OpenBSD Fork Bitrig Announced

With the goal of bringing more experimental development to the OpenBSD code base, a few developers have announced a fork named Bitrig. According to their FAQ, Bitrig aims to build a small system targeting only modern hardware and "be a very commercially friendly code base by using non-viral licenses where possible." Their first step toward that goal was removing GCC in favor of LLVM/Clang. The project roadmap shows their future goals as adding FUSE support, improving multiprocessing, porting the system to ARM, and replacing the GNU C++ library with LLVM's.

Comments

wtf is a bitrig?

[noh8rz3]

sounds like a place to keep my bitcoins...

Re:wtf is a bitrig?

[Anonymous Coward]

A bitrig is 1/8 of a byterig.

Re:wtf is a bitrig?

[Taco Cowboy]

A bitrig is 1/8 of a byterig.

If that's the case ...

1 Bytecoin = 8 bitcoins
.

Re:wtf is a bitrig?

[Liinux]

So, bitcoin really means piece o' eight?

Yarr!

Re:Theo is going to me sooooo mad

[fph il quozientatore]

"commercially friendly"

It would be a good definition for a stripper.

Re:Theo is going to me sooooo mad

[networkBoy]

I don't think he will be mad about that. Mad about the devs leaving, sure, but not about the commercial fork. If they contribute back to the main trunk, then I think all is well.

Seriously, Theo may be a bit aggressive, but he's not an idiot, the BSD license allows this more clearly than anything else out there short of public domain.
-nB

i386

[TeknoHog]

Bitrig will only target actively developing hardware and architectures such as i386 and amd64

How the fsck is i386 actively developing?

Re:i386

[blane.bramble]

Bitrig will only target (actively developing hardware) and (architectures such as i386 and amd64)

Does that help you parse it?

Re:i386

[gman003]

"i386" is OpenBSD-speak for the architecture variously known as "x86", "x86-32", "i686", "IA-32", and "32-bit Intel". Just as "amd64" is OpenBSD-speak for the architecture known to others as "x64", "x86-64", "IA-32e", "64-bit Intel", "Intel 64", and whatever VIA calls it.

Re:i386

[mcgrew]

Double woiosh to you, sir. It wasn't a parsing problem, he was pointing out an incredible stupidity. Look at the number -- I haven't seen an i386 since the early nineties.

Re:i386

[TeknoHog]

I did parse "i386" as "32-bit x86, including i686". I haven't seen too much development on that in recent years -- can you even buy such a machine any more?

Re:i386

[AlphaWolf_HK]

As others have said, though I'll add a bit more depth, is that i386 is the catch all for anything x86, with the exception of ensuring that it distinguishes from the 286 and below. The 386 was a major step up from the 286 and below due not only to being 32-bit, but also allowing protected mode and virtual mode operations, in addition to paging.

Virtually no modern software is adaptable to a 286 processor, whereas nearly all of them are adaptable to a 386, hence the common usage of "i386". As a matter of fact, intel actually didn't stop producing the 386 until around 2007. It was still widely used for embedded applications long after it was already obsolete.

Re:i386

[ChunderDownunder]

You'll probably find a 32 bit Atom netbook down at the department store.

Re:i386

[cthulhu11]

I had a phone conversation with a droid at Symantec last year, regarding the NBU that my employer insists on pissing money away on. After the one guy told me that I wasn't worth talking to since I don't have 4-figure signature authority, the other asked what platform we were using. I told them Solaris 10 x86 and he got all on my case about how I really needed to be using x64. If I'd been in the same room, I would have slapped both of them.

Re:Theo is going to me sooooo mad

[ArhcAngel]

If they contribute back to the main trunk, then I think all is well.

The double edged sword of the BSD License. I'm sure they will probably contribute back but unlike the GPL there is nothing legally to compel them to.

I wish them luck.

[Anonymous Coward]

This is a good "Put up or shut up" moment for BSD. For all the whining I hear about "Viral" and "Anti Business" licenses the various *BSD projects sure do have a meager adoption (Buisness, home, free or otherwise) compared to their GPL counterparts (Linux). I think an aggressive, forward looking BSD project would be great to have.

Granted, not all the most popular open source projects have "Viral" licenses (Eg - Most Apache foundation projects), but maybe.. Just maybe Linux's success is in part due to the GPL.

Some people feel the GPL is stealing something that they're somehow entitled too. In reality, it's more of an exchange. You give up the ability to have a certain business model, and in return you get the collective work of everyone else who's made the same agreement. You give up exclusive control of your source in return for a world-class, flexible, free, operating system with widespread uses. For free. With a BSD style license you're able to opt out of that "collective work" provision. You can take, but you don't have to give. As a result, the project does not grow.

It's probably in your long-term interest for the project to grow. I think the success of Linux proves this.

Re:I wish them luck.

[Anonymous Coward]

If you don't like it don't use the licence, and don't use GPL software. But don't whinge about not being able to rip it off for your own purposes.

Re:I wish them luck.

[Ichijo]

For all the whining I hear about "Viral" and "Anti Business" licenses the various *BSD projects sure do have a meager adoption (Buisness, home, free or otherwise) compared to their GPL counterparts (Linux).

Actually, OSX (Darwin BSD) is nearly twice as popular as Linux (9.0% vs. 4.9%).

Re:I wish them luck.

[serviscope_minor]

Actually, OSX (Darwin BSD) is nearly twice as popular as Linux

Actually, no it isn't. Unless you're restricting yourself away from servers and embedded devices. But why would you do that? This isn't a thread specifically about desktop operating systems.

Re:I wish them luck.

[101percent]

I haven't seen a troll like this in a long time. Cut and paste doesn't work? I spend the last 4 months of school doing vmware with RHEL and backtrack with a win7 host, and had no problems copying and pasting between any of these. Also, I just bought a 15" dell monitor off craigslist yesterday and running stock Debian 6 on a Thinkpad T60, plugged it in and it did mirroring 100% correctly. Tell MythTV users sound doesn't work reliably. Five guis are a mess? Do you listen to Tween Wave? If I apt-get source foobar I can build any debian package on my server and install it on my desktop. Documentation is worthless? The GNU Emacs Reference manual is like 1000+ pages. Any package which lacks a man page is a bug in Debian. Getting help online a problem? No one *owes* you help. Does your employer get free Windows tech support?

Re:I wish them luck.

[TheTrueScotsman]

It's "world-class" as a server OS.

Re:I wish them luck.

[serviscope_minor]

FreeBSD's SMP support is fucking incredible.

Is it any better than anything else? As far as I know, linux pretty much has the scalability record of 4096 CPUs in a single system image (SGI Altix).

Re:I wish them luck.

[killmenow]

in what sense do you consider Linux to be a success?

In the sense that it runs twice as many servers as Windows, roughly the same about of desktops as Macs (according to Steve Ballmer), and more mobile devices than any other OS in existence (where, btw, it is outpacing its rivals by a wide margin and now selling more units than desktop devices per year as nearly a million new linux-based (yes, Android is based on linux) mobile devices are activated every day. That sounds pretty successful. And it doesn't even include the embedded market, which you clearly know nothing about. So many embedded devices in use in many industries (the cable industry for instance) run Red Hat Linux and other distros.

By what metric is Linux "world-class"?

As of June 2010 the operating systems used on the world's top 500 supercomputers were: Linux 91.0%, Unix 4.4%, Hybrid Unix/Linux 3.4%, Windows HPC 1.0%, BSD 0.2%
That metric works for me. You apparently prefer ones with pretty pointy-clicky thingies.

Re:I wish them luck.

[rev0lt]

Cut and paste doesn't work?

It has been a while since I tried it, but with X/Xorg you can copy/paste text without an issue and you can copy/paste some kind of specific data (such as images) between applications that use the same toolkit. Everything else (formatted text, vector drawings, files) is a PITA.

Also, I just bought a 15" dell monitor off craigslist yesterday and running stock Debian 6 on a Thinkpad T60, plugged it in and it did mirroring 100% correctly.

So, just because it works for you, it must work for everybody? I had issues with older versions of Ubuntu on a Asus EE900, and with stock gnome you couldn't even click on the "ok" button of the display settings because it went offscreen (1024x600). It is somewhat easy to fix, but not very "user friendly".

Tell MythTV users sound doesn't work reliably.

The other day there was a rant one mile long regarding PulseAudio defects and how it killed low-latency audio in Linux. Many complainers were Linux users.

Five guis are a mess?

There is a consistent lack of consistency, yes. Partially because there is a multitude of toolkits available, and there is no single standard for almost everything. You gain choice, but you lose consistency.

If I apt-get source foobar I can build any debian package on my server and install it on my desktop

Yes, if they are roughly the same version. If they have eg. different glibc versions, you may have problems. And compilation flags in many linux distros can also interfere (try building an i7-optimized binary and run it on a p4), but that may also happen in other operating systems. AFAIK, both PC-BSD and MacOS have a somewhat elegant solution for it.

Documentation is worthless?

Yeah, it is. It is not that rare to find outdated manpages, features that lack documentation, or translation errors. You may not notice it, but if you are used to BSD-style manpages, you will have a shock.

The GNU Emacs Reference manual is like 1000+ pages.

Since Emacs isn't really part of the "base system", the manpages for Emacs are the same regardless of the operating system. I'm not familiarized with it, but size isn't really a quality metric.

Any package which lacks a man page is a bug in Debian.

Please have a look at the manpages of the base system of any BSD. Usually documentation maintainers are very strict, and "having a page" often isn't enough.

Getting help online a problem? No one *owes* you help.

I do prefer the BSD community (obviously), but we also have a bunch of arrogant bastards that will tell people to read the source. I'd say linux atracts a kind of "fanatism" BSD (usually) doesn't, but there are very nice and helpful people regardless of the operating system, so I wouln't expect getting online help for Linux or BSD to be a problem.

Does your employer get free Windows tech support?

Yeah, we actually do.

Re:I wish them luck.

[rev0lt]

While I'm a big fan/heavy user of FreeBSD, I've only seen reports of FreeBSD SMP machines upto 32 cores.

Re:I wish them luck.

[101percent]

I'm not a FreeBSD user, but I do know they spent some four maybe six years rewritting the kernel because they wanted better SMP. No, this does not make instantly "fucking incredible" like I said, but tracing the history back to grad students at Berkley decades following with some really smart developers, and following their release notes; no I haven't read the kernel source, but their major focus for half a decade was SMP support.

Re:I wish them luck.

[adri]

Sure, they set the record of 4096 operating CPUs, but the Linux kernel (much like the BSD kernel) has had large locks and serialisation points spread throughout the kernel for a loong time. ISTR that SGI Altix doing a computational workload (little kernel mischief). They got the CPU representation (ie, not a 32 bit bitmask) right and the scheduler working. But I bet you couldn't do 4096 CPUs worth of parallel disk or network IO.

FreeBSD, like Linux, has been recently focusing on modern IO and CPU intensive workloads, rather than the Altix "Demo" from years gone by. So ignore the silly SGI Altix datapoint and look at what people are using it for now.

(Yes, I'm a FreeBSD developer.)

Re:I wish them luck.

[TheRaven64]

A lot of FreeBSD code is shipping on every iOS, OS X, or Android device. Android's stack is mostly Apache and BSD licensed, with only a tiny bit of GPL. So, yes, just like that.

Re:I wish them luck.

[Bengie]

unless Mary looks at A and B and decides she doesn't like the GPL, re-invents the wheel and still creates C or C doesn't get created at all.

Or another case for BSD. Mary gets a contract from a company which says their few lines of proprietary code must remain secret. So Mary creates C and C.1(proprietary code). She releases C upstream for everyone else to use and keeps C.1 secret to honor the contract. All of the upstream people still benefit from the bulk of the work.

Remember, the biggest incentive to release your code upstream is that you no longer have to maintain it, which is much much much cheaper as your code would otherwise diverge from the trunk.

Re:I wish them luck.

[Eivind Eklund]

This is a good "Put up or shut up" moment for BSD. For all the whining I hear about "Viral" and "Anti Business" licenses the various *BSD projects sure do have a meager adoption (Buisness, home, free or otherwise) compared to their GPL counterparts (Linux). I think an aggressive, forward looking BSD project would be great to have.

Granted, not all the most popular open source projects have "Viral" licenses (Eg - Most Apache foundation projects), but maybe.. Just maybe Linux's success is in part due to the GPL.

Some people feel the GPL is stealing something that they're somehow entitled too. In reality, it's more of an exchange. You give up the ability to have a certain business model, and in return you get the collective work of everyone else who's made the same agreement. You give up exclusive control of your source in return for a world-class, flexible, free, operating system with widespread uses. For free. With a BSD style license you're able to opt out of that "collective work" provision. You can take, but you don't have to give. As a result, the project does not grow.

This is based on assumptions that don't hold water.

In particular, the primary assumption is that a significant fraction of contributions to GPLed projects come from companies that are forced to give these contributions, and that would not give these contributions if they could avoid it (as in BSD).

My impression (from having participated in BSD development and followed Linux development) is that contributions in this area is actually a larger fraction of development on the BSD side of the fence: Embedded systems companies take the BSD codebase and develop something proprietary with it, and give back the parts that aren't crucial. And logically, it would make sense: If a company feels they need to have proprietary parts, they don't touch the GPLed codebase at all; they just use either BSD or one of the proprietary microkernels.

What *does* affect contributions to BSDs is this myth of exploitation. The GPL has a very effective propaganda preface about "preserving freedom of users", incidentally ignoring that part of this preservation of users' freedom comes by denying some of those that could be users of the codebase the ability to become users. (Look at all the BSD users through Mac OS X.) This myth and propaganda clearly influence some developers.

It's probably in your long-term interest for the project to grow. I think the success of Linux proves this.

Yes.

However, the success of Linux has other possible sources than the license:

• The source code control system and project management led to "distributions", which allows rapid parallel experimentation.
• Distributions lead to more source code flow back and forth than different operating systems with distinct version control systems
• The Linux project structure made the project have a much more softly sloping "insider/outsider" distinction; the BSD structure with core team / committers / general public makes it harder to involve people, on a psychological level. (Everybody thinks things are the responsibility of the next inner circle, and then the core team think development is the responsibility of the community at large.) This led to easier recruiting on the Linux side.
• There are inherent size limits for communities at particular engagement levels (email overload); having multiple communities, in the form of multiple distributions, alleviate this.
• The initial bad support for low end hardware in the BSDs set a disparity in the numbers of users, and there are first mover advantages. People especially select relatively similar operating systems based on whether the operating systems run on the hardware they have, and with more people more drivers get written.
• Linux started with a reliance on binary packages for upgrades, while the BSDs started with a reliance on the ports system and building from source for upgrades. While source

Re:I wish them luck.

[arose]

Oh, there sure is a lot of formerly BSD code going around, I just don't give a shit as an end user because it won't prevent the manufacturer from EOLing my device by lack of software updates. Formerlly BSD code doesn't have nearly the same advantages as actual BSD code.

Re:I wish them luck.

[arose]

Yes, yes it is.

Re:No interest

[Applekid]

Freedom -- true freedom -- is about people having the ability to be assholes if they choose.

Re:could be interesting...

[Anonymous Coward]

You mean like this: http://mirror1.us.bitrig.org/release/snapshots/amd64/current/

Bad name

[davidwr]

They should've called it BitrigBSD.

Re:Bad name

[Anonymous Coward]

HotgritsBSD

Re:Bad name

[unixisc]

Actually, I like a BSD based distro not being called BSD. Just like some Linux distros, like Fedora, Ubuntu, Kubuntu, Mageia and so on, it's nice to have BSD distros not include BSD itself in the name.

Why not starting with FreeBSD?

[Conley Index]

Most points of their agenda are common with FreeBSD and some are already done there or actively been worked on. No one would stand in their way porting WAPBL from NetBSD (if done decently). Ok, stripping the base is (fortunatelly) not on the FreeBSD agenda, but making most of it optional for embedded needs is.

From their FAQ, "OpenBSD [...] has some of the best code around". Ok, but I still do not buy it. If they want to leave some of the conservatism that comes with the security focus of OpenBSD behind (from the article), I do not find a real reason why they started with OpenBSD.

Not that some more good, modern code with any of the BSD would be wrong...

Re:Why not starting with FreeBSD?

[rev0lt]

From what I've read (diagonally) from the article, the fork was started by OpenBSD developers. I guess they decided to fork something they already have good knowledge of. FreeBSD internals are quite different, and while NetBSD has some similarities, they have come a long way since the OpenBSD fork.

Re:could be interesting...

[Bill_the_Engineer]

Linux was getting too lame so I switched to OS X for desktop shit but I'm still looking for a good open source server. Sorry, I can't take Ubuntu seriously as a server, Debian suffers from ultra-cruft like trying to support obscure 90s processors and Red Hat is so expensive it defeats the purpose. So I've been planning to go to BSD but sort of procrastinating on it. This project looks just exciting enough to get me on board.

I know the parent is trolling, but for those who ARE looking for a good open source server there are the assorted BSDs, and of course CentOS, and Scientific Linux.

Re:No interest

[Goaway]

Why would those companies want to have to maintain their own forks and keep those up to date?

no SPARC support?

[Anonymous Coward]

I only used OpenBSD for SPARC hardware and it really belongs to the "big iron". Is this project aiming for the desktop? embedded platforms? Well, good luck with device drivers then. We already have linux for all that and you can't beat it in hardware support. So what's the point?

Re:no SPARC support?

[unixisc]

Frankly, I'm a tad disappointed by them deciding to restrict this project to just the x64 architecture, given the portability of both OpenBSD and NetBSD, As I note above, they could have preserved the portability aspect of it by doing it on NetBSD or even Minix.

Yeah, there are a plethora of OSs wanting to run on PCs. Sparc is one target they could have had. Another would have been Itanium. They might even have gotten some Intel and/or HP backing had they gone that route, and little competition, since the only OS's for Itanium are FBSD, Debian Linux and HP/UX.

Re:no SPARC support?

[gman003]

So what's the point?

1: License. Some people dislike GPL and other copyleft licenses, and demand something BSD-licensed or similar. I personally don't care, but for those that do, this is a Good Thing.

2: Just to be different. It's good to have *options*. I personally despise most Linux's init system. Too convoluted, too complicated, at least for my taste. Some distros, like Arch, have adopted a BSD-style init system. OpenBSD, and by extension Bitrig, also have a BSD-style init system. There's also the different package/port systems, and even the kernel is different. I'm not saying one is necessarily better, but being able to choose one that does it differently is a great thing.

3: Size. I've installed OpenBSD from *floppies* - a single 1.44MB floppy can boot OpenBSD enough to download the rest from the Internet. I wouldn't be surprised if some of the more "and-the-kitchen-sink-as-well" Linuxes can't even fit on a CD anymore.

4: For Teh Lulz: It's a free project. Do they even *need* a reason? Plenty of Linux distros have started off as "hey, this could be fun..."

Re:no SPARC support?

[unixisc]

POWER is right now limited to the games consoles - that's where their main market is, not AIXstations. As for Itanium, Intel's future plans include an 8-core Itanium2, and in NetBSD 5.1, they had a new Itanium source-only port. I wouldn't exactly call it dead, until Intel and/or HP issue an EOL for the platform.

One thing I'm not getting - why aren't the OpenIndiana guys supporting Sparc? If anything, that should be their first target - there are already a plethora of unixes, let alone OSs, for the x86 platform. Sparc, while it has 3 of the BSDs and 2 of the Linuxes, still could use something like OpenIndiana.

Re:no SPARC support?

[unixisc]

1. Every BSD distro there currently is runs on an x86, not to mention the various legacy x86 based Unixes that are already out there
2. Again, there is a plethora of distros targeting PCs. If someone is going to come out w/ something new, especially when based on a server unix like OpenBSD, then why not target orphaned platforms instead, like Sparc and Itanium, which have limited options to start w/?
3. I agree w/ the above AC - talking about floppies in this day & age when even CDs are going out of style seems very anachronistic. What next - 5.5" floppies? OTOH, high bandwidth to download the rest of the OS from ftp servers is not something nearly as ubiquitous Also, will the OpenBSD that fits onto a floppy necessarily include the correct network driver that enables one to download the entire OS from the internet?
4. Sure, but why not make the project more useful by making it an auxillary part of NetBSD, so that all ports of NetBSD benefit? And Minix as well?

Code reinvestment and positive feedback loops.

[bmo]

"be a very commercially friendly code base by using non-viral licenses where possible."

The advantages to Linux over BSD licensed operating systems is that improvements are reinvested in the code base, by mandate. This accelerates development at a much faster rate than we've seen with any of the BSDs since it is a positive feedback loop. Contrary to this, companies take BSD code, improve it, and tend to release nothing back. Because they don't have to. Look at OSX.

So now we have a project that is "focused on modern hardware and SMP" among other things. Compare and contrast to Linux which keeps up with modern hardware a lot better than any of the BSDs. I'm betting the goal of "keeping up with modern hardware" is going to fall by the wayside when they eventually discover how difficult it is when it's just them doing all the heavy lifting.

I also take issue with the "commercially friendly" jab. Linux is GPL, and it's commercially friendly. Sensible companies are not afraid one bit of using Linux. The ones who are don't understand what they're missing when it comes to the code reinvestment cycle.

--
BMO Downmods coming in 3... 2 ... 1...

Re:Code reinvestment and positive feedback loops.

[kthreadd]

The advantages to Linux over BSD licensed operating systems is that improvements are reinvested in the code base, by mandate. This accelerates development at a much faster rate than we've seen with any of the BSDs since it is a positive feedback loop. Contrary to this, companies take BSD code, improve it, and tend to release nothing back. Because they don't have to. Look at OSX.

Such as libdispatch, WebKit and LLVM/Clang? Just to mention a few.

Maybe I missed your point but just because Apple doesn't release their entire operating system as open source doesn't mean that they don't invest and contribute to open source.

Re:Code reinvestment and positive feedback loops.

[bmo]

Webkit isn't BSD. It's LGPL, because it came from khtml.
Libdispatch is Apache.
LLVM/Clang - oh look, you finally struck gold, a BSD license.

--
BMO

Re:Code reinvestment and positive feedback loops.

[kthreadd]

OK. I missed your point then. I thought that you pointed out Apples as an example of an organization that doesn't contribute back.

Anyway, they do contribute to fair amount of projects. Libdispatch even originated from Apple if I remember correctly.

Re:Code reinvestment and positive feedback loops.

[bmo]

Apple has a bunch of BSD code they've modified and never given out. I did not claim that they never do, but they tend not to.

Apple is under no obligation to contribute back. This behaviour was apparent when they tried to deal with the khtml crew and had no idea how to share code, causing a shitstorm and the khtml devs to reject their code. They had to learn how to share code with outside devs. Eventually this happened and we wound up with webkit.

What I'm trying to say is that the BSD license does not encourage the collegiality which I believe is the GPL's greatest strength. YMMV of course, but we've seen it time and again over the decades.

Like I said. For the things that Apple does not wish to share, they miss out on the feedback loop and code-reinvestment. Other people could call it network-effects.

--
BMO

Re:Code reinvestment and positive feedback loops.

[101percent]

We must be fair. Releasing Clang/LLVM is kinda a virtue of necessity. Once GCC went GPLv3, they kinda had no choice.

Re:Code reinvestment and positive feedback loops.

[101percent]

Shit, not releasing but rather adopting.

Re:Code reinvestment and positive feedback loops.

[avgapon]

Are you quite sure of what you say here? E.g. look at Andorid as an example of sharing back by mandate. Revealing some derived source-code and truly sharing it back (making it usable) is not the same thing. Also, regrading the hardware support, GPL has nothing to do (in direct sense) with many companies providing linux drivers for their hardware. You know, much more companies provide windows drivers (and even certify them with ms) and that's not because of gpl. It's just that historically linux has achieved (a lot) more popularity than BSD-derivatives.

Re:Code reinvestment and positive feedback loops.

[onefriedrice]

Apple has a bunch of BSD code they've modified and never given out. I did not claim that they never do, but they tend not to.

And the FreeBSD people would have absolutely not problem with this, assuming it were true. But alas, it isn't true. It just happens that most of Apple's changes aren't incorporated into FreeBSD because it just doesn't make much sense for them to be (e.g. the changes are particular to Apple products or their own operating system), but they do release those changes in the open source version of their OS (Darwin). The parts of Mac OS X that aren't open source or distributed with Darwin are mostly parts which didn't come from FreeBSD anyway. Other code released by Apple that is more general and more appropriate for FreeBSD (like LLVM) is used by FreeBSD.

Regardless, it's a common argument that the GPL has supposedly helped Linux become what it is, but that really short-changes Linux which is really just a kick-ass kernel regardless of the license. The reason Linux "won" over BSD-licensed alternatives was not because of the GPL but because of fortunate timing (USL v. BSDi was obviously a major setback for BSD) and also because Linux is a great kernel.

The GPL itself doesn't do anything to promote a strong free software environment. It just creates a lot of license compatibility problems within a community that would otherwise work better with less duplicated effort. It doesn't even succeed at forcing companies to "give back," it just forces the use of ugly hacks (see how binary blob drivers are implemented). The free software environment already incentivizes companies to "give back," without the need for complicated and incompatibility-inducing license terms, because of reduced maintenance costs (e.g. do companies really want to spend the money maintaining their own fork or patchset? really?), and there are also many good business reasons for a company to release their own code without the GPL supposedly forcing their hand (e.g. to promote interoperability with their products). It's interesting to recognize that most new code released by companies today are not copyleft-licensed but are usually licensed under an Apache license or the MIT or BSD licenses.

Stallman and the FSF should be given credit for the many positive contributions they have made to the free/open source community (GCC and the GNU userland are fantastic and well appreciated), but the GPL just isn't one of those positive contributions.

Re:Code reinvestment and positive feedback loops.

[DragonWriter]

What I'm trying to say is that the BSD license does not encourage the collegiality which I believe is the GPL's greatest strength.

BSD doesn't mandate releasing source code of derivative works that are distributed in object code form the way the GPL does; neither license does much one way or the other to "encourage collegiality", which is vastly more a factor of the particular personalities involved in a community than licenses.

Projects which attract a broad community of developers (individual or corporate) who have (or whose sponsors have) an interest, whether ideological or financial, in a more collegial development model end up with a more collegial development model. Lots of GPL proponents like to attribute what they see as successes of Linux in the OS area as an effect of the GPL specifically, but I think the idea that the particular features of the Linux community are closely tied to the GPL become hard to justify when you look at other areas where there are major GPL and non-GPL open source products (e.g., databases with MySQL on the GPL side and PostgreSQL [BSD] and SQLite [Public Domain] on the non-GPL side.)

Re:Code reinvestment and positive feedback loops.

[ChunderDownunder]

specifically what clause in the GPL3 would deem GCC inappropriate where GPL2 was acceptable?

Re:Code reinvestment and positive feedback loops.

[Raenex]

BMO Downmods coming in 3... 2 ... 1...

Oh look, it's the special-pleading-to-get-upmods trick.

Re:Code reinvestment and positive feedback loops.

[arose]

Was that licensing FUD tasty? Deploy GPLv3 software on whatever you like, what you can't do is ship it on systems where the user can't run modified versions. As far as I'm concerned GPLv3 code is a great litmus test, I'll go out of my way to avoid any vendor who go out of their way to avoid the GPLv3. I prefer hardware that won't be obseleted just because the vendor stops caring, among other things.

Re:Code reinvestment and positive feedback loops.

[arose]

With, yes. Installed on a User Product (as defined in the license) that can technically be updated, license violation (which is a fast road to copyright violation).

Re:Code reinvestment and positive feedback loops.

[TheRaven64]

Apple has a bunch of BSD code they've modified and never given out.

You claimed that twice, but didn't cite any examples. I can grab the sources for all of the FreeBSD-derived parts of libc and the kernel, for example, as well as all of use userspace utilities, from opensource.apple.com. There's little point in doing so, however, because Apple employs enough FreeBSD developers that most stuff that's sufficiently interesting gets pushed back by them already.

What I'm trying to say is that the BSD license does not encourage the collegiality which I believe is the GPL's greatest strength.

The 75+ companies represented at the Vendor Summit at BSDCan would disagree. The strong community was one of the most cited reasons for participating in FreeBSD development. The GPL doesn't force you to participate in the community, it just makes you do periodic code dumps. People who participate in a community because they feel that doing so is beneficial to them are a lot more valuable than people who grudgingly do so because a license forces them to.

Re:Code reinvestment and positive feedback loops.

[TheRaven64]

Two clauses. One was the broad patent license, which may result in accidentally licensing some patents that you didn't think you were licensing. The other is the termination clause, where if you are found to be in violation your license terminates with no mechanism for reinstating it once you return to compliance. The latter clause is the one that makes most companies nervous, as a simple error in distribution could result in their being unable to keep shipping their product.

Re:Code reinvestment and positive feedback loops.

[arose]

The other is the termination clause, where if you are found to be in violation your license terminates with no mechanism for reinstating it once you return to compliance.

That's the GPLv2 to the letter, v3 reinstates you after 60 days if the copyright holders don't say otherwise. GPLv3 is a strict improvement with regards to termination, I don't know where you are getting your information from, but I wouldn't trust them. And while I see why companies would worry about the patent clauses, letting a company contribute code only to be sued by them isn't really a solution either.

Re:Code reinvestment and positive feedback loops.

[Eivind Eklund]

Apple chose to invest in the BSD codebase because they could do this, and would likely otherwise have gone a completely different route (e.g, licensing vxWorks as a base.) So having Apple contribute all changes back was not in the cards.

Apart from that, I'm fairly sure FreeBSD was offered re-licensing on most Apple code for integration back into FreeBSD if we were interested (mail to a private mailing list); lack of takeup on this seemed to be that nobody on the FreeBSD side had the spare capacity to deal with all of this rather than Apple not being willing to give us code.

There was and is a ton of merge work that could be done for FreeBSD; NetBSD and OpenBSD had lots of worthwhile changes at the time, and nobody really bothered to merge most of those either. I did some effort on NetBSD/OpenBSD merging, and wrote some infrastructure, but never actually submitted much based on it. Darwin just showed up as one more source of changes to merge; one that had about as many worthwhile changes, but a bit more hassle license wise (need to send an email to get a license release to avoid contaminating.)

So, this is mostly boils down to infrastructure, code/project organization, and manpower - there's no grand licensing issue involved.

Re:could be interesting...

[larry bagina]

Take a look at DragonFly BSD -- it exists, Matt Dillon has a track record, and it's doing cool stuff (like HAMMER fs).

cheap GCC diss?

[UnknownSoldier]

Honest question: So what were the BSDs (Open,Free,Net) using to compile and run on x86 and amd64 before llvm/clang was around? GCC ?

GCC had its share of problems but this sounds a little ungrateful for what GCC has allowed hackers to do. An open source "good enough" compiler is better then a high priced closed source compiler that may or may not be available for your hardware.

Re:cheap GCC diss?

[kthreadd]

Honest question: So what were the BSDs (Open,Free,Net) using to compile and run on x86 and amd64 before llvm/clang was around? GCC ?

GCC.

GCC had its share of problems but this sounds a little ungrateful for what GCC has allowed hackers to do.

I'm only familiar with FreeBSD but I guess the situation is similar at NetBSD and OpenBSD. They are not ungrateful. They have found another compiler that they think is better for their needs. That simple. Makes good headlines though.

An open source "good enough" compiler is better then a high priced closed source compiler that may or may not be available for your hardware.

I agree, but isn't Clang open source?

Re:cheap GCC diss?

[101percent]

OpenBSD has interested in pcc, but I'm pretty sure they use an older GCC that isn't GPLv3 that they maintain. They kinda do this with a lot of their stuff like htpd and named, but not all for license reasons.

Re:cheap GCC diss?

[The Finn]

NetBSD has PCC working for a few ports. (http://en.wikipedia.org/wiki/Portable_C_Compiler)

Re:Theo is going to me sooooo mad

[Conley Index]

The double edged sword of the BSD License. I'm sure they will probably contribute back but unlike the GPL there is nothing legally to compel them to.

That is not a problem from the perspective of the BSD people. In their experience, code being contributed back only because of legal reasons is so rarely of the quality that anyone would consider merging it back to the original OS that it does not matter to worry too much about that code. Anyhow, there are companies that choose to contribute some of their changes back without legal obligation, which tends to be of better quality, since they want to have it included for whatever reason (for example not to have to maintain their own fork in rapidly changing regions of the code), while they do not consider working on GPL code for their own reasons.

It might be different for different projects.

Re:No interest

[kthreadd]

BSD is a glamour license. The small number of devs that love it want their names in source and help panels. Don't worry about it. OpenBSD has very few users, bitrig will have even less. Don't like it, just avoid any products that use it. Oh wait, you can't tell with BSD liftware.

A fair number of gizmos in your house probably runs it, or a similar flavor of it.

Re:Unfair coverage

[kthreadd]

The difference is that unlike those projects this one does not scuk

Scuk?

Re:Theo is going to me sooooo mad

[Anonymous Coward]

Methinks you need to read up on Theo's antics, before making such a statement. He wasn't happy about businesses using openssh without contributing back, and he kicked up a bloody fuss about when one of the openbsd people made a copy of a linux driver and then tried to "write out" the original of it (you can't, it's still a derived work), only the target for his ire was... the linux people, not the copyright-violator in his own precious bunch.

Theo lost any trace of sanity years ago.

Re:No interest

[gman003]

Another true freedom is having the ability to whine like a little bitch.

Re:No interest

[Microlith]

Why do companies write proprietary software? Because they think it gives them an advantage.

Re:No interest

[Darinbob]

Because it makes the code more widely used. Some people have a goal of making software open for everyone, others apparently have a goal that only their friends can use the software. Reinventing the wheel is a stupid idea, and yet GPL's goal is to force people to either join the commune or reinvent the wheel.

This is nothing new. This is how BSD has done it from the start, because BSD came from a public institution and was funded from taxpayer money and so can not legally, ethically, or morally restrict itself.

This is not about greed, it's about sharing. You probably can more even more money from GPL inhibited software.

Re:No interest

[Anonymous Coward]

I get the feeling you arent really aware of all the growing pains linux has at the moment because of all the going to desktop features they introduce and break useful stuff that worked for years (but is only used if you need to manage more than one system). BSD is more mature and usually doesnt break stuff with new additions.

UNIX family tree

[Picass0]

Whenever I see announcements of "We're creating a fork!!!" the first thing I think of:
http://www.levenez.com/unix/

Lots of tiny branches that just stop.

Re:UNIX family tree

[Saija]

I think that looks like evolution, you know, the tree of the species with some dying and some changing onto something else

latest binutils?

[staalmannen]

I found it strange that an aim in the roadmap was to support the latest GNU binutils [1].. I hope they are trying to adress that piece of GNU dependency too. There is the FreeBSD-project libelf/elftoolchain [2, 3] that could be interesting... [1] https://www.bitrig.org/index.php?title=Roadmap [2] http://wiki.freebsd.org/LibElf [3] http://sourceforge.net/apps/trac/elftoolchain/

Re:latest binutils?

[TheRaven64]

There's also an LLVM linker under development (lld), which we (FreeBSD) will probably be importing to replace GNU ld once it's working. We have a wiki page tracking all of the GPL'd stuff in the base system and our plans on how to remove it before FreeBSD 10.

Re:latest binutils?

[Bengie]

I hope 10 comes out next year some time. I really want to see PFSense using Netmap in the next 3 years.

Re:No interest

[kwark]

Nope, most run linux over here (TV, settopbox, routers, WAPs, DECT basestation, mobile phone). Maybe the washer, microwave or SIP phone are running a *BSD.

Re:Ummmm... you need a real goal first

[Anne Thwacks]

I mean with BSD if I experimented and found a way to dramatically improve it (say IO throughput or something) I probably wouldn't tell the world. I'd keep it secret for my project/products or sell the info.

That is probably because you are young and foolish.

If (corporate) you run a major project using your proprietary software on a bunch of BSD based servers, and you get your people to hack the OS code to fix a performance bottle neck, you certainly would (unless thick as two short planks) release it back, because if you do, then the hacks will (a) get thoroughly code reviewed and tested and (b) be introduced to the globally supported codebase, and consequently automatically introduced into all future OS updates, fully tested, and without your staff having to do a stroke of work.

If you are selling a service (VoIP, Money transfer, share dealing, etc), is the not like selling selling the software. The vast majority of OS users are not your competition, and no significant part of your competitive advantage is down to OS performance (if it is, you are doomed: short sell your company now)

Re:Theo is going to me sooooo mad

[jonadab]

> The double edged sword of the BSD License. I'm sure they will probably
> contribute back but unlike the GPL there is nothing legally to compel them to.

In practice, this only matters if the project is so stagnant that it doesn't actually matter any more after all.

If the project is active, the work of maintaining your changes (either by constantly updating your patches every time an upstream change breaks them or, if you prefer to go the clean fork route, porting over or reimplementing upstream changes that you specifically want) is so burdensome that any reasonably competent developer will WANT to get his changes incorporated upstream just so he can get off the maintenance treadmill for a bit and maybe have time to implement something else.

Why not starting with NetBSD?

[unixisc]

It would have been good for them to take their project and changes back to NetBSD, which might have been happy to use the improvements. As it is, there are a lot of legacy servers not based on x86 that could use this fork, so if it was too much work, then making the changes and then integrating it upstream into NetBSD might have been a better idea, and NetBSD could have made it available on all architectures. Another thing they could have done - take their changes, gone to Minix, and there, put their changes there, be it Clang/LLVM, and so on. Portability would also have been preserved, instead of being needlessly sacrificed

Re:No interest

[Kergan]

I'm mystified what the motivation would be to work on something like that unless its just another paycheck.

The same motivation that leads coders to contribute to GPL software, in spite of the fact that gazillions of other coders and designers and etc. make money on it without ever contributing anything back.

Re:Unfair coverage

[Kergan]

Scuk?

Dcik. :-)

Re:I thought the problems are being addressed

[unixisc]

I'd like them to enhance IPv6 support. One potential use for this distro could be something like pFsense or Monowall - make an IPv6 router/firewall that has all the security that OBSD has, along w/ the new compilers.

On multiprocessing, I'd think that DragonFly is already way ahead, so if things like PF, OSSH and so on are ported to DragonFly, it would work. They can even take Clang/LLVM there

Re:Theo is going to me sooooo mad

[erice]

If they contribute back to the main trunk, then I think all is well.

The double edged sword of the BSD License. I'm sure they will probably contribute back but unlike the GPL there is nothing legally to compel them to.

How does the GPL legally force people to contribute to the trunk? The source must be released, sure. But that doesn't mean you need to create patches, integrate, or even communicate in any way with the developers working on the trunk.

This fork appears to be open source anyway.

Re:No interest

[MartinG]

Including the freedom to take away other peoples freedom, I suppose?

Re:No interest

[bill_mcgonigle]

True freedom is anarchy. No thanks.

You're right, we really need to elect a group of people to coordinate the open source ecosystem to stop this messy chaos we have now. Folks who are smarter than us and can tell us what we all need to do, which projects should exist, and which technologies will win the future.

ahhahahahahahah.

Re:No interest

[bill_mcgonigle]

Freedom -- true freedom -- is about people having the
ability

to be assholes if they choose.

Yep, that's human freedom and what BSD is all about (well, except they try to force copyright on you - WTFPL FTW).

GPL anthropomorphizes code and gives it ultimate freedom. Different ends, different means.

Re:No interest

[magic maverick]

True freedom would be using the do what the fuck you want to license, or creative commons zero or similar.
BSD isn't true freedom because it requires you to keep the name of the original developer.

Re:could be interesting...

[rev0lt]

DragonFly BSD is HUGELY underrated. It is an aswesome project, and I'd love to see HAMMER in both FreeBSD and OpenBSD.

Re:No interest

[101percent]

OpenBSD has few users? How did you quantify this? There are no usage statistics collected.

Re:Theo is going to me sooooo mad

[buchner.johannes]

If they contribute back to the main trunk, then I think all is well.

The double edged sword of the BSD License. I'm sure they will probably contribute back but unlike the GPL there is nothing legally to compel them to.

How does the GPL legally force people to contribute to the trunk? The source must be released, sure. But that doesn't mean you need to create patches, integrate, or even communicate in any way with the developers working on the trunk.

This fork appears to be open source anyway.

It does not legally force people. But one customer is enough to let the cat out of the bag. So the company might as well.

More important I think is the update path. If upstream introduces a feature, you have to merge, making it very hard to keep up-to-date if you don't push your changes upstream. If OpenBSD is active enough and downstream wants those changes, they will also try to push their changes into OpenBSD -- it just makes things easier. In the case of the Linux kernel, it is just plainly impossible to keep a independent version.

Meager Adoption? No Growth?

[nurb432]

Um, you really don't have a clue, do you?

Re:could be interesting...

[csirac]

Nice troll. Funny that you picked Amiga, because this architecture failed to meet release criteria and was dropped from Etch in 2007.

Debian's stated goal is to be the "universal" operating system. They support any hardware that has active developers for it within the Debian project. At the same time, they don't allow a Debian release to be blocked by some architecture which lacks sufficient active developers.

Everybody admires Debian's packaging system. It's not really apt or dpkg though, that is awesome - it's Debian's packaging policies and their strict enforcement. If Debian is slow to release sometimes, it's because they're packaging and fixing so much buggy open source software, not the diversity of the architectures they support.

Re:Theo is going to me sooooo mad

[Anonymous Coward]

I'm sure they will probably contribute back but unlike the GPL there is nothing legally to compel them to.

Please provide a link so that I can download Google's changes to GPL'd software used on their servers. I'd like to run Gubuntu and use their optimizations to OpenJDK. The GPL-2 doesn't protect against Tivoization either, making the arguments for it mostly nonsense.

ARM...

[DECTerm]

I am curious if they are going to support Raspberry Pi too!

Re:could be interesting...

[TheRaven64]

Porting HAMMER is difficult because Dillon hacked up some bits of the VFS to make it work. These would need to be back-ported to FreeBSD for it to be possible. It also seems that HAMMER is basically dead from a development perspective now. I think it's quite likely that we'll port HAMMER2 at some point though, once it's stabilised.

Re:Theo is going to me sooooo mad

[unixisc]

Bitrig - does this remain under the ISC license, or have they gone w/ a different BSD license, such as the FreeBSD license or the MIT license?

Re:No interest

[Bengie]

One might say "taking credit" is the only real issue copyright should protect. copy someone else's work all you want, but don't you take credit for creating it. One could say BSD is the most free before detracting from the creator.

Re:No interest

[Bengie]

Better check your network stack on Linux. I'm sure it has a BSD logo on it.

Re:No interest

[Bengie]

Apple has given back a lot to BSD because GPL didn't want any of it unless it could get all of it.

Perfect is the enemy of good. GPL is a bit idealistic. Great ideal, can't argue that.

Re:No interest

[VisceralLogic]

Including the freedom to take away other peoples freedom, I suppose?

If someone modifies BSD source and distributes their product without distributing their own source, they haven't taken away anyone's freedom. Anyone who wants to use their binaries without having access to the source can make that choice. Anyone who demands to have source with binaries can go back to the original code. The developer has not taken away anyone's freedom, they have just chosen not to extend certain freedom.

Re:Theo is going to me sooooo mad

[ArhcAngel]

The day Google decides to sell those changes to a third party I'll send you the link they put up for the source.

Re:No interest

[Goaway]

If you are going to write something from scratch, you are going to have to maintain it yourself anyway no matter if the source is open or not.

That is a very different situation from using an existing open project maintained by other people. The big advantage there is that other people have done and will do your work for you to some extent. You lose that by making a non-public fork.

Is Btrfs its file system?

[unixisc]

Wouldn't one think so from the name? ;-)

Re:could be interesting...

[unixisc]

Well, Debian is not just doing Linux, but also has a FreeBSD and Hurd port in development. While their current kFreeBSD port is accompanied by their userland which is GNU based, they are also working on Clang, so they'll have that as well. These guys actually could have tried Debian as well to work w/ - maybe do a kOpenBSD project w/ their same goals.

Re:No interest

[arose]

If having the ability to modify the software you run isn't a freedom, then there's not much freedom in BSD licensed software (and software in general) to begin with. If having the ability to mofify the software you run is a sort of freedom, than it can be removed by excersing the freedom of the BSD licenses to not distribute source code. "Not extend a certain freedom" is functionally identical to removing it, all that remains is figuring wether or not one finds the freedom to do so more important than the freedom that will be optional. That has always been the balance of freedom when more than one person is involved but some people seem to believe that freedom is an absolute and ignore this.

Re:No interest

[arose]

GPL attempts to preserve as much freedom as possible along the entire chain of distribution, but that doesn't sound good when trying to do character attacks...

Re:No interest

[arose]

Care to show an example?

Re:No interest

[arose]

If reinventing the wheel is bad, then so is hiding away the one you grafted on, if it is not bad then let then why do you expect others to provide the base for the invention of propriatary wheels that will only need to be reinvented to not be continously reinvented?

Re:No interest

[Darinbob]

The problem is that you're assuming the entire world can just share software. It doesn't work like that. I would be fired tomorrow if I gave away the company's trade secrets to all competitors. But FSF hates that model, they want all businesses to be open source business, never have secrets, never have patents, never have copyright. Maybe someday that may happen but not today. The flaw with FSF is that their idealism smashes headlong into pragmatism and that they are more concerned with creating the perfect world than with actually getting coding done today.

Re:No interest

[arose]

The problem is that you're making assumptions about me. If your company sells copies of software in some form or fashion, then reinventing wheels is just about your job description. If you are going to make others reinvent the wheel it is a tad selfish to complain that they make you do so.

Nice Idea

[DaMattster]

It is a nice idea but the name makes me want to roll my eyes. Why not something cool like TornadoBSD or something along those lines? :D

Re:No interest

[rthille]

my Sony BluRay player is BSD.

Re:A DeRaat-free-OpenBSD??

[The Finn]

it's called NetBSD.