Reasons Why You Should Not Use FreeBSD

[LightBit]

I've been using Windows and Office on 8Mb... the only question is that it has been, er, SLOW. ZFS is an enterprise monster, it's design isn't a speed devil on it's own. So without huge RAM it would work but would be fairly slow.


Yeah, as for ZFS you're right but I seen dozen of complaints about it's stability though. So as for me I could use btrfs too, but I'm positive that it's devs don't call things stable when they are actually not, unlike FBSD guys who seems to be overoptimistic (because of lack any other alternatives?).

FreeBSD recomends 1GB RAM for ZFS.

Granted that nobody bothered to really rededign ancient on-disk structures and take into account modern state of art (unlike EXT4 did) it's still a piece of ancient stuff on steroids. At least, even ext guys finally got idea that extents are good idea and make things faster. Still not a case with UFS. Sure, this requires on-disk format overhaul. Something that FBSD guys failed to do.

UFS2 also supports extents.

[Uqbar]

Ah!

- Exchange (MAPI) and its groupware functionality missing.
- Telephony (ISDN to SIP gateways, Asterisk etc) missing.
- FreeBSD in the mobile space isn't too good.
- No Gluster file-system support.
- It doesn't yet support the latest graphics drivers (except for the NVIDIA binary driver, but the ATI and Intel drivers for FreeBSD are ages behind).
- The lack of suspend-and-resume for laptops.
- Incomplete ISDN support.
- Broken ports packages.
- It's easier to run Mac OS X than FreeBSD.

Apart the "broken ports", I would say the other points are "pros" to me to use AnyBSD!
I would not trash an OS because it doesn't have (YET) userland stuff to handle MAPI, ISDN or any other non-kernel feature!

[Korla Plankton]

At work I use a linux workstation but I spend most of my time in console on FreeBSD 8.1 servers. They are both fine. The FreeBSD servers are more stable, but if they weren't that'd be very strange since they're dedicated servers.

I could set up a perfectly functional FreeBSD workstation but why would I? It takes about 100 fewer steps to set up a Linux workstation.

I could set up a perfectly functional, secure Linux server but why would I? It takes about 100 fewer steps to set up a jailed server with FreeBSD.

BSD or GPL? C'mon, the differences are obvious. Any developer can pick the appropriate one for their project. Arguing about which one is more 'free' is ridiculous. No amount of ideological rhetoric will change the dictionary -- the freedom to not publish your modifications is obviously just that: a freedom. Whether it restricts other's freedoms down the line doesn't even enter in to it because people are 'free' not to use those restricted versions.

A problem should only have to be solved once. Solving a problem over and over is a waste of resources. That is the reason for the BSD license, so that someone with a problem can get the code needed to solve that problem. If a license restricts my usage of the code, as the GPL potentially does, then I may have to solve that problem again by writing my own code. That means the GPL'd software isn't getting used and it means there's another, perhaps incompatible or not-as-good solution to the problem out there.

Microsoft solved its TCP stack problem thanks to the BSD license.

Apple solved its crappy OS problem thanks to the BSD license.

These were real problems and they were solved because the BSD license makes the solutions available with no strings attached (other than maintaining copyright.)

I'm not saying that this is 'right' or 'better' than the GPL -- the licenses serve different purposes. GPL's goal is to maintain a free software ecosystem, solving problems is a side benefit (and a necessity, or else why use it?) BSD's goal is to solve problems and having a free software ecosystem is a side benefit (and a necessity, or else nothing would ever get developed.)

Of course, this is all just my opinion (and based on BSD 'philosophy' material I've read, way back when) but it sure makes sense to me!

[XorEaxEax]

A problem should only have to be solved once. Solving a problem over and over is a waste of resources. That is the reason for the BSD license, so that someone with a problem can get the code needed to solve that problem.

I generally agree with you but it's not as clear cut as you make it out to be. If a company (or individual) picks up BSD licenced code, fixes or enhances that code and doesn't return these changes then you will again have to 'solve the problem over and over' if other people would want to have those fixes/enhancements (i.e duplication of effort).

GPL 'fixes' this by demanding that any source code modifications must be made available along a distributed binary, but of course this demand in turn makes the code in question unsuitable for proprietary projects or projects which will not accept the GPL licence for other reasons (like it's viral nature).

Bottom line, there is no perfect licence (which in turn is why we have so many licences out there), I will agree though that if you release a piece of 'perfect' code that no one will have any reason to fix/enhance and/or you have no interest in any enhancements made to it by others then it makes perfect sense to release it under as permissive conditions as possible.

[RealNC]

My reason for disliking the BSD license is more selfish. If someone wants to base something on code I wrote and then make that proprietary, I want a cut. You want to use my work for proprietary software? Pay me and I'll grant you a proprietary license. If not, make your work fully open or gtfo.

People who work for IT companies and are already getting paid to write BSD code obviously don't care about this. Individuals who aren't getting paid to begin with should care.

[XorEaxEax]

If someone wants to base something on code I wrote and then make that proprietary, I want a cut. You want to use my work for proprietary software? Pay me and I'll grant you a proprietary license.

It's certainly a viable opportunity if your code is of interest for commercial proprietary ventures. The x264 devs are making good money this way, it's open sourced under GPL which means anyone can use it and also incorporate it into their open source projects, however if you want to use it in a proprietary project you can purchase a special licence from the x264 devs which allows this.

[Korla Plankton]

I generally agree with you but it's not as clear cut as you make it out to be. If a company (or individual) picks up BSD licenced code, fixes or enhances that code and doesn't return these changes then you will again have to 'solve the problem over and over' if other people would want to have those fixes/enhancements (i.e duplication of effort).

I don't disagree with you but it really is as clear cut as I made it out to be. It's the ramifications that aren't necessarily obvious.

What you describe is a non-open project adding a great new feature to solve some problem. They solved it for their commercial product but since it's such a great feature that everybody wants, its functionality eventually appears in a open project and the problem has been 'solved.' This is true whether you are talking about BSD or GPL (with the former fixing the problem for everybody and the latter fixing the problem for anyone who complies with the GPL.) So while the problem is solved more than once, the number of solutions is minimal and, once it is in the open, it doesn't need to be solved again. Unless it is only available under GPL and you are allergic to GPL for whatever reason.

Bottom line, there is no perfect licence (which in turn is why we have so many licences out there), I will agree though that if you release a piece of 'perfect' code that no one will have any reason to fix/enhance and/or you have no interest in any enhancements made to it by others then it makes perfect sense to release it under as permissive conditions as possible.

Again, no real disagreement here, though I don't think it's so much about the code being 'perfect,' it's about it successfully solving a problem, which means it gets used. If there is desire for improvement, the community will improve it. Companies may hire someone to improve it then keep those improvements to themselves but that doesn't prevent the community from improving it as well. That's a duplication of effort, sure, but a company will do what a company will do. It would make sense for that company to switch to the community-driven version once it is suitable for their purposes so they don't have to keep paying this developer to maintain their out-of-tree hackery.

Now one might say GPL avoids this by forcing the issue, but such companies wouldn't touch GPL software anyway so it's kind of a moot point. Also, just because a company is using BSD code doesn't mean they aren't giving back.

Anyway, we can get all esoteric and stuff but my main point was summed up in my previous post. That is to say, the BSD license and the GPL license are trying to achieve different goals. There's room for both. On that (and on most other things) I think we can agree

[Korla Plankton]

My reason for disliking the BSD license is more selfish. If someone wants to base something on code I wrote and then make that proprietary, I want a cut. You want to use my work for proprietary software? Pay me and I'll grant you a proprietary license. If not, make your work fully open or gtfo.

People who work for IT companies and are already getting paid to write BSD code obviously don't care about this. Individuals who aren't getting paid to begin with should care.

You went from talking about yourself to talking about 'individuals.' Sounds like you wanna restrict my freedoms!!!!111!!1!

PS in case it isn't abundantly clear, i'm joking. but i do think that if an individual wants to release under BSD, there's nothing wrong with that.

[0xBADCODE]

@0xBADCODE → Well, you're just yet another troll? ... I'm just going to do leave some "notes" not because of you (you just basically trimmed my message and answered what came out of your ass), but for the casual reader that may read your misinformation

And you're just another fanboy who tells that black is white because it makes their fetish to look better. Smells like double standards. And btw I have to trim things, else message would be uncontrollably huge.

This is where you show your lack of knowledge about the AT&T litigation that damaged the credibility of the BSDs for a long time,

Ahh, all fanboys are same. They prefer to see only one side of coin. You see, you've completely ignored that SCO has filed ton of lawsits against Linux, don't you? So according to your theory, Linux has been in disadvantage for a while too. Which hasn't prevented it from outrunning BSDs. And for being arrogant fanboy I would backstab you with your own weapon. You see, your theory assumes one fatal flaw: in real world nobody needs weaklings. So if some lawsuit could seriously harm some OS I simply would not use it anywhere in more or less critical places. Do you seriously want to tell us *BSDs are toy OSes which can be knocked off by single lawsuit so we're better not to use them for serious things? Or you've just admit you haven't noticed your logic is flawed? Please explain: why the heck just a single lawsuit should ruin the party for whole years? And what I'm expected to to next time some nut would file another lawsuit? You see, I cant afford to shut down all servers, abandon desktop and wait for year or two to see how it's resolved.

also it's a little bit of what XorEaxEax and blacknova said.Now, just to repeat: No, a license that allows privative forks and someoe exercises that, that's not an abuse ... a GPL'ed software like the ffmpeg case I mentioned, well ... that's proper abuse

Formally that's right. But as you see, in real world these "proper abuses" had their costs. Every time someone (ab)uses their right to keep things closed, mainline project loses to its competitors.

Now, you can jump to google and pretend that you already knew it.

That's a way too fat trolling and will not work: you see, I wrote "(ab)use" in the way which clearly indicates I knew this at the time of writing. Try harder next time and don't make stupid logic flaws

WTF? have you ever writed something like specfiles AAND a port? because I can ASSURE you, that this is plain fucking simple (it's a meta port example, but it's an example of syntax): http://www.freebsddiary.org/meta-ports.php and with that, not only the port, the port system can generate the package for you.

Can't you FBSD guys understand that for me OS isn't a fetish? And even if some technical operations are simple, they usually not related to my goals, targets and tasks. So it's just a SYSTEM MAINTENANCE OVERHEAD. Which takes my (limited) resources away from more useful goals. Which is bad. So I want to avoid it as much as possible. I'm not inherently against of doing something useful for opensource projects, but it have to lead to my goals rather than be some stupid technical overhead. The less overhead the easier my life.

I know distros with horrible repos that can't hold a candle to what FreeBSD offers (and I'm not thinking about *indie* distros)

You're right that some Linux distros could be even worse than FBSD. The "only" problem is that in no way I'm not forced to use them and I can choose what matches best with my goals and preferences. Sure, it's not easy for FBSD fans to understand that due to lack of distros. They have only one true way of doing things, everyone who disagree is hereic and should be trolled and fried. That's what I dislike in FBSD community, BTW.

So when it's linux it's ok, and when it's FreeBSD it's not?

You see, in no way I'm forced to stick to worst, least convenient Linux distros (that's where your logic meets EPIC FAIL). So I can just choose another distro and rock'n'roll. In FBSD you do not even have this option as there is no virtually "other distros". No, seriously, how you could even get idea I would use worst distros? I would use those which fit me best.

As for indies, Debian or Centos do not really report to corporations. They are running on their own, powered by their communities. So I guess they qualify as "indies", aren't they?

well, surely you are a popular person among the users of distros like Slackware or Arch. lol

I don't really like those distros. Mostly for the same reasons, BTW. They seems to offer relatively high amount of administrative overhead. Fortunately it's not like FBSD where you have to eat the sh*t at any cost. In Linux world I can just go away and select another distro which fits better. Simple.

Then you have pkg_add to install the binary packages that the FreeBSD offers to you, of course, you'll loose some customization, but you don't seem to care about that ... in any case, you can mix binary packages AND ports

As for me I like when all parts of system are declared as some packages and package manager undertakes package management hassle "by default". Something that FBSD fails to do for some poiltical or historic reasons. Btw, while this benefit haves huge advantages there are some disadvantages as well buy you seems to be completely unaware of them. This indicates you've never attempted to compare various approaches at all, sticking to your FBSD oldschool/barbarian approach.

And this is where your lack of knowledge shines.

Possibly. FBSD has been one of 1st system I used. It has been ages ago. I could forget something. I was not eager to learn too much about FBSD after figuring out Linux performs better for me as well. So I rather invested time and efforts in learning Linux and now I'm properly rewarded for my efforts.

Corrupt packages, package manager database and I assure you that you'll be S.O.L in a heatbeat mostly because linux puts everything in /usr (and with the merge, now even more) compared to base system of FreeBSD being in that folder and *everything else* being STRICTLY on /usr/local ...

You see, your theory is cool but in real world I only had 3 serious failures of package management systems in 8 years. That's per dozens of servers, etc. And package managers even told be how to recover. So it could be fixed in a matter of minutes. In fact, package manager is a very good tool to check and recover system most of times. It could validate file hashes to detect corrupt files (how you're doing that on FBSD?). It can be requested to reinstall packages to replace damaged files. However it system is very seriously damaged or suspected in security breach it could be faster and smarter just reinstall it or recover from backup rather than anything else.

so you can just purge package database, rm -rf that and have a virgin OS.

And who told you that your "virgin OS" is anyhow useful for me at all? On the other hand it is easy to instruct package manager to adjust selections to what I want, so I would have what I want in a matter of minutes. You see, I can have somewhat different view on what I consider "core part of my system" than FBSD guys do. In Linux it's not that hard to tweak package manager to do this for you, and all things are done in unified and universal way.

I wanna see you trying to fix that in linux WITHOUT leftovers. because yes, I CAN ASSURE YOU that what I mentioned leaves ... NO LEFTOVERS

If I would continue this strange logic a bit further, power facility could fail as well. Let's ignore electricity all together? Sure, you can even intentionally short circuit some wires to cause outage! Does this counds as electricity disadvantage?

[QUOTE]1) You can install any gcc version you like, by now, that is[QUOTE]
I can but it's utter stupidity to offer ancient stuff by default. I consider this as absolutely dumbest way to do license politics and get own leg shot as the result. Good luck with this approach.

You don't (you don't have to do WHAT I like to do).So don't blame your ignorance on others.

In theory FBSD is good OS. But when it comes to the practice it's either inconvenient as a h*ll or half-working or whatever else instead of "just works". You see, it lacks universal mechanics to do all this for all parts of system and installed software in uniform ways. So from my standpoint, dealing with all these quirk takes far more management hassle than it's desirable. I don't see a valid reason why some "core" part of system should be considered "more important" than things I installed. I never install OS just to exists and do nothing. Why this "base system" should exist? I don't need half of things inside and need dozen of things outside. Why I should be unable to manage all this in uniform way? Just because of some odd politics/relics? Nope, thanks.

and the system was of a comparable quality to what GCC4.2 produced.

The "only" problem is that rest of world uses GCC 4.6/4.7 which are superior to ancient 4.2. For some reason FBSD guys fail to understand that world haven't frozen when they decide to stick with 4.2. So I don't really care about comparison with 4.2. If you want a comparison, come on. Let's beat 4.6/4.7, okay? Not to mention GCC supports far more architectures I need and it's convenient to use the same set of compilers for different targets (less administrative/technical hassles again).

Ports and binary packages DOES have dependency information. And you have "metapackages".

And again, in theory you're 100% right. On practice all this just awkward, feature incomplete and isn't anyhow uniform for system and software. Does not stands comparison with more decent approaches.

Well, shame on your lazy you !!! you could have installed gcc on a newer version, lol.

You see, basically I want to have working system with minilal efforts. Having free BSDM session where I'm in passive role and system in active role isn't what I really wanted. It's so fortunate there are many OSes and distros. So I can just choose most convenient for certain task rather than fix all idiocies in the world myself.

It's on a port (x11/nvidia-driver) that downloads and install the privative driver ... as entertaining as any linux distribution.

The only "little" trouble on this way is that I have Intel graphics at notebook and AMD card on desktop. "Entertaining", eh?

As virtualization is not my area, I'll avoid that topic as much as I can ... when it comes to package management, well, it depends on your needs, it can do some really cute stuff that linux package managers can't (at least without some problems), but it also MAY add moar complexity to maintenance on certain cases. it depends on the scenario, and again : the right tool for the right job.

Okay, so Linux is a better tool for my jobs. And better for my hardware. Granted that my tasks are more or less common real world things, that's why I dare to share my views here.

I'll just ignore the unbearable and uninformed comments about ZFS and UFS2. I just can't find the patience to repeat myself and answer to what is clearly an uninformed aggression :S

Then maybe you could inform me what exactly "groups of cylinders" supposed to mean on SSD? Or even on modern LBA-based HDDs? And don't you think that ext4 with it's extent allocator beats UFS in speed any time of the day?

[vertexSymphony]

Can't you FBSD guys

Sorry dear, just to clarify:

alex@Sylbit:~> uname -a
Linux Sylbit 3.4.0-26-desktop #1 SMP PREEMPT Sun May 27 19:46:37 UTC 2012 (8353c9e) i686 athlon i386 GNU/Linux

P.S → I just gave up on you, it's pointless ...