Slashdot Mirror
The Roots Of BSD
drix was the first to write in with this "Standard fare roots of the BSD/hacker movement piece over at Salon. The picture of the FreeBSD devil guy is pretty cool." This is actually another chapter in Andrew Leonard's Free Software Project online book. Well written, but occasional errors (FreeBSD and BSDI have not merged, for example) cast doubt on some of the facts. Informed comment from people who were there would be appreciated.
Isn't is illegal to try and get root of a BSD system (at least, one that isn't your own)...
But there are four things that the Salon article missed that I'd quite like to see a "central" presentation on:
Back in the late '80s, I saw some folks in Ottawa playing around with it, with some paranoia going on over whether the lawsuit-happy would be stamping it out.
These started as varying approaches to the use of the "ashes" of BSD 386. I'm sure there's more of a story to it than that.
Perhaps with further commentary as to the "splits."
Which had pretty big plans, notably including filesystem efforts (journalling FFS, tmpfs, and such).
It's not clear what, of the final efforts on the academic side, have headed into active versions.
If you're not part of the solution, you're part of the precipitate.
beta survives as ``betamax'' and is used heavily in the broadcasting industry, and DATs are still used by tapers and for mastering in the pro-audio world. the algorithms developed for digital radio survive in MPEG2 layer III...
and btw... walnut creek and BSDi merged back in March, so FreeBSD is effectively backed by BSDi now.
NetBSD: the cathedral vs the bizzare.
To which CSRG work are you referring here?
STREAMS is a System Vism (influenced by the "streams" done by Dennis Ritchie at Bell Labs Research.
...as well as merging a fair bit of SunOS 4.x (the VM system, the VFS layer, the NFS code, and the dynamic linking mechanism, for example, although the versions in SVR4 had some additional changes, including renaming the as_hole() routine as_gap(), as I remember - yes, as_hole() was intentionally called that...) in as well.
OK -- I did goof. I should have written that BSDi had merged with Walnut Creek CD-ROM. I just got off the phone with Kirk McKusick and he straightened me out.
There is the expectation, according to McKusick that there will be some merging of the code bases between BSD/OS and FreeBSD. McKusick says that the source code to BSD/OS will be made available to FreeBSD committers, who will be able to take pieces of code, and once having integrated them into FreeBSD, change the license to a BSD-style license.
I'll make a change in the text ofmy story and log the correction in my revision log.
Editor, Salon Business & Technology
Salon.com
On the other side of things, Linux might not exist
;)
:), but it's a nice learning tool (even more so than minix, IMHO), and it was/is fun working on it. "
either if the FSF hadn't gone with Hurd running on Mach, but something "simpler".
The snowball might have started rolling several years earlier. Whether it would have ended up in a better result or not is impossible to say.
Same thing with BSD really; without the lawsuit it would have been quite possible that Linux would have a BSD-based TCP stack (sure there would have been licensing issues, but at that point they would have been pretty easy to solve,
Linux didn't even start as GPL)
Or maybe Linux wouldn't have started at all and we'd have a GPLBSD for those who don't believe in the BSD license (with a GPL-style license that allowed the BSD advertising clause)
Btw., you just gotta love embarassing quotes from the past
"/I/ think it's better than minix, but I'm a bit > prejudiced. It will never be the kind of professional OS that Hurd will be (in the next
century or so
- Linus in December 1991
Comment removed based on user account deletion
Actually, this process is pretty much Darwinian. In the context of biological evolution "best" really has no meaning other than "survivors". The best survive, because they're best at surviving. Tautological, but there you go.
So publicity is a survival mechanism that companies use in the Darwinian world of the marketplace. There are many different successful strategies, and many niches to occupy. Just like in biology.
Not that there aren't differences between technological and biological evolution. Heredity isn't quite as clear in technology, for example.
"I believe that the cult of the particular brings only death - for it bases order on likeness." St.-Exupery
Well, you got closer, but still no cigar.
System V R4.0 was the result of merging a few parts of BSD into System V R3.2 plus providing a compatability layer for much of the rest. It was done by AT&T USG (Unix Systems Group) and Sun (under contract). (STREAMS was already part of System V R3.2, and actually had grown out of the "packet driver" work Bell Labs had started years before Berkeley sockets; Dennis Ritchie himself had devised STREAMS as a way of accommodating a variety of different proptocol stacks and network drivers, though he wasn't happy with the USG's adaptation of it.) This formed the basis for Solaris 2.x. But the USG continued, adding security and SMP features to System V and improving the VM system, resulting in System V R4.2. There was actually a considerable divergence between this system and Solaris 2; Sun did their own SMP and security enhancements starting from the System V R4.0 code base (which they ultimately bought the rights to). Unix and the OS development part of USG was sold to Novell, which marketed System V R4.2 as "Unixware." It was, IMHO, a much better system than Solaris (at the time), but Novell simply couldn't stand to support a product that competed with NetWare, and after a few years of letting Unixware wither on the vine, sold what was left of the USG and System V R4.2 to SCO. SCO wisely dumped their own System V R3.2-based technology as fast as they could, but by that time Linux was becoming a strong competitor--and we all know what's happened since.
I've used professionally every product mentioned above (with the exception of NetWare), and I've used BSD from release 2.4 (which ran on PDP-11's).
It's sad when posts composed of guesswork and half-truths get moderated up by moderators who have even less of a clue. Like one of the earlier posters on this topic, I'm just about ready to abandon Slashdot as the noise has simply drowned out the signal at this point.
Sun became a company in 1982 (when Bill Joy left Berkeley for Sun). BSD was there Waaaaay before that.
SunOS 4.1.x and before (Solaris 1.x) descended from BSD. SunOS 5.x and above (Solaris 2.x) is SVR4, which was developed mainly by Sun and AT&T by merging the best of both codebases (BSD and SVR3).
Get your facts straight before posting, please.
-Kevin
My posts don't reflect the opinion of my employer, and my employer's opinion doesn't influence the content of my posts.
The missing piece in aleonard's chapter is about the erratic course AT&T took regarding UNIX during the 1980s. When it finally decided that it was worth promoting as a salable commodity, licensing and use restrictions tightened considerably, but they failed to extend the OS in the directions that it needed to go to meet the burgeoning needs of local and connected networks (which then became known as the Internet).
Perhaps most important was the development of NFS, which was introduced formally by Sun but based directly on work by the CSRG. Another important building block was Berkeley sockets/STREAMS. These are the things that distinguished Berkeley from AT&T UNIX in the mid-1980s and caused sysadmins who were not encumbered by AT&T purchase requirements to go with the Berkeley flavor during that foundational period. In Cuckoo's Nest, Cliff Stoll alludes to some of these differences from his work as a part-time sysadmin at LBL.
Finally seeing the commercial potential in the late 1980s, particularly driven by corporate markets moving to Oracle and other UNIX-based business applications, and the growing importance of Sun, Apollo, HP and other entrants in both the server and workstation markets, AT&T was faced with two facts in its pursuit of a payoff for its languishing UNIX product: (1) its inability to succeed in the retail systems market against more experienced competitors like HP and more eager ones like Sun; and (2) the ongoing breakup of the Bell System under MFJ III.
Consequently, AT&T sold UNIX off to Novell, in one of the classic examples of the "greater fool" theory of marketing, since Ray Noorda and his merry band in Utah had not Clue 1 about what to do with it. Novell's Univel subsidiary was set up to put together a repackaging called Unixware which never really got a foothold. The only good thing about all this for Novell was that they eventually enticed Eric Schmidt over from Sun to run the company. Schmidt, Berkeleyite to the core, flung the doors wide open to IP and eased away from IPX, and Novell has been able to find a role in the modern corporate market for servers and directories when it was almost guaranteed that the company would sink without a trace in the mid-1990s otherwise.
But the fight over the intellectual property rights of the AT&T and Berkeley flavors was heating up even in the late 1980s. Probably the best coverage of the ensuing battle was in UNIX Review columns over those years, and I hope aleonard will review those as his book project goes forward.
Through a rather complex and messy process, there was a showdown between Novell and UC Berkeley, the very end of which is described in the FreeBSD handbook capsule history.
For about 18 months, it was entirely unclear whether an open UNIX would be possible; this was the period when 386BSD was basically frozen and Linux and other now-forgotten "free U**xlike" things were being worked on. And the reason that those were happening was the continuing expansion of the DOS and Windows 3.x market which brought about decreasing costs and increasing capabilities to desktop machines. Desktop UNIX on the Intel platform only really became usable with the faster late-1980s 80386, and was still basically a toy before 1990; most desktoppers were running SunOS 4.x boxen or maybe AT&T, HP or Apollo workstations to do local development and the very earliest forays into what became ISPs.
The legal battle over the status of UNIX allowed a critical mass to converge on development of Linux, which was far enough ahead in 1994 that even my Bay Area friends were probably installing it more than the BSDs (with the exception of Berkeley grads of course!). The "distribution" concept promoted most effectively by Yggdrasil and Slackware played a major role in this, because small-PC UNIX players no longer basically had to be kernel hackers by necessity.
There's also no question that the *BSD groups develop with more of the "cathedral" mode than the "bazaar" mode, but that may be an appropriate niche-ification as we go forward. Certainly those of us with more an affinity for the Berkeley flavor will continue to lean toward *BSD than Linux with its stronger SYSV heritage. But in reality, the differences really are a matter of preference, not capability.
Bill Gates Is My Evil Twin.
Let's just say that there's been a lot of claims made, and no real decisions yet, shall we? You know what media coverage can be like, even friendly media coverage.
My blog: http://www.seebs.net/log/ --- My iPhone/iPad app: http://www.seebs.net/seebsfrac/
See also Keith Bostic's interview at O'Reilly's FreeBSD DevCenter. He is one of the BSDi founders and runs Sleepycat Software (BerkeleyDB - the base for Perl's DB_File Module).
--
>The Chapter Twenty Years of Berkley Unix: From
/catalog/opensources/book/kirkmck.html
>AT&T Owned to Freely Redistribuatble by Mcusick &
>Co. does a very good job at describing the
>movement and development from the AT&T based
>Berkley Unix to the Free versions that we have
>today.
The full text of this article is available at http://www.oreilly.com
go to rpm.org and get your head out of the sand. all the code is there along with a detail spec. RPM is NOT a proprietary scheme - its just a tar file with a specfile attached. i suppose you consider .deb file format to be proprietary too.
It's been there for a while. I'm just browsing through 4.3BSD-Reno source. Link here.
Bill Joy was some kind of freakin genius and since he had access to AT&T UNIX he was able to run with it and create a greatly improved version. But then he dishoners the open source methodology because most of the eyes looking at it wouldn't see the nastiest bugs and wouldn't provide code that was good enough. That statement really illustrates how he doesn't understand open source (in the esr nomenclature since we are talking about the "many eyes" benefits in his essays) because the "many eyes" theory is NOT that thousands of eyes will together find the bugs... it is to make it POSSIBLE for that one genius out there who can fix the bug to do the job... just like Joy was that genius who could really get his head around BSD UNIX code. When I heard him give his opinions on open source at a Sun/Java conference last year in Seattle, even after going over how he was able to improve UNIX in the '70s I was really put off.
Agreed. I prefer this fox logo:
http://www.early.com/~emackey/linux/
At least a fox is agile, fast, and cunning. A penguin is just fat and slow (well, except in water), and just hobbles about. Yeah, that's what I want to think of my software as...
Please adopt the fox...hey, maybe someone should make a distro just to gain popularity for this logo...
It's 10 PM. Do you know if you're un-American?
WTF are you talking about?? That's a /guy/ fox...
It's 10 PM. Do you know if you're un-American?
Well, no.
SunOS came from BSD. SunOS4 still pretty much looked like BSD.
Then Sun came out with Solaris 2 (and renamed SunOS4 to Solaris 1 after the fact) and began the
merger of BSD with SVR4.
Now SVR4 was the last gasp of the AT&T UNIX dynasty. Widely considered to be "The UNIX Standard" at the time. It, of course, was descended from SVR3 (3.2 to be specific) with the addition of some new technology from Bell Labs (STREAMS instead of Sockets--for the not-invented-here crowd).
Which is all well and good, as far as tangents go, but why do I mention it, you ask... Well the SVR3.2 (and subsequent SVR4) programming manual set was available in regular computer book stores and is written in a nice "standardsy" language. Just perfect for someone who had this idea to clone UNIX. Thus Linux has a System V base (which is part of the gulf between it and BSD).
But now the twist... Since we started the Sun, and talked about how Solaris was merging SunOS4 with SVR4... (which after the head on collision of Solaris 2.4 turned out alright from the point of view of Solaris (2.)8).
Think about this: Solaris is like merging FreeBSD and Linux! (As the BSD and Linux zealots run screaming back to their respective battle lines and hide.)
Pleasent dreams...
"What's the point of going abroad, if you're just another tourist..."
I think Leonard missed a major issue in why Linux has become "where the action is" in many ways, and that is the license. It's my impression that the ability to put one's work into software that will permanently be free is a major motivator for many people working on Linux, and they would not have the same motivation to work on BSD where their work could be incorporated into someone's closed proprietary software. And that this has helped create the great surge of development activity around Linux as opposed to BSD, which in turn is the main reason why Linux has become so popular (in spite of what may be the fact, and at least is my technically-uninformed impression, that BSD kernel still has many advantages). That is, if the best kernel won, BSD would have won, and perhaps would still win; but instead the best license has won (with the belief that the kernel will ultimately catch up). I'm an outside observer. What do those on the inside think? Is GPL vs. BSD a major (the major?) factor in the ascendency of Linux over BSD?
Yes. George Dinolt and I got our version of UNET up on his pair of Onyx boxes, connected over 9600 baud serial lines running SLIP.
John Nagle
He has no name... A lot of people think he's named Chuck though, apparently due to some misinformation from Walnut Creek. See this link
In 1985 Intel was preparing to release its first 32 bit processor. BSD's Keith Bostic was appoached by an Intel employee who requested that BSD be ported to the Intel 386. Whether through snobbery or stupidity, Bostic dismissed the request. Like his friend Richard Stallman, Keith Bostic was convinced that the future of personal computing was in the Motorola 68000 series processors. If Bostic had been a little more humble and accepted the Intel challenge, BSD would have had a six year headstart on the competition and Linux probably would never have been invented. It just goes to show that a swelled head doesn't imply extra gray matter. Quite the contrary. By the time Bostic came around the Intel side, it was too late. The lead time had been squandered and Linux was on its way to its legendary success.
Also, some things put off for 2.6 will help tremendously:
What I would like to see from the BSD community:
I'm sure I've missed lots of things... corrections are obviously welcome.
-l
Help cure AIDS, cancer, and more. Donate your unused computer time to worldcommunitygrid.org. Join Team Slashdot!
How Berkeley hackers built the Net's most fabled free operating system on the ashes of the '60s -- and then lost the lead to Linux.
:)
I don't know if "Lost" is an appropriate term. Sure, there are more Linux users out there than BSD (I am one of them.) I am installing OpenBSD on the system I am currently piecing together. If it was not for Linux I would never have known about BSD. True, if not for Linux BSD would probably be at the forefront where Linux is now.
Still, is that really important? The most important thing about free software perhaps (IMO) even more important than the "free" is the compatible file formats. BSD and Linux are pretty much cousins, file formats are not a problem. BSD is very much like Linux as we all know, but it is not Linux, and thats a good thing. Some buisnesses may not like the GPL, but they need the Unix model, and like free software ideals -just not the GPL- Enter BSD.
To all the GPL zealots out there, I think that we need both: The BSD licence and the GPL. This promotes competition, and that is good.
So wish me luck on the install!
Try to hack my 31337 firewall!
I know this is probably heresy, but I think the BSD Devil is way cooler then a freakin Penguin.
WWJD -- What Would Jimi Do?
(Smash amp, burn guitar, take home the groupies)
3COM dropped support on UNET and TCP/IP around 1983, instead pushing their own, now-forgotten protocol suite for PC LANs. We finally switched to BSD's networking on the VAX when 4.3BSD came out, and even then, it had lousy interoperability with non-Berkeley TCPs. I had to fix the thing myself, for which I got a minor mention in the 4.3BSD release notes.
The big advantage Berkeley had is that they could give their work away. UNET sold for about $5000 per CPU, just for the protocol stack.
John Nagle
The current dominance of Linux over BSD leads to the interesting thought that Linux may be doomed to the same (relative) 3rd string status as BSD eventually. Both OS's were formed and are maintained in a similar manner, and both have the same weaknesses that has been killing BSD for the past 10 years.
It really is a question of strong leadership. When Joy left the BSD movement the problems really began, and now that Linus is working for Transmeta, how long will it be before he too drops his creation in favor of newer (and much more profitable) enterprises?
I hate to defend Micros~1, but Gates' leadership is the primary reason the company is so strong. Same for Apple & Jobs. You can not have long-term success w/out leadership, and no one took over that position with BSD. If Linus goes, who will replace him?
I work for an ærodynamics company, and despite our superior product, we may soon be filing chapter 11. Why? Because our brilliant co-founder left to work for a breakfast cereal company (of all things!), and despite the new CEO, nobody can take his place as a strong leader.
The prestige of parenting a brilliant idea is wonderful, but it seems to me that most will choose to use that prestige achieved to gain a more lucrative position for themselves, dropping their creation like a dirty diaper.
I actually never realized just how much BSD has influenced the free software movement until I started noticing just how many parts of Linux systems inherit from BSD. The whole socket abstraction to TCP/IP (and other protocols) came from BSD, basic utilities like renice, write, and others as well. This may not sound like much, but you just have to read the source for things like IRC clients or other net apps to realize just how pervasive that BSD idea of sockets is. Plus, I live on renice so much that I can't imagine life on Linux without that little contribution from BSD. :-)
---
mikre he sophia he tou Mikrosophou.
There is no doubt that the importance of BSD to the free software movement has been vastly understated. What is disappointig about the article is that it focuses as much on Joy's personality rather than the incredible accomplishment that BSD was.
It proved that software projects could be distributed yet centrally managed--a fact that needed to be established before telecommuting could become mainstream. Sure, Joy is interesting, but the fact that a culture beyond open-source owes its existance to BSD is completely understated.
ByteMyCode.com: A Web 2.0 code sharing community.
Well, if you want to know about factual errors, how about the one where the slashdot article refers to the Berkeley Daemon as "the FreeBSD devil"?
:)
1. He's a daemon, not a devil.
2. He's BSD's, not FreeBSD's.
My blog: http://www.seebs.net/log/ --- My iPhone/iPad app: http://www.seebs.net/seebsfrac/
Ah, BSD, looking back it is like looking into another world altogether. I remember other products from the same period that were "the best" and yet, those too are long gone.
History looks back not on the best, but on the survivor. Beta Video tapes, DAT Audio, Digital AM/FM Radio, all have been "the best" and all have simply died.
The causes are varied but they all share a common thread. Microsoft realized very early on that if you want to survive it matters not if you are the best, but rather that everyone recognizes that you are "it". Sony blew it with Beta because they did not allow general propegation of their standard. The VHS format was given away for free and adopted instantly by the pornography movement in the US and became the instant standard.
BSD never made themselves a public entity. Linux has fought tooth and nail to make themselves visible. Outside of the computer professional field I doubt if anyone has heard of BSD, free or not.
Unfortunately, it is the public's awarness that determines a products viability, and most importantly it is the public perception that a product is "used" that makes it indeed used.
Take voting in an election as a good example. People want most of all to vote for the winner. So, whether they understand, believe in, or agree with, a candidate is moot. They will vote for the candidate that they believe will win. MS was preceived as having "won" the OS wars way back in the late '80's, even though MacOS, OS/2 and others were far far ahead of Windows 3.0.
In Summary: Publicity Pays, big time
Later . . . . . . WebBug