same XFree86, schedular performance, threads performance, etc …
if there was one site dedicated to keeping us updated on all these things for the most popular OSes – then that would be very useful
I am by no means a guru at either linux or bsd, though from my experience bsd is much easier to work with. both seem very similar, with bsd more organized and stable, with linux more bleeding edge. Simple things like kernel changes and software updates are a joy on bsd, a pain often times on linux. Little things like directory structure and consistency of convention make bsd a much easier distro to work with in my humble opinion.
That being said, bsd lacks the size and power of the linux community. I very much doubt that in time that bsd will be able to keep up with linux. Even now little things like emerge make me wonder how long i can hold on to bsd.
for now… bsd is the best thing going. and i hope it says that way
This ” Linux is more popular than BSD ” affair must be good for analysts and reporters. Otherwise, they wouldn’t always do their best to fuel a fire that doesn’t exist.
This attitude is prevalent in the news : when it’s not “a mob lynched somebody somewhere”, it’s “the war between companies X and Y escalates”. To find another instance of this sort of business tactics, look no further than the bottom of OSNews website : there’s a mention of an article published by eWeek that is titled “Symbian vs. Linux : Who Will Win in Wireless ?” I’ve read it and according to the content, the topic should have been : “Why did Motorola sell their Symbian shares ? ”
Another way to attract people is to say things like “IBM and Novell versus individual coders : who will make open source enterprise ready ?”
All this noise has nothing to do with the operating systems qualities. It’s more about catching the eyes of readers.
“why is it hard to find concrete statistics on performance and stability difference between the BSD’s and the main linux’s?”
Mostly because there are so many different configurations that would skew the results in favor of either system. It would have to be a *massively* standardized suite of tests to be sure.
“That being said, bsd lacks the size and power of the linux community. I very much doubt that in time that bsd will be able to keep up with linux. Even now little things like emerge make me wonder how long i can hold on to bsd. ”
To be honest, I doubt that you are right about that. I don’t see BSD, Linux, Mac OS or Windows going away. They all have something anchoring them to their respective markets and userbases. As populations grow, so will adoption of all of them.
Never tried BSD however I have not had any problems with Linux yet, so maybe that’s why I keep using it. That’s one of the reasons why I got rid of MS Windoze.
I can’t get to the link.
Yep, here too.. I will try again in the next hours..
The greatest fault and the greatest strength of Linux is the GPL. On one hand, it allows the freedom we have today. On the other hand, it doesn’t actually let you “sell” it in the traditional way and thus makes commercial exploitation harder.
BSDs are in my opinion nicer for certain critical servers. Linux for desktops. Personally I’d prefer adminning a BSD box over a Linux one but the choice of software is bigger for Linux. So for example if you want a game server, you’ll go with Linux most likely. If the application is not OS-dependant, then go for whatever offers the best performance. BSDs and OpenBSD in particular have a great security track record. On the other hand, BSDs tend to lack good SMP support.
In conclusion: use whatever you want or need and don’t flame people because their choices are different
This article touched on most everything. I’d say the #1 reason Linux “pulled ahead” of 386BSD and the other Net/2 derivatives back in the day was the USL lawsuit. Linux seemed to start gaining its massive mindshare among the higher ed/research crowds at that time, especially at NASA. The USL lawsuit did irreparable damage to *BSD’s mindshare, as Linux mindshare was increasing exponentially that whole time.
The end of the article probably makes the best point of all:
Linux was able to take advantage of the corporate trend toward selling services because it was “in the right place at the right time,” he said.
Linux’s tremendous mindshare over *BSD isn’t due to its technical merits. It isn’t due to the GPL somehow being inherently superior to the BSD license. Linux was simply in the right place at the right time. Its tremendous mindshare was the reason it received corporate backing.
Mindshare is an excellent thing to leverage. Mindshare is the reason why OS/2 and DesqView, both DOS compatible and substantially superior from a technical, were unable to gain enough mindshare from DOS/Windows 3.1 before the release of Windows 95 to prevent their eventual decline into obscurity. Mindshare is the reason Apple is not dead yet.
Without mindshare, FreeBSD’s role in business will be relegated to backoffice servers… but that isn’t necessarily a bad thing. It’s more or less where it belongs.
Hmmm.. To me the USL lawsuit seems more like a cleansing, because all non-free files were identified and rewritten.
So now with *BSD everyone has a well-documented consistent, very stable, clean, efficient system that allows for any commercial project you can dream of, with that license.
(though I don’t believe any of the SCO license problems with Linux, either ;-))
That said, I like the GPL for fairness value, and I see that Linux currently is a bit more bleeding edge and enjoys much more commercial support (like drivers, even though they aren’t free sometimes).
FreeBSD is keeping up fast, though, and seems to gain many new users.
I was introduced to Nix through OsX. My daughter baby-sat for a software engineer who was doing magic with Linux. What a great problem to have. BSD or Linux?….and I get to play!
The permaculture that is nix makes me smile.
As to why some corporate players are so enamored of Linux, Murphy cited IBM as a case in point.
… IBM <*is* using NetBSD as their OS of choice for some of their more powerful terminals — but never openly admit they do.
Well, I use both, FreeBSD and Linux (in my case, Slackware) and have always experienced difficulties to choose one of them.
Some times for reasons that are not easy to explain I pick one and know that the other is equally capable for the job.
Here, in OS news, one time I blamed Linux as being more lenient, to not enforce standards, thinking that FreeBSD was more System V compliant and focused. If you have to keep your application source for both, you probably had swallowed some distasteful pills. But by now I know I was wrong in many and important points.
Still, I would like to see a more cooperative attitude from the developers of both projects, establishing a standard that can be accomplished in a near future (file system layout, system calls, etc.). That should be awesome.
Hey guys,
By the way, Gentoo has started *BSD* ports. So soon, you’ll have a stage1 tarball to install Gentoo/OpenBSD or Gentoo/FreeBSD, with portage/emerge in OpenBSD. It’s really in the early stage, it’s been announced this week or last week in the GWN. But it’s something that devs are working on right now. So you will have the the best of both worlds in one. I just can’t wait for that.
I am a linux user myself, but have played around a bit with FreeBSD earlier as a server os. (where i didn’t do the install myself)
I have seen lots of people so far praising *BSD over linux, but i have yet to read just ONE good reason as to why this is so. (no, “it is more system v like” or “i like the file system layout better” does not count IMO)
From what i have read it (=freebsd) seems to be harder to install and setup than linux, and i can’t see how it can be easier to use once setup, considering it uses the same software for most of the things. It does have the ports system, which many people seem to praise, but i just can’t see how that is superior to tools like urpmi or apt-get (or graphical frontends to those)
So can anyone actually point at something that makes running bsd much better than running linux?
Troels: When people say easier to use, they should probably say easier to manage. For sys admins, I find BSD to be much simpler to manage. One example, you probably have seen the rc scripts of RedHat, which are a frigging mess. Now look at /etc/rc.conf in openbsd, it’s really clean and straight forward, you want to disable something, you change SENDMAIL = YES to SENDMAIL = NO in /etc/rc.conf. try to do that with redhat. Sure there’s the nifty windows-like GUI program to do it, which in the end does everything for you. but it just bites. you have less control over what you can do with it.
also, when you use rpm-based distro, you have to use the options the packager decided were best for you. you have no idea what was compiled in. with BSD, you can change that, as well as in Gentoo, which is a distro that really looks like BSD.
Those and others, are examples of things that make it easier to manage/run/use a *BSD box. but that’s just my opinion, i don’t mind reading a lot of docs before being able to tweak the .conf text file. Some just might prefer the RedHat/SuSE/Mandrake-gui like style of configuring things. you just use the one your prefer.
I agree.
The one you left out was, “BSD is a complete system, kernel and userland”.
I say so what, what real advantage is gained by this?
I also agree with, tech_user about “concrete technical performance differences”.
Kingston replied, Mostly because there are so many different configurations that would skew the results in favor of either system. It would have to be a *massively* standardized suite of tests to be sure. That is an excellent comment. Which led me to think why don’t we make FreeBSD the baseline to compare distro’s such as Redhat, UnitedLinux Flavour, Mandrake, etc.
In those performance reviews of a standardised test could be done when a newer version of a specific distro is released. These reviews can say how much faster or slower Distro A is than FreeBSD. Once many distro have been reviewed in such a way we can see if there are variation between distro’s.
I realise this is a big task, and won’t be done because no one could care less. I guess only enterprise customers really care enough.
“From what i have read it (=freebsd) seems to be harder to install and setup than linux, and i can’t see how it can be easier to use once setup, considering it uses the same software for most of the things.”
I have found that FreeBSD is easier to set up than any version of Linux I’ve tried, and I do install things often enough to know what I’m talking about here. FreeBSD’s installer ‘sysinstall’ is small, fast, flexible and a real joy to use. Debian’s installer (ugh), and Red Hat’s Annaconda just don’t compare.
Let’s not even get into RPMs vs FreeBSD packages. Debian’s packages have some good things over FreeBSD (binary upgrades come to mind), and falls short in other areas.
“It does have the ports system, which many people seem to praise, but i just can’t see how that is superior to tools like urpmi or apt-get (or graphical frontends to those)”
Like I said above, apt-get has some handy functionality that FreeBSD packages currently lack, or implement poorly. FreeBSD’s packages are integrated into the ports ports collection however, in a way that while I believe is possible with Debian, is not a fun way to spend an afternoon. I dare you to try to upgrade all of Debian’s packages from source, and then to do the same from FreeBSD and tell me with a straight face that Debian is better there
“So can anyone actually point at something that makes running bsd much better than running linux?”
I can only name a few things that I have experienced personally. I’ve had fewer stability problems on BSD than on Linux, on a wide range of hardware. (not all of it mine sadly, but at least I get to play with things
I’ve not yet had major security issues (yet) with BSD (excepting a horribly broken installation of FreeBSD 5.0-CURRENT a while back, totally my fault
, where I’ve had my various Linux boxes rooted more than once, despite my best precautions. The BSDs are also IMO more well laid out, thought out, and implemented than is Linux and GNU. Once you get the basic idea behind how things are done in BSD (any variant) you don’t have to go searching for things or spend long trying to figure things out.
BSD documentation is also orders of magnitude better than anything that has ever come from the FSF. Like the software, it’s well organized, and easy to browse through.
Just some things to ponder, but as always, it’s best to try things for yourself, and use either what you like, or what works for you. For me, BSD does both.
The *BSDs and Linux should exist together – eventually, the day will come when you can have your favourite distro’s tools – e.g. gentoo/portage etc.as another poster mentioned sitting on top of a FreeBSD, GNU Hurd or Linux kernel, on an x86/x86-64 or PPC/POWERx machine, and not worry about the difference.
You can choose the kernel, or your systems vendor can choose it for you. Everyone is more or less working towards providing a POSIX/UNIX userspace with the details of the kernel largely hidden.
I believe it is vital to have these projects run in parallel. Competition along with cross-pollination is practically guaranteed by the licenses used in these projects, and the pace at which they are evolving seems to be scaring the hell out of the commercial software industry who still can’t figure out why this is even happening to them.
I’ve learnt so much using Linux/*BSD,UNIX and others, found a great set of communities, and I am a better person for adopting free software licensing for my own work and applications.
I have been given the freedom to try out different options and make my own decision on which one fits my needs, and I hope everyone will always have these choices.
Fact, I’ve used both FreeBSD and Linux. The last Linux I used was Mandrake 8.1 But I’ve used FreeBSD since 1.1 and Linux slackware 1.2.
Both have served me well but I always went with FreeBSD because it worked better for me.
The thing that got me to side with FreeBSD was when I was trying Redhat 5.1 and I ran into the glibc problem which need to be used for this or that rpm etc. It was a head ache..
And when it came to multitasking I felt Freebsd worked better. Time and again on fairly sturdy machines, if I were building a Linux Kernel that was about all I could do. Don’t worry about playing an MP3 or doing any editing work with emacs etc. It would jump and pause way too much compared to using FreeBSD.
Even this I feel was something to do with building a proper custome kernel to optimise the thing. But when I went to IRC to ask about it. I got answers from people running the Linux room that kind of gave me pause.
‘Why would you want to do something when your building a kernel?’ or ‘You shouldn’t be touching your machine while it’s building a kernel it takes way too much cpu time’ etc. I kid you not this was almost verbatum the responses I got.
Mind you I’ve had unpleasant replies from FreeBSD folks too. but I’m not complaining about the people part. Just the way I experienced the OS’s.
Same machine, build a kernel with make -j 4 and I’m still playing mp3’s just fine and I get a kernel in 5 minutes. I go to /.’d servers and find they run Linux most of the time. So your milage may vary and I am glad if you like Linux.
I have to admit, I prefer OpenBSD by far to FreeBSD. But one major GOOD, even VERY good point that FreeBSD has. It’s the FreeBSD handbook. Which can be consulted even if you use Linux, because it will point out at least the logical steps to achieve something. Kudos to that.
From what I’ve seen also, *BSD tests more their code before releasing it. This is why they still ship for example, apache 1.3.28 instead of 2.x. Which makes it more secure, since they audit it a lot more. Also, seems like anyone can submit patches for Linux. To submit a patch to *BSD (at least OpenBSD), it will be audited by other devs as well. If your patch is accepted, you might be baby-sitted by other devs for 1-2 years before they let you submit other code by yourself. All those things put together, makes better OSes I guess.
I guess, the real argument in your summary is more of about the personal prefer because you only talk about how easier to use Linux is than BSD… 🙂
BSD is always easier than Linux for me. But, that’s not only reason why I love BSD. It’s more than just the personal prefer, so the other replies already have cover mostly. 😉
I’m using OpenBSD now, and I stated before that it is in many ways nicer than my native FreeBSD. It was basically a ten minute install from start to finnish, with no need to load modules for sound cards or tweaking config files for my mouse wheel to work in X. Sweet.
You are right about the FreeBSD handbook though, and it was one of the very first things that caught my attention when I was struggling with Linux (what seems like ages ago). OpenBSD’s FAQ isn’t too shabby itself, but it does need a little work.
As far as the Apache thing though (if you are refering to it’s status on BSD right now), I think that it is only a non 2.x version on OpenBSD, and it has something to do with the Apache team making some unfair assumptions about kernel threads, and not anything to do with your earlier argument about the BSD folks testing code more thoroughly than others (which I do believe myself for what it’s worth, but that’s OT).
no … i really don’t like this. its the one thing i really don’t like about the BSDs … and i think Slackware does this too.
its fine if yu only ever have named and sendmail. its nice and simple and cruft-free to edit in or comment out “SENDMAIL = yes”.
but this is not scaleable. is there is definitions file anywhere that tells me the options? does “SENDMAIL” do anything? what if i put in “POSTFIX = yes”? or “MYDAEMON = bananas”? see? soon it loses its appeal. anyway, what is actually run when the BSD encounters “SENDMAIL = xyz”? i’m sure i can find out – but its not immediately obvious – and it never was by design – you were only ever meant to change “YES” to “NO”.
having a directory structure like /etc/rc.d/, the contents of which are scanned at start up and the contents of which are executed makes the system more general, more scaleable/entendible, and in my opion, simpler. i think its called “Sys V” startup.
the issue of redhat having horrible scripts is a separate issue. i know mandrake’s scripts are terrible.
i prefer “correctness” and “elegence” over such hacks as /rc.conf. and i wish NetBSD would adopt this more scalable bootup style – that would be elegence heaven!
I’m not very familiar with the BSD License but with the GPL any opensource developer retains the copyright to his/her code.
Does the BSD Style license allow for similar protections? if not then maybe this is part of the issue..
***
Linux’s tremendous mindshare over *BSD isn’t due to its technical merits. It isn’t due to the GPL somehow being inherently superior to the BSD license.
***
I’ll take issue with two of your points.
One could certainly argue that in the areas that Linux and BSD can be compared, BSD is more mature. However there are many areas (such as SMP, filesystems, userland) where the BSDs are embarassingly behind or hopelessly dependent on GPL’d Linux efforts.
As for the GPL vs. BSD license, the “all forks eventually return” nature of the GPL puts much more effort behind the main thrust. The various bitchy BSD schisms keep all their efforts weaker than they need to be.
“I’m not very familiar with the BSD License but with the GPL any opensource developer retains the copyright to his/her code. Does the BSD Style license allow for similar protections? if not then maybe this is part of the issue.. ”
Copyright is copyright. The ownership of the code in question still lays with the copyright-holder.
I haven’t used FreeBSD or NetBSD alot, and FreeBSD very much recently, but I have been watching them a bit. I remember one of the arguments that used to surface as an argument for Linux is that drivers for new hardware usually made it into the Linux kernel quicker than the BSD dists.
Is/was this ever true? Or is there pretty good cross porting of drivers across them?
How is the OpenBSD and the FreeBSD installers nowadays? Do they do a pretty good job of autodetecting hardware on newer systems? SATA support? AGP? Video Cards? Etc…
I’m honestly not trolling here, just curious, as I mostly use Linux myself ;-).
” I’ll take issue with two of your points.”
Well, three actually…
“However there are many areas (such as SMP, filesystems, userland) where the BSDs are embarassingly behind or hopelessly dependent on GPL’d Linux efforts.”
Sigh. Another person who posts before trying something.
“As for the GPL vs. BSD license, the “all forks eventually return” nature of the GPL puts much more effort behind the main thrust. The various bitchy BSD schisms keep all their efforts weaker than they need to be.”
BSD licensed code is freely shared, just like GPL’d code. BSD code has fewer restrictions as well, not preventing people from doing as they will, as long as the license’s terms are obeyed. (Basicall you can’t sue the writer if it breaks, and you can’t claim it as your own work).
“The various bitchy BSD schisms keep all their efforts weaker than they need to be”
Not so. Those are no more prevalent today among the BSD groups than in the various Linux groups. In an evolutionary sense, the BSDs are far more sucessful as BSD code can be found in many more places than can Linux code. The diversity is staggering. There is BSD code *in* Linux, as well as Windows, Mac OS, and every commercial UNIX you can think to name.
We can talk some more once Linux has been around for some twenty+ years
Troels: I have seen lots of people so far praising *BSD over linux, but i have yet to read just ONE good reason as to why this is so. (no, “it is more system v like” or “i like the file system layout better” does not count IMO)
Documentation is perhaps the primary reason. Not only does FreeBSD have extensive online documentation of all aspects of the system, but the entire system is in itself self-documenting thanks to an extensive amount of comments present in every script.
Hmmm.. To me the USL lawsuit seems more like a cleansing, because all non-free files were identified and rewritten.
The remaining AT&T code left in the 4.4BSD Lite and Net/2 forks was extremely minimal. The actual removal of AT&T owned (and therefore legacy) code occured long before the USL lawsuit with the creation of BSD Lite.
The GPL is the greatest strenght and the greatest weakness of the “Linux development model”. The same can be said about *BSD, except it’s a different license.
one reason to use BSD over linux is better, more stable and consistent IPSEC and IKE daemon support. the isakmpd from OpenBSD (works on netbsd, and some people did some work porting it to freebsd, i did some work for a company on this once). at that stage it was far superior to linux 2.4 and the freeswan add-on. there was certainly no support onl inux for non-static ipsec connections (aggressive mode, with username associating to networks). that hopefully may change with kernel 2.6
One could certainly argue that in the areas that Linux and BSD can be compared, BSD is more mature. However there are many areas (such as SMP
Yes, FreeBSD is “behind” Linux in terms of SMP. This comes with a number of caveats such as FreeBSD is moving to libkse/thr for multithreading (and several applications can already be built with libkse/thr) and also that Linux is aiming for big iron scalability, whereas FreeBSD is optimizing for smaller (ie. 1-2 processor) systems.
filesystems
Absolutely not. If you believe this is the case you are gravely mistaken.
UFS2 is a 64-bit extents-based filesystem. It uses soft updates to maintain filesystem consistency in the event of a power outage. Soft updates have an extremely positive effect on filesystem performance… by caching metadata operations it is possible for large metadata intensive operations (such as recursive removal of a directory) to be carried out instantaneously from the user’s perspective. Journaling, on the other hand, has negative performance characteristics.
That isn’t to say that Linux doesn’t also have 64-bit extents based filesystems, however one of these, XFS, had to be hacked on in horrible ways (by building their own page buffer directly into the VMM) due to Linux’s discontiguous VFS.
From a purely technical perspective, it would’ve made much more sense to port XFS to FreeBSD, as FreeBSD’s VFS is much more similar to Irix than Linux’s is.
userland
The GNU userland is the single most abominable part of Linux. glibc has scrapped its heap implementation 3 times trying to improve heap performance on Linux, but unfortunately since the Linux VMM is a moving target they still haven’t produced a decent implementation. Linus has ended up in arguments with the glibc developers as people submit patches to work around glibc bugs that Linus felt should instead be fixed in glibc.
glibc is constantly making changes which break backwards compatibility, and worse is doing as little as possible to adhere to de facto standards, instead choosing to follow the letter of the law of written standards, even when the wording of these standards is somewhat vague. Most notable in my mind was the decision to implement fpos_t as a structure, which, while POSIX compliant by the letter of the standard, detracts from the de facto integer implementation used on virtually every other platform.
The userspace applications of FreeBSD was selected by Apple for use in conjunction with MacOS X. I think that should say something about their quality. If you have any specific examples of how the FreeBSD userspace applications are “embarrasingly behind” their GNU counterparts I would like to know.
where the BSDs are embarassingly behind or hopelessly dependent on GPL’d Linux efforts.
In none of these areas is BSD dependent on GPL code. The only notable GPL program the BSDs are dependent on is the gcc toolchain.
It happened in 1992 and was resolved in February 1994 (http://freeunix.dyndns.org:8088/site/bsd.shtml). Linux at the time of the resolution of the lawsuit was at version 1.0. What are you all talking about? Real multiplatform development started long after that and the kernel never really became a serious contender until version 2.0. BSD derivates had their chance to dominate the market. However, companies are reluctant to invest, in contrast to the Linux kernel. The big difference between the two is obviously the licence. No company in their right mind is going to invest in a kernel and user space tools (GNU project) that can become proprietary at any time. No IBM or SGI are going to donate a journaling system if some other company can immediately usurp the effort and add it in its own products without giving something back. Don’t delude yourselves… it is the licence that counts…
Come on, say it with me… *FREE* BSD. You can’t rightly have a proprietary product with *FREE* in the name now can you?!
Seriously though
damn I say this alot) does the fact that Microsoft can take FreeBSD source code and make a proprietary product make FreeBSD any less free? Think hard about that one. BSD source is everywhere, it’s free, and it’s not going away. Apple has done so, and guess what, there’s still our good old friend FreeBSD, and it’s still free.
Oh, and code has in fact flowed back into the BSDs. For a quick example, check out the math libraries.
<quote>
From what i have read it (=freebsd) seems to be harder to install and setup than linux, and i can’t see how it can be easier to use once setup, considering it uses the same software for most of the things. It does have the ports system, which many people seem to praise, but i just can’t see how that is superior to tools like urpmi or apt-get (or graphical frontends to those)
So can anyone actually point at something that makes running bsd much better than running linux?
</quote>
BSD is much easier to keep updated once you’ve configured cvsup and make.conf. I can update the entire base OS by typing “rebuild”, which I’ve aliased to
“cd /usr/src && make update && make world && make kernel && mergemaster”. On my dual Xeon it’s done in 30 minutes and I can use the system the entire time. It’s so painless I could do it daily if I wanted to be anal retentive.
I can upgrade any port simply by typing “portupgrade -rR gnome2” or whichever port I want. Building any port via ports is much easier than compiling on Linux since the ports system handles dependancies for you (debian notwithstanding). apt-get gives me a binary compiled to someone else’s desires. My system is custom tuned to my cpu and each port is customized the way I want. I haven’t actually done an Install upgrade on my FreeBSD 4.8-stable box since 4.2 was released.
How does glibc becoming more posix compliant spell doom to the GNU tools or the Linux kernel in particular. Also, sometimes breaking compatibility is necessary and better especially if in means more POSIX/UNIX complaince. Even gcc has broken compatibility with earlier versions. The resultant effect of that wasn’t too catastrophic. Backward compatibility sometimes restricts innovation and advancement, the Windows 9x series anyone.
I tend to agree with AC that the GNU tools are is an order of magnitude better and many instances more mature than the FreeBSD userland tools. To compare the GNU tools and the FreeBSD userland tools is a joke, in my opinion. Performance wise, there’s no visible difference between FreeBSD and Linux. They are both solid and evolving in different ways.
Finally, MAC OS and other commercial ventures who choose to use the FreeBSD kernel did so not because it is better as you seem to infer, but because the BSD license enables those commercial ventures to rip off the efforts of the open software without given back to the community or even releasing their closed source code that was once open. So, their choice of FreeBSD isn’t based on performance or whatever you seem to allude, but on the fact that the BSD licenses encourages commercial entities to sell open source code and closed/proprietary source code, ala MAC OSX.
Regards,
Mystilleef
=) I’ve done this. I set up a FreeBSD 4.x-stable box that ran cvsup and rebuilt the entire system (didn’t install it though) over and over again and walked away for nearly a week. Near constant load and it didn’t crash. It was very impressive.
The command:
“cd /usr/src && make update && make world && make kernel && mergemaster”
does the following:
1. fetches all the latest source for whichever branch I want.
2. Rebuilds all the system binaries (ls, bind, vi, etc)
3. Builds a new kernel
4. Updates/merges changes in all configuration files.
gnu_command –too-much-typing-for-my-liking.
Gah. GNU long options. Man I hate those… *terrible* for scripting.
“Finally, MAC OS and other commercial ventures who choose to use the FreeBSD kernel did so not because it is better as you seem to infer, but because the BSD license enables those commercial ventures to rip off the efforts of the open software without given back to the community or even releasing their closed source code that was once open. ”
Sorry old chap, but it was actually for *both* reasons, the permissive license, and the technical maturity.
http://developer.apple.com/darwin/projects/darwin/faq.html
“BSD is widely respected as clean, robust, and maintainable code. There remains a strong developer community that knows the code base very well and continues the work started at UC Berkeley. In addition, the BSD license is very open, which has made it easy for us to leverage its compelling core technology to enhance the Mac OS.”
Don’t bash each other as you are from the same UNIX family
We are actually on the same page. FreeBSD is as good as Linux and vice-versa. However, I’m positive Apple decided to choose BSD over other alternatives, for cost/benefit reasons which heavily favored by the BSD license, rather that technical merits. Corporate entities hardly give any crap if code is clean or not, whatever that means.
Business Mentality-101
“Hey, this is free code, people say it’s solid, robust and clean. Oh, and we could take it and make it closed source and make a fortune out of it. Also, for our marketing plot, we’ll tell them its a clean, robust, secure and maintainable FreeBSD code. Hmmm…I can hear the registers ringing. *Corporate executives grin*”
As a Slackware user I do not understand what you are talking about when you say.
its fine if yu only ever have named and sendmail. its nice and simple and cruft-free to edit in or comment out “SENDMAIL = yes”.
but this is not scaleable. is there is definitions file anywhere that tells me the options? does “SENDMAIL” do anything? what if i put in “POSTFIX = yes”? or “MYDAEMON = bananas”? see? soon it loses its appeal. anyway, what is actually run when the BSD encounters “SENDMAIL = xyz”? i’m sure i can find out – but its not immediately obvious – and it never was by design – you were only ever meant to change “YES” to “NO”.
Slackware does not work like that.
Well designed.
The BSD code is easy to read, and is easilly modified. It is well documented, and is very portable. As it evolves, fewer changes need to be made as it was well designed in the first place, often needed mere tuning rather than the wholesale removal of subsystems as is all too common in Linux.
BSD has been scrutinized for decades, and is therefore far more mature than is Linux, and more security holes have been eliminated. Linux is on the bleeding edge and quite often it shows. I can remember when Linux 2.4 came out, and the fury of swapping that constantly occured. It was madness. How did that ever get into a mainstream release?
I’ve never seen or even heard of such a thing in BSD. I’m sure it has happened, but it was so long ago that I was half my current height. There are things I like about Linux. It has its charm. I just don’t like having to spend so much time caring for a system I can’t depend on to be up and running and ready to go, when I have more important things to do.
ummm…..
u do know that the bsd in mac os x comes from nextstep and thus back from late 1980s…. when apple acquired nextstep for the os, it released the core technology under darwin name and has since then put even apple developers as the freebsd code cvs access….
they hired john hubbard (well known freebsd hacker; served on their 9 member council, i think)… don’t think that just because apple is a commercial company that it is not ready contribute back to opensource…. (hello, rendevouz, khtml, etc, anyone??? they very actively involved in opensource, even on xfree devel list… of course they have their own interests, but that doesnt mean that they havent helped out opensource any less)
<< hey, i already posted all this apple sh*t in another thread, i am not doing it again… >>
of course, when nextstep was seeking a unix kernel, the bsd liscence must have struck out at them…. and i see nothing wrong with that…. if i was the head of a commercial company, i would think the same way too…. that isn’t a weakness of bsd, but one of its strength…
now, IMO, the gpl is the reason for linux’s success…. i mean, if a company did not have to submit code, would it really want to? the gpl forces people to do so…. the viral nature of the gpl forces companies to work straight and pay attention to what gets released….
linux hype is also another one of the reasons for its success…. companies want to implement linux for its name…. its a viscious cycle in which companies hear of linux hype, adopt it, thus paving the way for others to adopt because some company adopted too…
bsd license is much, much more free…. when people say freedom of software, u cant get much free than the bsd license…. to the argument that the developer wants his software to always remain free… well, it is free even if some company decides to make the code properitary…. it is still available under the bsd license…. sure the enhancements wont get open…. but thats their anyways, isnt it??? i mean, THEY enhanced… they just give u credit for giving them the original code…. so u already got your credit…. Whats the problem?!
now, what has caught my admiration is that the bsds continue to persevere and release excellent and innovative code without the hype… (((i mean, the ip stacks and other code that helped to propagate the internet is just one example; so bsd does and will continue to impact our daily lives))) and maybe thats how it should be…. maybe it should be in its own little world that provides for the rest of the world…. i mean bsd code is HEAVILY found in linux, apple, (rumored) windows, UNIX System V, etc…
let it continue then… i m happy with the bsds at their current positions…. the tight knit community is very friendly; i just love it…. there is a bsd out there for every job and everyone (net for portability and clean architecture that i dare to challenge anyone to beat; open for robustness and secuirity features that shame other unices; free for its rapid development, innovative path (well that one is for all the bsds), and general purpose strength in server and desktop)….
meanwhile, linux has by far the best hardware support of any unix… it is the most widely used and has gained many developers….
still, i dont think that the bsds are going anywhere in their importace and absolute novelty and innovation….
frankly… it shouldnt be any other way…
Two minor corrections, Jordan Hubbard, not John.
“meanwhile, linux has by far the best hardware support of any unix… it is the most widely used and has gained many developers…. ”
More drivers does *not* equal better hardware support!
I’ll take *better* drivers any day.
The BSD code is easy to read, and is easilly modified. It is well documented, and is very portable. As it evolves, fewer changes need to be made as it was well designed in the first place, often needed mere tuning rather than the wholesale removal of subsystems as is all too common in Linux.
Yes, the word clean has often confused me as it’s usage differs amongst individuals. Any proficient C programmers can read any Kind of C written code. How one C coded algorithm is easier to read than orders remains a myth to me. The BSD code has been constantly evolved for 30 years.
Have a look at the Kernel-2.6 code for current changes to the Linux Kernel architecture. You don’t need to blowup the whole kernel to make changes in it in the current incarnation of the current kernel. So your argument is somewhat outdated.
BSD has been scrutinized for decades, and is therefore far more mature than is Linux, and more security holes have been eliminated. Linux is on the bleeding edge and quite often it shows. I can remember when Linux 2.4 came out, and the fury of swapping that constantly occured. It was madness. How did that ever get into a mainstream release
Don’t you think it’s a little unfair to compare a kernel that has been constantly hammered over 30 years with one that is a little over a decade old? In my opinion, comparing Linux with the BSDs is not only embarrassing to BSD, it further shows the strenght of Linux, the underdog OS only 5 years ago. Besides, a lot more individuals scrutinize the Linux kernel than the FreeBSD.
It only makes sense for more exploits to be detected in the Linux kernel. Do you really think if FreeBSD had as wide a userbase as Linux does, more security exploits wouldn’t be discovered in FreeBSD? I think not.
I’ve never seen or even heard of such a thing in BSD. I’m sure it has happened, but it was so long ago that I was half my current height. There are things I like about Linux. It has its charm. I just don’t like having to spend so much time caring for a system I can’t depend on to be up and running and ready to go, when I have more important things to do.
To make a statement like the above is premature. We are supposed to have a paranoid mentality not a confident mentality with respect to security or stability or whatever. Any system can be hacked. No OS secure. Yes, *BSDs included. The only secure system in the one no one can reach and is not connected to a network.
Most importantly, an OS is only as stable and secure as the skills and knowledge of the person administring them. Linux is nowhere as mature/old as BSD is, but it is definitely comparable and usable and any of BSDs are. Perhaps, this is what has impressed the many Linux enthusiast. Or perhaps not.
Regarding
Mystilleef
Economically speaking, operating systems are suspectible to natural monopolies. An operating system with larger market share creates more revenues, so more money can be spent on its development. Besides, larger developer share means more outside contributions to the operating system. Compared to ordinary software, operating systems are especially suspectible to monopolies because the value of an OS strongly depends on the programs available, and incentives to develop a program for an operating system are proportional to the market share.
The relative success of Linux as compared to BSD as well as the continued dominance of Microsoft Windows is due to the natural monopoly effect.
However, I believe that monopolies on operating systems are harmful for innovation, so I would not want an OS, even if it is linux, to monopolize the market.
thank u for the corrections… im just sleepy is all…
“Yes, the word clean has often confused me as it’s usage differs amongst individuals. Any proficient C programmers can read any Kind of C written code. How one C coded algorithm is easier to read than orders remains a myth to me. The BSD code has been constantly evolved for 30 years. ”
Clean code can be readilly understood by non-C programmers. That’s a powerful distinction in my book.
“Have a look at the Kernel-2.6 code for current changes to the Linux Kernel architecture. You don’t need to blowup the whole kernel to make changes in it in the current incarnation of the current kernel. So your argument is somewhat outdated. ”
I have seen it, I have used it. There are improvements yes, but it is still a mess.
“Don’t you think it’s a little unfair to compare a kernel that has been constantly hammered over 30 years with one that is a little over a decade old?”
Get with the program dude, that’s what both you and I are doing.
“In my opinion, comparing Linux with the BSDs is not only embarrassing to BSD, it further shows the strenght of Linux, the underdog OS only 5 years ago.”
When I first got into Linux (about five years ago in fact), I suffered for it’s immaturity. When I use Linux now, I suffer the same way. Linux is a dream for people who like to fix things. Despite my love of playing with these things, I do on occasion need a mature and robust system.
I hardly see cause for embarrasment for the BSDs, or those who develop them. Who pioneered Open Source? Who laid much of the groundwork for the formation of what would be the Internet? Virtual memory? Portability? Oh, right, your mistake. BSD.
“Besides, a lot more individuals scrutinize the Linux kernel than the FreeBSD.”
Considdering the remote possibility that this point is true, you are missing the point here. It doesn’t matter if every pair of eyeballs in the world take a gander at the Linux source code at any given point in time, as it is constantly being re-written. Continuity and evolution are not something that anyone I know associate with Linux. Revolution and discontinuous development better describe the chaos from which Linux springs. Few things can be thoroughly tested, because nothing stays around long enough for the scrutiny to matter.
“It only makes sense for more exploits to be detected in the Linux kernel.”
Yup.
“Do you really think if FreeBSD had as wide a userbase as Linux does, more security exploits wouldn’t be discovered in FreeBSD?”
Different development models, more thoroughly tested code, evolutionary advancements? I think that FreeBSD would still come out on top.
“To make a statement like the above is premature.”
Uhm, no. The FreeBSD VM has existed for years. Nothing as bad as what happened with the early 2.4 Linux kernels would *ever* make it into a production release.
“We are supposed to have a paranoid mentality not a confident mentality with respect to security or stability or whatever. Any system can be hacked. No OS secure. Yes, *BSDs included.”
I am paranoid, and I do tend to think in worst case scenarios. I also study, read, and research, and experiment. The relative immaturity and flakyness of Linux compared to BSD is more and more apparent. I am well aware that no software is perfect, but I also know that some software is better (quality wise) than others. I’d take BSD over Linux any day.
“The only secure system in the one no one can reach and is not connected to a network.”
Not entirely true, but I’ll let you have that one.
“Most importantly, an OS is only as stable and secure as the skills and knowledge of the person administring them.”
That my friend, is only part of the picture. Poorly written and buggy code works it’s magic independantly of a users (lack of) ineptitude. You are too hard on admins and users, and not hard enough on the developers if you can say what you just typed with a straight face.
“Linux is nowhere as mature/old as BSD is, but it is definitely comparable and usable”
For many tasks yes, but I’d be damned if I were to run a buiness based on it.
“Perhaps, this is what has impressed the many Linux enthusiast.”
Perhaps. But sadly it seems that most only ever hear one side of the story… yours.
“Or perhaps not.”
I have never tried BSD I have used SCO which sucked bad but Linux has been a bliss. I for one am never going back to windoze after using linux and I dont see BSD on the plate for me ut wont mind trying OSX
ok, what does SCO have to do anything with anything in this discussion…
we are talking about bsd and linux; not unixware, bsd, and linux….
============
<quote>
I have never tried BSD I have used SCO which sucked bad but Linux has been a bliss.
</quote>
============
i mean, bsd and sco have nothing to do with each other unless u consider that probably **some** of unixware still might contain bsd code <probably the best parts of the code ;-)>
p.s. good to know that u wanna try os x 🙂 but wait until panther comes out and then try…
“bsd license is much, much more free…. when people say freedom of software, u cant get much free than the bsd license…”
If you take the view that a license that has fewer restrictions is more “free” than a license with more restrictions, then yes, BSD is “more free” than the GPL. That seems like a simplistic view though. What sense does it make to grant a “freedom” to take away other peoples freedom? Does “freedom” to own slaves really seem like freedom?
Clean code can be readilly understood by non-C programmers. That’s a powerful distinction in my book.
Your book will be any interesting read.
I have seen it, I have used it. There are improvements yes, but it is still a mess.
A mess? Anything above hundreds of thousands of line of code can easily be categorized as a mess. Being a little more specific wouldn’t hurt.
When I first got into Linux (about five years ago in fact), I suffered for it’s immaturity. When I use Linux now, I suffer the same way. Linux is a dream for people who like to fix things. Despite my love of playing with these things, I do on occasion need a mature and robust system.
Have you looked into Lindows, lately? All the window managers I know of have come a long way since the hay days. To claim the Linux has remained the same for the last five years, is not only ridiculous, it is false. And what are fixing? The kernel? the window manager? drivers? the gnu tools? Please, let’s get past these fanboy statements.
I hardly see cause for embarrasment for the BSDs, or those who develop them. Who pioneered Open Source? Who laid much of the groundwork for the formation of what would be the Internet? Virtual memory? Portability? Oh, right, your mistake. BSD.
Yes, only for an underdog to catch up and takeover.
Considdering the remote possibility that this point is true, you are missing the point here. It doesn’t matter if every pair of eyeballs in the world take a gander at the Linux source code at any given point in time, as it is constantly being re-written. Continuity and evolution are not something that anyone I know associate with Linux. Revolution and discontinuous development better describe the chaos from which Linux springs. Few things can be thoroughly tested, because nothing stays around long enough for the scrutiny to matter.
Yeap, change is the only thing constant in life. Change or stay in limbo. That’s the reason Linux is pioneering the open source movement. If there weren’t technological revolutions as you would put it, then will still be using abaccuses today. Change, or continue to play catch up.
Different development models, more thoroughly tested code, evolutionary advancements? I think that FreeBSD would still come out on top.
Yes, because FreeBSD has the best development model. *sacarsm* Can anyone hear fanboyism here?
Uhm, no. The FreeBSD VM has existed for years. Nothing as bad as what happened with the early 2.4 Linux kernels would *ever* make it into a production release.
Technically, FreeBSD didn’t have a VM for years, it was ported from its forefathers. And similar catastrophes have occurred in the existence of BSD do your research.
That my friend, is only part of the picture. Poorly written and buggy code works it’s magic independantly of a users (lack of) ineptitude. You are too hard on admins and users, and not hard enough on the developers if you can say what you just typed with a straight face.
So are you alluding the Linux kernel is poorly written and buggy and written by incompetent developers, or not as competent BSD developers? You are getting more interesting.
For many tasks yes, but I’d be damned if I were to run a buiness based on it.
Yeah, say that to IBM. Don’t you think you are underestimating Linux? Linux can handle any task FreeBSD can. No, we really don’t need to argue about this. This is so 5 years agoish. Go figure it out from supercomputers to watches. From the Athlon64 to the your calculator, Linux is everywhere dude and it is preffered. That says a lot, doesn’t it. Now, how many watches do you know that run on FreeBSD? Even TIVO uses Linux in it’s hardware. If FreeBSD were that flexible, you could have hope of it’s prevalance in more businesses than Linux already commands.
Perhaps. But sadly it seems that most only ever hear one side of the story… yours.
I have no story. Only that FreeBSD sucks as much Linux does. You can deal with it, or leave in denial. Good day.
Regards,
Mystilleef
First of all, I should point out that I enjoy BSD and Linux equally, for different reasons, and appreciate their differences as ‘character’ and not ‘failings’. Or, to put it another way — neither is perfect, but they’re both pretty damn good and getting better every day.
Kingston:
“The various bitchy BSD schisms keep all their efforts weaker than they need to be”
Not so.
http://www.netsys.com/netbsd-port-sparc/1995/05/msg00090.html
http://www.monkey.org/openbsd/archive/misc/0207/msg01843.html
Of note are the following facts:
1) 33.33% of the major BSD branches (Free, Net, Open) were created when an immature windbag used the word ‘cock’.
2) The best packet filter I’ve ever had the pleasure of using (pf) was born out of stereotypical BSD-style politics. I haven’t decided yet whether this is good or bad.
“The only secure system in the one no one can reach and is not connected to a network.”
Not entirely true, but I’ll let you have that one.
You were making a fair bit of sense until you said this. The statement to which you were responding verily IS entirely true.
🙂
bsd license is much, much more free…. when people say freedom of software, u cant get much free than the bsd license….
I think it’s more accurate to say that the BSD license is less restrictive, whereas the GPL licence is more free. Tomayto tomahtah.
<quote>
I think it’s more accurate to say that the BSD license is less restrictive, whereas the GPL licence is more free.
</quote>
yea, my bad; like i said, i have been sleep-deprived…. i shoulda said that bsd license is less restrictive…
HOWEVER, i don’t understand how the GPL is more free. If developers are forced to release any changes out to the public because of the license, then it is definitely not free for the developer. He/she does not have the freedom to choose what to do with their modifications… how is that more free than bsd license where u have the proper freedom of choice to do whatever you want with your modifications?
i think i am a little OT so i’ll stop their…
and btw, what does “Not as many BSD zealots as I expected” mean? 😀
lol, from my personal experience, ive seen linux zealots to be the greater of the two evils… god knows with their world dominations schemes and all 😀
anywho, i dont claim loyalties to anyside… i just appreciate systems for what they are (except licenses, they bug me, but if the system is good i am willing to toss that aside)… and i am quick to support the other side if the arguments seems to me to be getting one-sided…
a good debate is good for everyone’s soul 🙂
Linux is cool, but it has some annoying bugs that are so dangerous in a 365x24x7 server. From March to date I had to update the kernel 2 times because two bugs in netfilter.
I think that there is much Linux hype around. Cool? yes. The better? nope. If you want to have a secure, ever up server, perfectly tuned, you should try *BSD.
Thanks for the argument, it’s been insightful and fun
This picture was taken a few days ago from a campus computer. I went to check my mail and found that I was having trouble. It would seem most of the load is coming from the VM as there just wasn’t enough ram. However, the system never crashed.
This system runs a 2.4 kernel and I think it shows just how far linux has come. In fact, I watched the load go up as far as 222, it may have gone further but I was busy talking to someone about the cause of this load.
While I believe this shows that Linux has has come a long way both in terms of stability and robustness, I think the VM could’ve been a bit more graceful, but in spite of that, the system handled the load pretty well. I almost want to put a 2.6 test kernel on here just to see what happens.
This is not a quad cpu system, but rather dual Xeon’s with HT.
http://audioseek.net/~hiryu/huge_load.png
PS: Love FreeBSD too but I’m sick of unfounded bias.
Ignore the title of the window, that top is for audioseek.net, NOT panzer.audioseek.net which is my personal NetBSD alpha system on my personal DSL.
Wow. At least Mystilleef (IP: 66.71.219.—) has a clue.
“Of note are the following facts:”
Your statements are hardly factual.
“1) 33.33% of the major BSD branches (Free, Net, Open) were created when an immature windbag used the word ‘cock’.”
Uhm, no. The first two of the major branches (NetBSD and the FreeBSD) diverged from 386BSD because 386BSD’s author would no longer develop it. The FreeBSD group were more patient than the NetBSD folks, and so waited longer to see if 386BSD was indeed a dead project.
OpenBSD spit from NetBSD initially due to differences of opinion, but that has no bearing on sharing between the two today. Get with the program
“2) The best packet filter I’ve ever had the pleasure of using (pf) was born out of stereotypical BSD-style politics. I haven’t decided yet whether this is good or bad.”
Er, no again. The packet filter that was included with OpenBSD before was not freely modifyable, and this is not an acceptable fact for a part of a free software project. Politics had nothing to do with it.
“{The only secure system in the one no one can reach and is not connected to a network.”}
Not entirely true, but I’ll let you have that one.
You were making a fair bit of sense until you said this. The statement to which you were responding verily IS entirely true. ”
I’ll admit, I was tired by this point and I seriously misread the first part of the comment that I was responding to there. A system not on a network can still be attacked however (local users), and that was my line of thinking. I’m the first person to admit that I am not always on the ball at that time of night.
“bsd license is much, much more free…. when people say freedom of software, u cant get much free than the bsd license…. I think it’s more accurate to say that the BSD license is less restrictive, whereas the GPL licence is more free. Tomayto tomahtah.”
When you really think about it, “less restrictive” means “more free”. Anyone is “more free” to do as they wish with the code.
Just a few more things,
“Technically, FreeBSD didn’t have a VM for years, it was ported from its forefathers. And similar catastrophes have occurred in the existence of BSD do your research.”
Technically, when FreeBSD came into being, the copy of the 4.4BSD (Mach) VM became the FreeBSD VM. Your staement makes no sense.
I do in fact do my research, and not too many days go by where I am not reading the code of the modern BSDs, Linux and ancient UNIX. It’s an odd hobby of mine.
“So are you alluding the Linux kernel is poorly written and buggy and written by incompetent developers, or not as competent BSD developers? You are getting more interesting.”
Not at all. My only claim in this regard is that the development model employed is flawed, and that the architecture of Linux could use more planning. The coders on both sides of the fence are top notch.
“I have no story. Only that FreeBSD sucks as much Linux does. You can deal with it, or leave in denial.”
Everything has its advantages and its disadvantages. From what I’ve seen, BSD has fewer disadvantages than does Linux. If you are happy with Linux, then I am happy for you. I could never be, and I am not in denial.
But thanks again for the argument. You did have a few really good points in there, and even though we don’t see eye to eye, I do in fact appreciate your input.
I can’t help but notice that a lot of people here are obviously not talking from personal knowledge or experience, but are most likely merely passing on the usual memes that are thrown around in these discussions. It’s the exact same phrases that are used over and over again, discussion after discussion.
I’ve grown tired of these ‘discussions’ myself.
simpler to manage. One example, you probably have seen the rc scripts of RedHat, which are a frigging mess. Now look at /etc/rc.conf in openbsd, it’s really clean and straight forward, you want to disable something, you change SENDMAIL = YES to SENDMAIL = NO in /etc/rc.conf. try to do that with redhat.
Ahem:
# /sbin/chkconfig sendmail on
# /sbin/chkconfig sendmail off
# /sbin/chkconfig –list (show all services and runlevels)
# /sbin/chkconfig –level 235 sendmail on (override default)
This Redhatism is actually one of the features I wish was moved over to Debian. It makes things sooo much easier.
also, when you use rpm-based distro, you have to use the options the packager decided were best for you. you have no idea what was compiled in. with BSD, you can change that, as well as in Gentoo, which is a distro that really looks like BSD.
I’ve lived through a libc update, several broken gcc updates, a binary format change and some significant RAM defects that caused serious issues when compiling my own binaries. The last thing I need is to do is go back to the hellish nightmare that is QA’ing your own binaries.
That being said, recompiling all of Redhat to optimize for a platform isn’t really all that hard. It is just time consuming. The same is true for Debian (big easier actually with apt).
Indeed a great argument Kingston. 🙂 They are both great operating systems. And I do agree they both have their advantages and disadvantages. I don’t think Linux is better than FreeBSD or vice-versa.
I only wish more hardware vendors and software developers have the Unices in mind when they develop there software. That way individuals will be at liberty to choose whatever operating system they enjoy using the most.
Nevertheless, an excellent argument.
Regards,
Mystilleef