Slashdot Mirror
BSD For Linux Users
noackjr writes "Matt Fuller posted among his rants a great introduction and explanation of BSD For Linux Users: 'It's been my impression that the BSD communit{y,ies}, in general, understand Linux far better than the Linux communit{y,ies} understand BSD. I have a few theories on why that is, but that's not really relevant. I think a lot of Linux people get turned off BSD because they don't really understand how and why it's put together. Thus, this rant; as a BSD person, I want to try to explain how BSD works in a way that Linux people can absorb.'"
I'm sure Darl McBride is going to quote you quite liberally. Thank you very much for HARMING both communities.
--
*Art
It this an arrogant statement or what ? This sort of moderate arrogance, of showing off some alleged superiority, together with the constant efforts to undermine Linux from some BSD enthusiasts, don't really help me get interested on BSD.
But the reason I use Linux rather than BSD is simple: I personally prefer the GPL rather than the BSD License. Let alone the argument of which is better, I heard both sides several times, and I prefer the FSF's philosophy. This doesn't imply that I don't like the BSD philosphy. If the Linux kernel didn't exist, I would definitely be running KDE on to of some flavour of BSD ...
Why do I run BSD?
It Just Works. When I install the system, a set group of pieces are there. Period. When I go through the upgrade process, it works. When I install an addon package, it works.
It supports practically every piece of hardware I've thrown at it in close to a decade of use. And that support Just Works, it doesn't flake out or fail from one version to another.
I can choose whether to stick with tried and tested release versions, to track the -STABLE branch, to track the -CURRENT branch... however close to the edge I want to live. And even the bleedingest edge usually works just fine. I can take a system that hasn't been touched in 3 years, and bring it up to date with minimal pain.
When a problem comes up, whether it's security or stability or performance or even just aesthetics, it gets fixed. And because of the development methodology, I can get those fixes later that same evening. All the time. When a new feature gets added, I can grab it right now, not wait a week or a month or 6 months until the next release. I can see, whizzing by in my email, every change as it's made. Or, when something breaks, I can fix it myself, and be able to send the fix right off to somebody who can get it incorporated into the system right away. I've done it before, I'll do it again, and so have thousands of other people.
It's designed to all fit together. Not just munged after-the-fact to go together OK, but designed from the ground up to be a single coherent system. And when something doesn't fit, it's a bug to be fixed, not just "part of the game".
It's had quirks. It's had vagueries. It's had out-and-out bugs. It's had bizarreness. It's had inexplicable behavior. It's given me misleading error messages. Sometimes, it out-and-out breaks things that worked yesterday. And all those will happen again in the future. But, every time, somebody goes right ahead and notices, and discusses, and fixes it. I read the mailing lists, and I see these things go by. I see the problem brought up, I see the people who know the area discussing it, I see solutions proposed and discarded and tested and agreed upon, and I see them committed into the main repository. And usually by the time I see it happen, I can already get my hands on it.
It's just a level of consistency and transparency and usability lightyears beyond anything else I've seen. Its development path is aimed at technical merit and correct solutions, not just sexy features and workable hacks. And I think that's the only way to be sustainable long-term.
Why should you run BSD?
Well, I don't know. Does what you have now meet your needs? Then you probably don't have an urgent need to change anything.
There're plenty of essays and rants on there about why X works better than Y, for any values of X and Y. I don't want to write another one. If and when you do get time or inclination or resources to try BSD, I want this essay to help you understand why some of those differences are there, and what the result of those differences is.
More info
All the BSD's sites (FreeBSD, NetBSD, and OpenBSD) have extensive documentation available on their systems. Because the base system is integrated, all things dealing with it can be authoritatively documented in a single place. They describe the strengths and weaknesses of the individual systems, how to install them, how to upgrade them, how to admin them... Everything you need.
In the FreeBSD docs in particular, the article on Explaining BSD is similar to this essay, in that it describes BSD and its historical context. There's all kinds of documentation, from introductory documents describing the basics of FreeBSD and Unix-type systems, to detailed descriptions of the inner workings of the kernel, and everything in between. Scroll down the page.
And, of course, the FreeBSD Handbook, and the FreeBSD FAQ are both stuffed with information you never even knew you needed. Don't leave home without it.
The BSD's also all have mailing lists and news
...Linux isn't UNIX. It's an independent project, despite what SCO thinks.
I wonder how funny it would be to tell SCO that BSD is a direct derivative work from AT&T UNIX...
Do not look into laser with remaining eye.
Linus T. said one in a Cnet interview that Linux development was 'organic' not engineered.
Too bad the author didn't link to that cnet interview.
As you mentioned, Microsoft uses this code, too. The difference is that the motivation behind Microsoft's additions is based upon power and greed in a feverish attempt to maintain their monopoly.
If you work on BSD code, you could possibly be a contributor to OS X. Certainly there is a sense of pride in this.
However, as in Microsoft's case, your contributions can also be twisted in dark and selfish ways without your knowledge or control. In this case, one is not proud, but deeply saddened by the misappropriation of which you spoke.
"Condescending" is in the eye of the reader.
This week, I had a developer log a reply in our bug tracking system as "... maybe because I am a developer it is 'intuitive' to me".
This was in response to an issue raised on a pop-up calendar with no less than 9 functional elements and 2 invocations to change a date field on a web page.
Now that's condescending.
just get along?
I tried FreeBSD once, it seemed okay. I agree that the BSDs are probably a better, more reliable operating system than Linux. However, Linux was SO much better than Windows, that any improvements BSD could make seem minimal at best, especially since most of the server and end-user software is identical.
For example, someone once told me that BSD is much more stable than Linux. Assume that's true for a minute: I've NEVER had a Linux system crash, except for hardware failures, and I've been using Linux since 1998 or so. Okay, so maybe BSD is more stable than Linux, but the time spent learning BSD isn't worth the "extra" uptime I'd have, since the extra uptime is approximately zero.
As I mentioned, the end-user software is mostly the same, but I've heard Linux has more variety and more hardware support. My printer is supported under Linux (it's an HP USB printer), but I don't know whether HP is putting resources into BSD support. Same goes for my Palm Pilot. Since everything is working now, and I'm happy with how well everything works, why should I want to switch to BSD?
And why can't we have our own preferences anyway? Why do BSD users have to bash Linux and vice versa? The two are a lot more alike than different, at least compared to Linux vs Windows or BSD vs Windows. (It's almost like brothers fighting...)
--RJ "Firmly sitting on the fence"
"Because there's nothing on Linux that most Windows users think they need or want, yet how many people are standing on street corners praising Linux and all it's glory over Windows? More software is written for Windows, and there's alot more support available."
You're overlooking philosophical differences between Window and linux. Which is a lot more different the the philosophical differences between BSD and Linux.
Also momentum. Windows is loosing places to look for new users, Linux is still gaining them. Mean while BSD seems to be just kind of gliding along.
Could just be my impression, however I can't think of the last time I saw BSD mentioned in any trad mag. for any industry as an alternative to windows.
I want to make something perfectly clear:
This post is in no way anti BSD. I haven't used BSD in years.
The Kruger Dunning explains most post on
Below is from the SFU licenses you get with SFU 3.5
The utilities bc, ci, co, cpio, csplit, dc, diff, diff3, gawk, gzip, gunzip, ident, merge, nl, rcs, rcsdiff, rcsmerge and rlog are covered under the GNU General Public License, here reproduced. In accordance with section 3b of this license the source code to those utilities is available from the Services for UNIX World Wide Web site, http://www.microsoft.com/windows/sfu.
Help fight continental drift.
Most Linux users don't bother with BSD because of the license. At any time a corporation can come along, take the code, and use it in their closed source product. With the GPL the Microsofts of this world have a harder time stealing the code and reinforcing their monopoly.
If I worked at a large software corporation and was hellbent on shoring up my monopoly at least cost I would thank the lord that there are fools enough releasing software under the BSD license and prey that Linux and the GPL would just go away.
I don't mind contributing to Linux and GPL'd software for free but doing the same with BSD or closed source software is lunacy.
Release BSD under the GPL and I'll try it out, otherwise, forget it.
It's not technical. It's political.
5 And finally, the GPL would be democracy. Democracies are hardly ideal, they are slow, they waste a lot of energy, they infight a lot, but in the end there is NOTHING better.
Refresh my memory, how is the GPL not a "shared source" license again? How can I create a derivative work of GPLed code without sharing my code?
Oh, and you are mixing your metaphors. Anarchy, fascism and democracy are systems of government. Communism is a system of economics.
Serve Gonk.
I think with Linux's popularity, people often fail to understand that generally the same desktop environments, utilities, and such work exactly the same on the BSD's. Especially in the case of the ever-so-portable NetBSD, it allows one to have a Linux-like desktop on pretty any platform imaginable that's powerful enough.
Now, commercial/binary application support, and variety of device drivers available are the main areas were the BSD's are still playing catch-up.
Love many, trust a few, do harm to none.
On the "advise" (really more like repeated nagging, which seems to be fairly common from BSD enthusiasts) of a friend, I went with OpenBSD when setting up a server a year or two ago. Even after getting over the culture-shock, I ended up not liking it very much. I've gone back to debian, which I'm confident is every bit as secure, and is easier to admin.
/etc. Having to go searching for httpd.conf in /usr/local/etc or maybe /var/www/conf *really* bugs me.
I found the installer to be utterly cryptic, filled with unexplained single letter commands. I should not need a $40 reference book to install the freaking OS. Once it was up and running I was fairly comfortable, except for the propensity to scatter config files/etc around the filesystem. Maybe debian has spoiled me, but I expect to find *all* config files under
Then of course there is their reputation for security. They would have people believe that there has never been a root exploit; in reality, that only actually applies to the "base" system. I do like that the secure by default thing; having to specifically enable services is a good idea(getting the choice while installing is better though).
Which is ironic as current some engineering branches are trying to understand and duplicate efficient and fault-tolerant organic systems.
" the reliability of any given system is often largely a function of the person running it."
You're absolutely correct here. The most stable OS on Earth can be brought down by someone with enough access and a lack of knowledge. In the same fashion, I'm sure a freshman administrator could take an OpenBSD machine and end up with it being less secure than an unpatched Windows 95 box. The documentation for most OSS projects tends to be lacking, out of date, incomplete, and sometimes outright wrong. The reason for this is rather obvious - writing documentation is perhaps the only activity more boring than reading it. Thus, someone who's read all there is to read in the docs for a particular OSS project can still have little to no idea what they're doing with it. It's not anyone's fault, it's the simple matter of no one being paid to write good, strong documentation. That's one of those jobs that sucks so much that you've almost got to be paid more for doing it than for writing the actual software just to get motivated. Some in the elitest crowd may contend that the source code is, itself, all the necessary documentation.
To be perfectly honest, when I first started working with FreeBSD (I needed a reliable DHCP and tFTP server), I screwed it up so badly that it had to be completely wiped about once every few days. These days, I can sit here and laugh at some of the ridiculous things I was trying to do, because I know better from working my way through problems.
"the ones that matter have been "as reliable as running water" because I chose to keep them off the bleeding edge"
As I replied to another comment, I didn't mean to put down Linux at all. Linux is pretty reliable out of the box in most cases, and keeping yourself off the bleeding edge is a smart way to make sure things stay that way. My comments were centered more around the communities, the thinking, and the habits associated with FreeBSD and Linux. Although there are some instances where the OSs themselves evidence my point (as in firewire, CD/DVD burner support, etc), it's the users who really generate the mentality. As I said, there's nothing wrong with being on the bleeding edge if that's your style. For many like me, though, the running water is plenty.
Right now I use a combination of FreeBSD and Windows 2000 at home and at work for the reliability of one and the feature-richness of the other. If Linux gets to the point of Windows 2000's compatibility and usability (usability for me, not in general - I know it's usable and has been for quite some time), then I'll re-evaluate things. Ideally, I'd like to see FreeBSD make some progress towards desktop-readiness, as that would get me down to using one OS. I also recognize, however, that FreeBSD's intended market is now, and probably always will be, the server market.
-- "Government is the great fiction through which everybody endeavors to live at the expense of everybody else."
Are you aware of the 503 Service Unavailable status code? It comes complete with an optional Retry-After header.
Furthermore the license allows proprietary software to "steal" source code and use it. The combination of these problems leads to a somewhat inferior OS.
Now, Apache uses a BSD style license but they have an open development model which allows them to take advantage of a very large developer pool in order to stay ahead of their competition. In fact although proprietary versions of Apache exist which perform better than the official releases, SGI has put out some open source patches which generate even larger performance boosts. This is the reason why they have such a strong showing in terms of market share.
BSD once had potential but the procedural problems they are experiencing hurt it when it comes to the market. I suspect that this is probably in part because the BSD teams are not interested in such things, and that is a shame... In fact, although I labeled it as an inferior OS, this is not due to lack of progress within BSD -- it has been progressing somewhat, but rather because all the improvements they make tend to be quickly copied by their competitors AND they lack the developer pool to stay ahead of this game (a problem which does not exist in the Linux or Apache communities, though for somewhat different reasons).
I don't think that there is enough widespread support for BSD to save the operating system. What must be done is an opening up of the development process OR a GPL-style restriction on redistribution. In many ways I favor the former.
Even in a worst case scenario, I don't see BSD completely dying. I think the developers are less into competition and more into a sort of idealized cooperation. As a result, even if BSD becomes more marginalized, I don't think that it will die outright. It will most likely outlive Netware, for example.
Well, a bit of a history lesson for you. Once upon a time, a young reseacher at CERN named Tim Berners-Lee invented a nifty system for developing technical reports. HTML and HTTP was not all that brilliant. What was brilliant was the decision to release the protocols and the first barely mature reference implementations into the public domain.
Why was this brilliant? Because within a few years, dozens of folks were tripping over each other to produce better implementations based on the house that Tim built. Meanwhile, protocols that had languished under various copyright restrictions faded away and vanished. Without that intense competition to build better implementations of an open protocol (including proprietary implementations like MSIE and Opera) we might still be a fragmented internet community with some people stranded in vax notes, some people doing gopher, and some doing Lotus Notes.
I can say that for me, the single worst thing about the *BSDs is the way the kernel is bound into the base system.
My FreeBSD experience went roughly thus:
Maybe I was unlucky. But the lesson I learned from this is that when monolithic `designed' systems go wrong, you're screwed. Modular systems, by contrast, give you the option of swapping out faulty components, even if the resulting system is slightly less slick than a monolithic system.
I now use Debian Stable for high-availability systems, and Gentoo for high-performance systems. Gentoo is, of course, the modular system par excellence, giving you total control over all aspects of the packages you use. Debian is less so, but gives you the security of knowing that your system has been thoroughly tested under all the conditions it is likely to encounter, and that with a modicum of judgement, updating whenever new packages appear will give superlative security without compromising stability.
And, of course, if the latest kernel screws up your system, you can downgrade to the previous version with absolute freedom. To me that's worth far more than whatever an integrated system can provide.
Anyway, I consider myself somewhat of an advanced *nix user, even though I've only been doing it circa 2001. I started out with linux, of course, and liked it very much. I've seen the both the system and the open source community make tremendous leaps and bounds in just these few short years, and I must say that linux is absolutely a fine operating system.
Despite this, I was also compelled to try FreeBSD; I'm always interested in trying new things and tinkering around as much as I can with things I know little about, so naturally BSD was the perfect target for me. Being used to linux, I found the subtle differences in BSD very attractive; I wanted to get in on this because of the useful knowledge would provide as well as the rich history behind it. Also, a tiny bit of zealously may have played a part ("Wow, it's really unix!) ;)
I had a touch and go realtionship with BSD for the past several years, but it wasn't only until this one that I got a little more serious about it. Putting linux down for a while, I delved deeper into the use and maintanence of the system, gaining the kind of more intimate knowledge that the casual user probably doesn't have.
What happened was that I loved absolutely everything about it; it was very different than linux! Instead of groping around for strange tools and trying to fight a quirky system with which, due to constant mutation, there was always some uncertainty as to the correct way of doing things (are you running distribution X, Y, or Z? What version of software package A do you have, and are you using libraries B or C?).
With BSD, I found a system that was consistent, reliable, and very practical in its design. If you wanted to do something, there was a right way to do it, and this way always works (and works well, at that). Package management was top notch, and upgrading, maintaining, and generally managing the system was a quick, easy, and painless process. It is often said that BSD is not "user friendly", but I guess it's true when they say unix is "picky about who its friends are." To a moderately advanced user like me, BSD was more friendly and easy to use than linux! There aren't as many programs that will hold your wee-wee for you so you don't miss, but all of the information you need to know in order to do a certain thing is always readily available and right up front, beckoning for you to do it yourself and do it right. (some linux distributions are much better in this respect than others)
It must be said, however, that linux is a very fine system in its own right, and not to be looked down upon in any way! As Matthew says in his article, the big differences are mainly philosophical and cultural, and I have just found myself very much at home in this culture and very agreeable towards this philosophy.
What puzzles me is all of the rampant slandering of BSD that goes on by linux users (only a small percentage of this is meant to counteract the slandering made against linux by BSD users). Yes, there are common myths (like the ones Windows users hold towards linux, which make us all cringe), and yes, there are many misconceptions afloat.
Regardless of personal preference, the fact remains that BSD is also a very fine operating system, and does a few things a lot better than linux. Of course, linux does a few things a lot better than BSD, so when choosing between the two it's important to learn what those things are and how important they are to you.
I don't need to achieve "world domination" in order to get good software. No matter what Microsoft does with it's copy of FreeBSD, my copy is still intact, along with the copies of those who develop it. It would change nothing in my life.
If a hardware manufacturer would give back his changes to you, that would change your life (a little bit).
Take another example. "Alice in Wonderland" is in the public domain. It is not copylefted. So along comes Microsoft and exploits it. They change the wording, the plot, dumb down the puns, etc. Then they release it under a proprietary license, charging a fee to read it (not merely a fee to purchase the paper it's printed on). Further, let's assume that their version of AiW becomes more popular than the original. Does this change anything? No, not really. It's sad that so many people are choosing to read the bastardized version rather than the original, but the original is still there and unchanged. There are no restrictions in place anywhere that prevent people from accessing it. Anyone who wants to can go get it and read it. It's still in the public domain.
I think I understand your point very well. This is what I meant with "island". Consider this: it's not a big difference if you read the 1865 version of Alice or the 1965 version. But software is constantly evolving. Your piece of public domain software will be worthless 10 years in the future - if that's all you got, then you got nothing, but the maintainer of the popular proprietary branch got everything - mostly from you.
These arguments aren't new, and I don't think that we will discover anything new if we go on. In my opinion we don't have to "resolve" this conflict. Though I count myself to the GPL and the Linux camp, I also use BSD and am glad that there is this other implementation of UNIX with its different license. I wouldn't want BSD to change. I want these two distinct communities to stay and to grow, because I don't like single points of failure.
Um, except for that logical flaw in that software and ideas are non-rivalous resources (unlike the dodo and the bald eagle btw.) No matter how much an idea is exploited, you can never run out of it. There is no need to "defend" something that is limitless.