Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
BSD and commercial Forks

By kraant in News
Tue May 02, 2000 at 04:42:53 PM EST
Tags: Round Table (all tags)
Round Table

Did you ever notice how rarely the *BSDs get commercial forks, and how even when they do, the company in question almost always ends up releasing it's source back to the BSD community?

Well I've been thinking about the topic and want to see what other people think of why this is...


I am a techie in a largely Macintosh environment and recently was told that one of the other techies wanted to pick my brains on *BSD since he was going to get his hands on MacOS X. As a side issue I mentioned that there were no guarantees on how similar it would be to FreeBSD since they'd probably forked it off in a fairly major way.[1]

I was told that Apple had released most of it's code back to the world.

But what occurred to me as he said that was that, despite the lack of any legal protection, I have never seen an obvious[2] fork of a BSD that didn't eventualy release a large[3] proportion of it's code back to the BSD community. Which makes me wonder whether the militancy of some GPL advocates[4] is neccesary.

This is also a very important point to examine since, if the GPL is ever found to not be legally binding, then whether linux and other major GPL projects remain free will depend on whether the dynamic that BSD has can apply to other free software projects, or whether it is because of BSD's unique circumstances that it remains free.

I have a couple of theories on why BSD remains free...

  • Firstly there is the momentum and support of a project's developers. If anyone forks the code and doesn't release the changes back to the BSD community, the BSD code will deviate further and further from the now-proprietary code that the company appropriated, negating a large part of the advantage of working with a free project. since...
  • A large amount of resources that would otherwise not be required are needed just to incorporate any changes made in the free fork (and from working on a couple of team projects without any CVS or equivalent I know personaly what kind of hell it is trying to patch some changes you made independently onto another part of the code that someone else has been altering)
  • Also, I think that in general BSD attracts a higher standard of contributer. The two freeBSD contributers I know personally are both SysAdmins, one for a major ISP, the other for the Computer Science department at my local uni. This is opposed to the linux developers I know who are all either students or 16 year old hackers[5]. And since BSD development tends to concentrate on the really important things like a rock solid kernel and security rather than all the fancy gadgets like Gnome, this means that unless a person's coding ability is somewhere around godlike[6] they wouldn't be that useful... And a vast majority of the people who would be good enough to be useful would tend to already believe that code should be released back to the community... which brings me to
  • Cultural values and mores. Because BSD never got the media overexposure that linux suffers from, by the time people even think of trying to install it they already know about gnu, the gpl, free software. And really unless you are heartless scum it's pretty hard to disagree with the spirit of free software even if you do disagree with the implementation. In essence it is fairly unlikely for BSD to get a commercial fork because it is arcane and underground enough that to get a working product, uasbale by the average user, without incorporating GPLed or otherwise copylefted software would probably be more work than creating a simpler OS from scratch. And if it is too arcane for the average user then the only people who would want it would know that they could get a free version with source code, making commercial forks impractical.

So what does everyone else think are the reasons for BSD not disintegrating and getting commercialised in the way that caused RMS so much grief at MIT AI labs? And which GPL projects would survive if the GPL and copyleft licences were found to be invalid?

Personally I think that large projects which need to be very reliable and have a lot of developer momentum such as linux or apache would continue without too much of a hiccup but smaller, more application- or GUI-based projects would soon find themselves having nonfree forks that would make some CEO stinking rich...

daniel - who feels kinda depressed now and really should go do some homework

[1]Little did I know just how wrong and yet at the same time right I was, because I didn't know that Apple had decided to use the Mach kernel at the time. I hadn't been paying attention to it and the last I'd heard of MacOS X it was based off freeBSD to the point that you could telnet into it and you'd see it announce itself as freeBSD :P

[2]By this I mean where it's a variation of the original program not someone borrowing a slab of code

[3]Normally only leaving totally trivial non-core technologies unreleased

[4]Mainly I am talking about those irritating trolls who go on about how evil BSD is because it isn't GPLed

[5]Not to put down either of these two fine professions ;)

[6]I am definitely not competent enough to assist a BSD project :P

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Related Links
o MacOS X
o most of it's code
o GPL
o free software
o so much grief
o linux
o apache
o Also by kraant


Display: Sort:
BSD and commercial Forks | 34 comments (34 topical, editorial, 0 hidden)
It's an interesting question, but i... (2.00 / 1) (#9)
by leshert on Tue May 02, 2000 at 01:06:49 PM EST

leshert voted -1 on this story.

It's an interesting question, but it seems as if the poster has already made up his mind...

Good write-up. Personally, I'm not ... (2.00 / 1) (#4)
by Ozymandias on Tue May 02, 2000 at 01:07:13 PM EST

Ozymandias voted 1 on this story.

Good write-up. Personally, I'm not that fond of BSD - simple personal preference. Still a good article.
- Ozymandias

FUD at it's finest. ... (1.00 / 1) (#6)
by kovacsp on Tue May 02, 2000 at 01:26:04 PM EST

kovacsp voted -1 on this story.

FUD at it's finest. Who cares? GIve it up already.

Woop, good article.... (1.00 / 1) (#3)
by alisdair on Tue May 02, 2000 at 01:42:00 PM EST

alisdair voted 1 on this story.

Woop, good article.

heartless scum, eh? ;-)... (1.00 / 1) (#7)
by thelaw on Tue May 02, 2000 at 02:02:55 PM EST

thelaw voted 1 on this story.

heartless scum, eh? ;-)

Nothing more than one man's theorie... (1.00 / 1) (#11)
by dgay on Tue May 02, 2000 at 03:34:52 PM EST

dgay voted -1 on this story.

Nothing more than one man's theories. Nothing really backing it up except for opinions.

I'd like to try BSD at some point. ... (1.00 / 1) (#8)
by Saint Zero on Tue May 02, 2000 at 03:52:23 PM EST

Saint Zero voted 0 on this story.

I'd like to try BSD at some point. Long article, for an opinion piece. I don't like this recurring "BSD is for mature people/Linux is for hacker kids..." crap.
---------- Patron Saint of Nothing, really.

Yes, this is an interesting theme t... (1.00 / 1) (#1)
by RangerElf on Tue May 02, 2000 at 04:09:30 PM EST

RangerElf voted 1 on this story.

Yes, this is an interesting theme to ponder.

I used to advocate the GPL and used... (3.00 / 1) (#10)
by drdink on Tue May 02, 2000 at 04:15:59 PM EST

drdink voted 1 on this story.

I used to advocate the GPL and used Linux, and then I read the GPL. Then I read the BSD license. Then I tried FreeBSD. Oh my god, what was I smoking? Linux was nice, but BSD offers much more stability, trust, and maturity. I also like the majority of the BSD community, though there are some people who are complete asses. You can find help quite easily on just about anything. BSD's forks also work to maintain a relationship, such as library and binary compatibility. Before you say "BSD sucks", "Windows sucks", "Linux sucks", I suggest you try them all at length first and then make your decisions.

Re: I used to advocate the GPL and used... (none / 0) (#16)
by Inoshiro on Tue May 02, 2000 at 07:21:22 PM EST

Disclaimer: I have run Linux (2.0.x, 2.2.x, 2.3.x), Win2k, WinNT, Win9x, and OS/2, but I haven't run FreeBSD (although I've played with OpenBSD locally).

I must say, I don't know where you get "FreeBSD is stabler" from. On a workstation level, even Win2k is stable enough (although once you "pollute" it by installing Non-MS sw, the stability bleeds away like helium from a balloon). On a server level, I've never had a problem with Linux.

5:13pm up 66 days, 15:29, 8 users, load average: 1.00, 1.00, 1.00

It works for me :-)



--
[ イノシロ ]
[ Parent ]
Re: I used to advocate the GPL and used... (none / 0) (#20)
by drdink on Tue May 02, 2000 at 09:18:17 PM EST

FreeBSD is more stable in my personal experience and in the people I've talked to who have used both Linux and FreeBSD. BSDs make a better server OS than Linux, imho. It can run BSD, Linux, and IBCS2 binaries. Unless it has been fixed, Linux can only handle files up to the size of like 2GB. FreeBSD doesn't have this limit. Places such as media production houses would have problems with this limitation. If, in the rare situation, the kernel was to crash, FreeBSD has the nice feature of being able to do a crashdump to the swap partition which can later be read and the problem can be debugged with gdb. FreeBSD also doesn't use flatfile password files. The Linux /etc/passwd and /etc/shadow can't compare to the speed of /etc/pwd.db and /etc/spwd.db. Linux also has the /proc filesystem. This, in my opinion, is ugly as hell. You do not replace system calls with a filesystem. Instead, make a program such as the FreeBSD 'sysctl' command that gives a MIB-style interface to view/set values. This makes it faster, cleaner, and probably takes less memory too.

[ Parent ]
Re: I used to advocate the GPL and used... (none / 0) (#24)
by Inoshiro on Wed May 03, 2000 at 01:46:49 AM EST

Tsk, you have some misconceptions :-)

Reasoning behind no kernel debugger: Linus Torvalds doesn't like to keep kernel debuggers in the kernel because unless you are a developer, it won't help you at all. Also, he thinks it's better to fix the source of the problem (stare at code, attain greater understanding of problem and reengineer) than simple fix the symptom.

For the proc file system, Linus believes it's easier to allow humans to read the dirs and be able to tweak things at will, instead of forcing people to use strange numeric sysctls. The same applies to devfs, which is being put into the kernel to allow things to be made manipulatable via any text editor... rather than only sysctls and API calls.

Speed doesn't matter when you cannot fix your passwd and shadow equivs without using strange binary tools.

I like the cleaner, more human usable text interface view of Linus, our Glorious Leader :-)

As for the file limit, they're gone in 2.3.x (AFAIK).. on 32-bit archs. 64-bit ones (Alpha, Sparc64) never had that limitation :)



--
[ イノシロ ]
[ Parent ]
NeXT tried to make a proprietary, c... (4.50 / 2) (#5)
by your_desired_username on Tue May 02, 2000 at 04:21:13 PM EST

your_desired_username voted 1 on this story.

NeXT tried to make a proprietary, closed source fork of gcc.
Because of the GPL, they were not able to distribute their fork as closed source, and gcc got an objective-C front end (which it still has today) out of the deal.

The company that produced the first C++ front end for gcc (which has since been wholly re-written) also tried to make a closed source fork of gcc as well; again, they failed, and gcc got a c++ front end.

I do not know about freebsd, but gcc would be much poorer if it was not gpl'd.

Note that many comercial unices (IRIX and sunos for sure) are known to have BSD code in them.



This is an interesting topic. There... (3.70 / 3) (#2)
by raph on Tue May 02, 2000 at 04:35:49 PM EST

raph voted 1 on this story.

This is an interesting topic. There is a lot of misinformed speculation about forks out there. My own personal view is that forks happen when one set of maintainers is (at least in the opinion of the forkers) mismanaging the project. If you look at the history of the Great Forks (gcc/egcs, emacs/xemacs, *BSD, ncurses), I think you'll find that theory has more predictive (retrodictive?) power than anything having to do with licenses.

Also, I think giving code back to the community is a quid pro quo - it means you'll get support from the community. For an operating system in today's ridiculously complicated environment, that's essential. Otherwise, you're likely to have something that runs on too little, too late (BeOS, anyone?).

Re: BSD and commercial Forks (4.50 / 2) (#12)
by Emacs on Tue May 02, 2000 at 05:14:58 PM EST

Also, I think that in general BSD attracts a higher standard of contributer. The two freeBSD contributers I know personally are both SysAdmins, one for a major ISP, the other for the Computer Science department at my local uni. This is opposed to the linux developers I know who are all either students or 16 year old hackers

You are making a dangerous assumption here if you think that being a sys-admin makes you a competent developer. In fact I can tell you that the sys-admin at my place of employment (I work for a software company) has never written a line of code, and the windows developer I eat lunch with knows next to nothing about setting up a netork and keeping it running. In other words, a sys-admin and a software developer are two very different beasts. The skills needed to administrate a network do not translate to developing complex software applications

Now on to point two. I think it's also a very dangerous assumption to say that BSD attracts a higher caliber of developer(than Linux). If you are basing that assumption on the teenage hackers you know then you are using faulty logic. In fact you are using no logic at all.



Re: BSD and commercial Forks (3.50 / 2) (#13)
by inditek on Tue May 02, 2000 at 05:29:52 PM EST

it seems half the posters, especially those with one-liner complaints, didn't read any footnotes. neither did the dude bitchin' about the author making himself as competent as a developer. sheesh.

anyway, yes - it does seem the author made up his mind.... and he hit the nail on the head:

probably the most compelling reason commercial developers return their code to the community is to continue to reap the benefits of its openess.

apple should benefit greatly because of two things: BSD is a free OS, so they'll (and their more savvy users) will be able to benefit from other contributions and BSD is more tightly controlled than linux, making it more stable and secure in many peoples eyes. granted, certain fixes and changes take longer - but it's a great free OS and a good option for commercial enterprises to develop from.



Re: BSD and commercial Forks (none / 0) (#17)
by mattm on Tue May 02, 2000 at 08:22:33 PM EST

Yeah, and I think the author shouldn't have used footnotes in the first place! If readers need to read your notes to get the general idea (as opposed to the minor details) of what you're saying, you're using them incorrectly. Unless you're Douglas Adams, which Kraant here definitely is not.



[ Parent ]
Re: BSD and commercial Forks (4.50 / 2) (#14)
by Alhazred on Tue May 02, 2000 at 05:50:16 PM EST

I find it very amusing that the author uses BSD as an example. There are so
many forks of BSD out there it would practically be impossible to count them
all!

Practically EVERY commercial Unix started out as a BSD fork. Then we have
netBSD, freeBSD, BSD this, BSD that! Granted many of them are quite similar and
they seem to trade code often, but since it is open source code that isn't all
that surprising... In fact if you look closely at the Linux kernel you will see
a few bits of BSD in there too.

Personally I think BSD underwent a LOT of unproductive forking. Certainly a lot
of man hours had to go into patching changes to one flavor of BSD back into the
code bases of the others. IMHO the reason for this is the very closedness of
participation traditional in BSD projects. If a project is open to the
participation of all, then it won't fork. gcc/egcs and emacs/Xemacs are also
prime examples of the same phenomenon. Small groups of developers walled
themselves off from the rest of the community. Other developers felt the need
to participate, and also believed their needs and concerns were not getting a
hearing, so they forked. 

Personally I don't think this has much to do with which license you use either.
gcc was GPL, yet it was forked, and so was BSD with its "BSD license". The more
open the "Uber Hackers" that run a project are to ideas and contributions from
outside, the less likely a fork is, pure and simple. Linus is a genius at this.
Sure the Linux kernal has its "core group" of developers, but its a MUCH larger
group than BSD, and its MUCH more open to outside suggestions. Some people
might say BSD is more stable etc, but the fact is, Linux has 20 times the
installed base. 

That is not dead which may eternal lie And with strange aeons death itself may die.
Re: BSD and commercial Forks (3.00 / 1) (#15)
by your_desired_username on Tue May 02, 2000 at 06:49:47 PM EST

I think you missed the fact that this story is about proprietary, closed source forks of BSD.

There are plenty of these as well (the comercial unices you mentioned), but note that most of them were not forked off of recent BSD source.

Since kraant was talking about closed source forks, the license does matter.

Also, egcs/gcc have rejoined; see gcc.gnu.org.

Since you mentioned gcc, see my earlier post about how the gpl has benefited gcc.

[ Parent ]
Open forking matters too (none / 0) (#29)
by kmself on Wed May 03, 2000 at 02:50:48 PM EST

Though the article focusses on proprietary forks, open forking of projects is also significant, particularly when forking occurs under incompatible licensing terms. This essentially splits development effort and divides developer base, and is long term harmful.

License transitivity under BSD means that it's possible to have licence C and license D which are both compatible with license B(SD), but not with each other. Project developed under C and D can both benefit from work done under B, but not under one another (without cleanroom reimplementation of code).

This is a theoretical, not an empirical argument, and I don't know the specifics of licensing say, with the Apple Public License (which is heavily BSD influenced, IIRC). However, the theoretical argument doesn't apply to the case of the GPL -- GPLd code is compatible, from a licensing standpoint, with GPLd code, always has been, always will be. Code forking necessitated by licensing incompatibility isn't possible.

--
Karsten M. Self
SCO -- backgrounder on Caldera/SCO vs IBM
Support the EFF!!
There is no K5 cabal.
[ Parent ]

Re: Open forking matters too (none / 0) (#30)
by your_desired_username on Wed May 03, 2000 at 04:36:21 PM EST

GPLd code is compatible, from a licensing standpoint, with GPLd code, always has been, always will be. Code forking necessitated by licensing incompatibility isn't possible.

Fortunately not, but something vaguely similar does occasionally happen.

Eben Moglen, RMS, and others believe there is legal reason to believe that the licence to a piece of code can only be defended by the copyright holder, or an authorized representitive (say, the copyright holder's lawyer).

For this reason, the FSF insists on copyright assignment for all contributions to most (all?) 'official' gnu projects, such as gcc, emacs, glibc, bash, etc.

Emacs and Xemacs are both licensed under the GPL. However, Xemacs contains much code that has not been copyright assigned to the FSF. I have read that this played a part in the emacs/xemacs split, and is still an important reason (along with some architectural decisions) why they do not rejoin today. See http://www.xemacs.org/About/XEmacsVsGNUemacs.html

As for the BSD-derived codes with incompatible licences problem, I do not know of an example, but I would imagine that if more comericial unices start open sourcing their code you may see such an example, as many of them are rumored to still contain code from the early BSDs, and possibly from more recent BSDs.

You might also find such an example if you compared SCSL licensed code with APSL licensed code; unless I am misremembering these licenses, they are incompatible.

[ Parent ]
Re: BSD and commercial Forks (none / 0) (#33)
by Anonymous Hero on Sat May 06, 2000 at 10:41:16 AM EST

Modern BSDs forked off of 386BSD for one reason. 386BSD development stalled. NetBSD and FreeBSD forked off that project with different goals. FreeBSD supported only i386, NetBSD tried to support anything with a processor. Later OpenBSD forked off NetBSD because of personal differences, but targeted security.

Personally, I don't see how that hurt anything. Each system has an area where it specialises, and each system does that job well. It has nothing to do with the licensing.

As for linux, it would be just as easy to fork the linux kernel. The GPL doesn't prevent that. It very well could happen some day, but I certainly don't think it will hurt Linux, as the new kernel would have to have a different niche than the original one, and will bring new ideas to the table.

In short, Forking != bad

[ Parent ]

source and licenses (3.00 / 1) (#18)
by mattc on Tue May 02, 2000 at 08:59:54 PM EST

I'm not sure where you get the idea that Apple is embracing free software. The source they have released is under their own Apple license, not BSD or GPL.

I also don't see any advantage to releasing under the BSD license instead of the GPL. If you release under the BSD license companies can just steal your work!.. and if you are using the "modern" BSD license they don't even have to give you credit for it!

Re: source and licenses (none / 0) (#31)
by Anonymous Hero on Fri May 05, 2000 at 08:23:09 PM EST


Why do you have a problem with people (including closed source projects) using your code? Why on earth would you write it if you didn't want people to use it?

[ Parent ]
Re: BSD and commercial Forks (3.00 / 1) (#19)
by analog on Tue May 02, 2000 at 09:04:20 PM EST

Okay, just some random pot shots as they occur to me:

But what occurred to me as he said that was that, despite the lack of any legal protection, I have never seen an obvious[2] fork of a BSD that didn't eventualy release a large[3] proportion of it's code back to the BSD community

AFAIK, there are several, but other people seem to have addressed that already. I would also warn you not to assume that the only way to use BSD code and not contribute changes back is to fork the operating system; I believe Windows9x networking owes a debt to BSD, for instance...

This is also a very important point to examine since, if the GPL is ever found to not be legally binding, then whether linux and other major GPL projects remain free will depend on whether the dynamic that BSD has can apply to other free software projects, or whether it is because of BSD's unique circumstances that it remains free.

Complete poppycock. If the GPL is ever found to be not legally binding (and while I hear this a lot, I have yet to hear anyone give a good reason why it would happen), then it reverts to standard copyright and it will be illegal for anyone but the author to distribute GPL licensed code. Since just about every piece of GPL software I've ever seen states that it may also be distributed under any later version of the license, the FSF would only have to make the modifications needed to make it valid to take care of the problem.

Also, I think that in general BSD attracts a higher standard of contributer. The two freeBSD contributers I know personally are both SysAdmins, one for a major ISP, the other for the Computer Science department at my local uni. This is opposed to the linux developers I know who are all either students or 16 year old hackers[5]

The sky is blue. The denim in my jeans is blue. The sky is made of denim. The author also appears to want to compare BSD system developers with Linux application developers. Apple, meet orange.

For those who haven't figured it out yet, let me help you with a few concepts. The GPL is a good license; it's targeted at a specific process, and it does its job well. The BSD license is a good license; it's targeted at a specific process, and it does its job well. Linux is a good operating system kernel; the utilities used to flesh it out are also good. Ditto BSD. I use both daily, and I see no significant differences between them in my usage (ymmv). There may be instances where one is a better choice than the other; there are probably far more instances where either choice is just as good as the other, and it comes down to personal preference.

Well; I feel better now. I'm going to go ponder some important questions, like what I'm having for dinner tonight...

Licence or OS (4.50 / 2) (#21)
by Dacta on Tue May 02, 2000 at 09:26:25 PM EST

Are you trying to compare the BSD licence vs the GPL, or are you comparing the Linux kernel vs the BSD kernel?

It seems like the second one, and it seems like you've already made up you mind.

And since BSD development tends to concentrate on the really important things like a rock solid kernel and security rather than all the fancy gadgets like Gnome

Firstly, the idea that the (Free, mostly)BSD kernel is better in some way than the Linux kernel is put forward all the time. I have never seen any evidence to support this.

Take the old claims about how FreeBSD networking used to be better than Linux networking. Most people accept this as true, and it may well have been. I've never seen any benchmarks to back it up, though, and I've looked.

I found a paper ( http://www.usenix.org/publications/library/proceedings/sd96/full_papers/lai.txt) about some benchmarks back from 1996 (Linux 1.2x), that compared Solaris(x86), FreeBSD and Linux did show that FreeBSD had faster networking, but that Linux does well on system calls, context switching, and pipe bandwidth. Its performance on small-file workloads with intensive metadata manipulation is an order of magnitude faster than the other systems.

I remember some benchmarking someone did on serving a static page, a perl CGI script and a MOD_Perl page from Apache on FreeBSD and Linux. (It was posted on Slashdot in 1998 or 1999). FreeBSD served maybe 10% more static pages than Linux (which I guess shows better network performance), but Linux got a lot more (maybe 25%, I can't remember exactly) page hits on the other two.

It's the same for the "Linux has better/faster multiprocessing support than FreeBSD". Everyone knows that, but how? Where's the number?

As for the claim about BSD developers concentrating on the Kernel rather than "Fancy Gadgets like GNOME" - these are the same people who jump on anyone who says that GNOME is for Linux, and always point out that you can run GNOME on BSD. You can't have it both ways, you know.

Anyway, these kind of stories always annoy me. I use to use NetBSD on some sun3/50s, so I aren't an anti-BSD bigot or anything, but I think the title of this article was misleading. I was hoping to read some informed discussion about why commercial comapnies do return forked code but it seemed to be more of a BSD vs Linux thing.

Re: BSD and commercial Forks (3.00 / 1) (#22)
by Anonymous Hero on Tue May 02, 2000 at 09:34:40 PM EST

You're right on a lot of points, but you might forget that XFree, under the BSD license, has been repeatedly scavenged for code by large corporations who then charge money for it (ahem, Xi Graphics....). This is perfectly legal. However, in the BSD operating systems themselve, this basically doesn't happen at all.
One other thing: BSD lets companies use the code without getting teir legal departments in on the deal. Most of these companies release their code back to the world. That's why many FreeBSD contributors are more active than their Linux counterparts - often, they work for businesses which have a vested interest in making the code better.
in conclusion, the BSD does not spell out that companies must release their code, but it doesn't need to because 90% of the time they do so anyway. Why is Linux not really suitable for embedded environments? Because companies have been afraid to use it, since much of their code would need to be released GPL. They choose the BSDs, which results in better BSD code most of the time.
And wrt BSD forking - in my experience, it hasn't really been a problem. The four "classic" BSDs are Open, Net, Free and BSDI, but BSDI has been scrapped in favor of Free and Net is dying a slow death to be replaced by Open (aimed at the same market). Of course, a new fork is coming down the road - Darwin - which will work on Macs and PCs but won't have X or any other graphical environment. On the other hand, I think it will probably catch on in server rooms where Free is not suitable.

Re: BSD and commercial Forks (none / 0) (#27)
by orabidoo on Wed May 03, 2000 at 07:48:50 AM EST

there *have* been commercial forks of BSD; most of the old Unix systems were exactly that, but they were later replaced by SysVr4 as the industry moved that way. SunOS, BSDi, Ultrix and NeXT were all based on BSD.

[ Parent ]
Re: BSD and commercial Forks (none / 0) (#32)
by Anonymous Hero on Fri May 05, 2000 at 09:17:20 PM EST

NetBSD and OpenBSD do NOT target the same market, and NetBSD is far from dead. Have a look at this page for a few reasons why it is not dead. Shaded

[ Parent ]
Re: BSD and commercial Forks (4.00 / 1) (#23)
by kraant on Tue May 02, 2000 at 11:19:24 PM EST

Originaly I was going to flame some of the more kneejerk reactions to the article but thinking about it I've decided that continuing the rational discussion and just ignore the more inflammatory comments and reply to the people who have valid points. (Anything meant in humour I will reply too later individualy assuming I have time ;) )

Lehert writes

It's an interesting question, but it seems as if the poster has already made up his mind...

Well yes and no. I am basicaly making guesses given the evidence I have as to why open/free/netBSD are still coherant and developing and most importantly free projects that's why I called them theories think of it as a starting point for a discussion.

Inoshiro

Disclaimer: I have run Linux (2.0.x, 2.2.x, 2.3.x), Win2k, WinNT, Win9x, and OS/2, but I haven't run FreeBSD (although I've played with OpenBSD locally).

I must say, I don't know where you get "FreeBSD is stabler" from. On a workstation level, even Win2k is stable enough (although once you "pollute" it by installing Non-MS sw, the stability bleeds away like helium from a balloon). On a server level, I've never had a problem with Linux.

5:13pm up 66 days, 15:29, 8 users, load average: 1.00, 1.00, 1.00

It works for me :-)

Would it make you feel better if I said that VMS is stabler than the *BSDs? ;)

On a more serious note there's nothing wrong with using linux it just sacrifices a bit of functionality that it's users generaly don't need (extreme stability) for a functionality they do need (better support for a variety of hardware and ease of use)

And as a side issue I don't see why so many people took this article as a linux vs BSD article It's a total side issue. I will admit to having some personal bias in thinking that the design philosophy of the *BSDs tends to be better than the "It works for me" ;) attitude of linux. But except as a comparison of a copyleft project that is most similar to the *BSDs linux was largely irrelevant to the main thrust of my article which was that the *BSDs remain free despite not using GPL... why? and if copyleft in general was found to have no legal grounding at all then what kind of copyleft projects would be able to remain free and how?

analog

AFAIK, there are several, but other people seem to have addressed that already. I would also warn you not to assume that the only way to use BSD code and not contribute changes back is to fork the operating system; I believe Windows9x networking owes a debt to BSD, for instance...

I do believe I implied why the release of this kind of source is unneccesary for *BSDs to remain free (as opposed to the free software community in general

If it helps you get your head around it do you realy think that the BSD developers would want to use the microsoft tcp/ip stack? ;)

Complete poppycock. If the GPL is ever found to be not legally binding (and while I hear this a lot, I have yet to hear anyone give a good reason why it would happen), then it reverts to standard copyright and it will be illegal for anyone but the author to distribute GPL licensed code. Since just about every piece of GPL software I've ever seen states that it may also be distributed under any later version of the license, the FSF would only have to make the modifications needed to make it valid to take care of the problem.

I'm more worried about the entire concept of copyleft to be found to not be legaly binding under contract law... I can just imagine Bush giving this a helping hand if he is elected president... More nonsensical law have been enacted and precedents been set for me to not at least think about this is a possibility.

"it will be illegal for anyone but the author to distribute GPL licensed code" But then it wouldn't be free even tho since the author released it under the GPL in the first place it would probably remain free. (This is an important distinction)

For those who haven't figured it out yet, let me help you with a few concepts. The GPL is a good license; it's targeted at a specific process, and it does its job well. The BSD license is a good license; it's targeted at a specific process, and it does its job well. Linux is a good operating system kernel; the utilities used to flesh it out are also good. Ditto BSD. I use both daily, and I see no significant differences between them in my usage (ymmv). There may be instances where one is a better choice than the other; there are probably far more instances where either choice is just as good as the other, and it comes down to personal preference.

You're preaching to the choir here. ;) I'm more interested in why *BSD remains free when according to the rationale behind the GPL it shouldn't.

dacta

Are you trying to compare the BSD licence vs the GPL, or are you comparing the Linux kernel vs the BSD kernel?

Neither. The comparisons were only in support of the main thrust of my article

Anyway, these kind of stories always annoy me. I use to use NetBSD on some sun3/50s, so I aren't an anti-BSD bigot or anything, but I think the title of this article was misleading. I was hoping to read some informed discussion about why commercial comapnies do return forked code but it seemed to be more of a BSD vs Linux thing.

Read the actual article not the people doing the "clearly sir" flames there's 4 main points I make in why the *BSDs remain free only one makes a comparison to linux one other mentions linux in passing... If you think those points are incorrect have a go at writing your own :P Alhazred

Practically EVERY commercial Unix started out as a BSD fork. Then we have netBSD, freeBSD, BSD this, BSD that! Granted many of them are quite similar and they seem to trade code often, but since it is open source code that isn't all that surprising... In fact if you look closely at the Linux kernel you will see a few bits of BSD in there too.

Sorry my bad. I was using *BSD, BSD as shorthand for open/free/netBSD not the more ancient versions I apologise for the confusion caused.

daniel
See how annoying it is to use brackets instead of footnotes? now which do you prefer? :P
--
"kraant, open source guru" -- tumeric
Never In Our Names...

Re: BSD and commercial Forks (none / 0) (#28)
by pwhysall on Wed May 03, 2000 at 07:59:08 AM EST

<blockquote type=cite> Would it make you feel better if I said that VMS is stabler than the *BSDs? ;)

On a more serious note there's nothing wrong with using linux it just sacrifices a bit of functionality that it's users generaly don't need (extreme stability) for a functionality they do need (better support for a variety of hardware and ease of use)

Don't be confusing stability and functionality. They're not the same, and they're not interchangeable.

And FWIW, VMS is more stable than any other sub-mainframe OS, OS/400 notwithstanding :)
--
Peter
K5 Editors
I'm going to wager that the story keeps getting dumped because it is a steaming pile of badly formatted fool-meme.
CheeseBurgerBrown
[ Parent ]

I use FreeBSD at work (4.00 / 1) (#25)
by imp on Wed May 03, 2000 at 02:01:44 AM EST

I've worked at several places that use FreeBSD in their products. why do they do this? There are several reasons. First, the BSD license allows us to make parts of the system proprietary. This is critical in any business interprise that uses cutting edge technology. With Linux you have to hope that Linus' pronouncement of loadable modules is good enough to protect you legally if you chose not to release part of the system. with FreeBSD (or NetBSD or OpenBSD) you don't have that worry.

There are a lot of other reasons as well, but I'll not get into them here.

Both places that I've worked (Pluto Technologies and Timing Solutions) have contributed back to FreeBSD most of their changes to the base OS. Why did they do this? Common business sense. The common base OS is what enables these companies to make their money. It isn't a compeditive advantage to fix bugs. It is in the company's best interest to contribute back changes to the base OS. It lessens the support load on that company if all their tweaks (or a lot of their tweaks) are available out of the box when the next release rolls around.

There are several commercial closed versions of FreeBSD being shipped today. You don't necessarily get the source when you get a box, but there may be tweaks not in the main source tree. The way the FreeBSD project is setup encourages people that do this to contribute back to the FreeBSD base those changes that are mature, stable and tested. This limits how far these forks go away from the FreeBSD tree they came from. They are close enough that people just say that they are FreeBSD and no one notices.

This is a different balance between the carrot and the stick than the GPL based systems have. In the GPL model, you get a lot of the stick (eg release the mods or else). In the BSD model you get a lot of the carrot (package the source in such a way that it is acceptible to the FreeBSD project and it gets integrated back, making your future workload less). Companies that understand this do well, reduce their long term maintanance cost and they don't have to worry about the stick beating them up for those parts of the system they have to keep proprietary.

Others may disagree with this approach, but I like it. I'm biased however. I'm a FreeBSD committer, the FreeBSD Security Officer and devoted user of FreeBSD.

Warner Losh

Re: BSD and commercial Forks (4.00 / 1) (#26)
by el_vez on Wed May 03, 2000 at 06:38:16 AM EST

i would say that NFR IDS is a commercial spinoff
it uses OpenBSD, but doesnīt mention it anywhere, except on boot.
(ok, i could be wrong. i didnīt actually read the license agreement :)

dacta wote:

"...benchmarks back from 1996 (Linux 1.2x), that compared
Solaris(x86), FreeBSD and Linux did show that FreeBSD had faster networking"

Iīve read somewhere that from kernel 2.0 and on, there isnīt much diffrerence between
BSD and linux networking in terms of speed



Re: BSD and commercial Forks (none / 0) (#34)
by Anonymous Hero on Tue May 09, 2000 at 01:16:41 PM EST

Running Netscape 4.5 on FreeBSD 3.1 with fonts set to New Century Schoolbook 12 pt and Courier 12 pt a lot of the text is too small to read. I think you would have a nicer site if you would let me *** The User *** :-) :-) have the text be the font size I want. ---------------------------------- I would rather scroll than squint. ----------------------------------

BSD and commercial Forks | 34 comments (34 topical, 0 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest Đ 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!