Hacker News Reader

Comment by musicale | Hacker News Reader

I don't entirely understand why Linux took over and left BSD as an also-ran (except for Apple which uses BSD.) They're very similar in terms of core functionality and there was even Debian/kBSD that ran the Debian userland on a BSD kernel.

Was it the specter of lawsuits (which ultimately had no effect?)

Was it the "cathedral vs. the bazaar" (which didn't really cause that much of a divergence between *BSD and GNU/Linux distros in terms of core functionality?)

Was it GPL vs. permissive licensing (which seems like it would favor BSD in terms of adoption?)

View on HN

My theory is it was a combination of things. From my experience, Minix was the first Unix-like system I was able to run on my PC. And Linux was sort of a spiritual successor of Minix.

Then there was the fact that Linux was just a kernel, and needed a distribution to put together a working OS. Whereas BSD already was a full system. Now that would seem to favor BSD, however the userland for Linux was the GNU tools. Where BSD userland was mostly identical to other Unix systems (seeing as half of them were BSD based), GNU was (in my recollection) mostly an improvement over the standard Unix userland utilities. Furthermore since Linux was only useful if you had a distribution, a lot of distributions popped up. Now you could have a situation where BSD was more popular than any given Linux system, however the quantity of all the Linux systems ended up being more than BSD, with a lot more innovation from all the distribution creators. Each distro could specialize in a given area, you could have hobby-oriented ones, enterprise targeted, embedded... you wouldn't get that from any of the BSDs as they wouldn't have the resources to branch into all those competing specialties.

Exactly this, and then all those hyper specific pieces of Linux software would get picked between distros and eventually became common, and well, that's history. I got into Linux after it had beaten out BSD for popularity, but having ran into the BSD versions of some utils on macOS, yeah, the GNU variants are far superior in almost every regard. Darwin's variant of dd still doesn't even accept the status=progress flag(or any equivalent), for instance.

That’s valid in many cases, but the dd example isn’t really fair, because you’re comparing modern GNU to ancient BSD. I don’t know how often macOS pulls from upstream FreeBSD, but I will say that relatively recently I saw a man page in macOS that still referred to ARPANet…

Today’s FreeBSD does support status=progress as can be seen here: https://forums.freebsd.org/threads/watch-status-of-running-d...

MacOS only pulls fragments from Free/NetBSD, and in many places it's still OSFMK showing up on the surface, so sometimes you find out that you need to program as if you were on BSD 4.3 Reno still...

And that is enough to be UNIX (TM).

Ah, I was unaware (but should have assumed) it was an ancient version. Sadly, though, that's the only exposure to BSD I and many others have.

Press ctrl-T while dd is running to sent it SIGINFO. IIRC dd status=progress was added so that dd would have something like ctrl-T on non-BSD systems. Linux lacks SIGINFO so it isn’t very widely known, nor are there many 3rd party programs that support it.

On linux either SIGUSR1 or 2, I forget which, will do it too, it's mostly just an annoyance with there not being a flag for it. The only signal I can reliably remember is to quit programs...

The distribution thing was key because a lot of people would not have had networked computers at the time. The first version of Linux I installed was Slackware. I downloaded 27 disk images for 3.5" floppies at the university on one of the handful of computers we had with an internet connection. I then biked home to install it. One of the disks turned out to be corrupted so, I had to take the bike back to fix that. I had no internet at home. That PC never got connected to a network. Slackware was great. It worked. I was able to get it going with the bundled documentation and some howtos that I printed at the university. Without that, I would have gotten stuck quickly.

Later, CD-ROMs became popular and I bought a red hat CD in a store. Cd burners were not that common at the time and I did not have access to one. Also, 700MB was a lot of space at the time. My hard disk was smaller than that. Red Hat came with a lot of software (more than I could actually install) and that quickly became the key selling point for distributions: lots of software and a convenient way to get it installed by people that did not necessarily have a lot of unix experience. The competition between Linux distributions was fierce. You had Slackware, Red Hat, SUSE, Debian, Mandrake, and more. Some of these were commercial companies that were making money from selling CDs. They had budgets for marketing and distribution, they payed people to write nice installers for them, they spent a lot of time testing things and fixing things that needed fixing, document things, offer support, etc.

BSD did not really compete much in that space. I never actually used it so I can't vouch for how easy or hard installing that would have been but I bet it would have been harder to install and probably did not have the same level of hardware support for the simple reason that far less people used it. I don't think I ever saw any BSD cds in the store. Was anyone even selling such cds commercially? They had the software but they lacked the distribution.

386BSD vs Linux came down to a single thing for me: Support for MBR partitions. I was aware of both systems from the beginning and started to prepare trying both of them from around January 1992. What I had available was a 486 PC with OS/2. So in the end it was quite simple: With Linux and its support for partitions I could use that 486 without losing OS/2 (this was at work, after all, I couldn't ditch that system - I used it for production). I first booted and executed Linux from a floppy, and soon I could install Linux on a separate HD partition (and still boot from a floppy, for a while).

But 386BSD needed to take over the disk, and that was out of the question. That was the showstopper for me.

And shortly after that, my x-terminal broke and I switched to use that particular 486 PC as a replacement, a hack version of X11 came around early so that was actually possible. I was a little involved in testing new drivers and filesystems and so on, and by that time I was firmly into the Linux camp and have stayed there since. My original Unix experience was mainly from SunOS.

More likely it was device support.

You could get something like the Yggdrasil distribution that seemed to support a pile of video cards, network cards, CD-Roms etc. that were hard to get working on 386BSD. You have to remember that for most people, getting X up and running was a real effort, let alone patching the kernel to support whatever weird CD-Rom you might have.

I picked up 386BSD first, but quickly switched to Linux (around '94 I guess) simply because the community had more of a stake in Linux for providing support for different devices.

Ding, ding, ding! Winner, winner! Chicken dinner!

It's about the drivers. Getting traction with a new OS is all about driver support. Which is what makes launching a new OS hard, assuming your goal is to be popular. You have to inspire an army of driver writers.

But was that better driver support simply a snowballing consequence of Linux already having been ahead by then? If so (and IDK if it is so) then the original explanation would have to lie elsewhere.

There was quite a bit of weirdness surrounding the BSDs back then due to AT&T code that was replaced, it kind of became moribund after some dispute over patches that I didn't follow too closely, then those patches became forks (FreeBSD and NetBSD). By the time the smoke cleared, XFree86 was ported to Linux and the general perception was the Linux community had their stuff together, while the BSD guys were fighting over nothing.

I don't think it helped that the BSD distributions were very opinionated about their userlands, where in Linux you had lots of choice.

In hindsight, it seems to me the BSD guys simply blew their shot by infighting at a time when early adopters were clamouring for a Unix like, and didn't care about the BSD purity war.

That was where the snowball started. Linux had no real advantage other than the cooperative nature of the community.

Sounds like it's getting back to people then. It still seems like a strange puzzle to me.

But it seems plausible that someone, somehow, managing to get one important driver into Linux that was missing from BSD might have helped to get the ball rolling, enabling "Linux has better driver support" to become a self-fulfilling prophecy, encouraging driver writers to write drivers for Linux (rather than BSD) and for Linux to include them. Presumably BSD eventually largely caught up, but by then the race had already been lost.

It would definitely have been something like that. There weren't an infinite number of devices for PCs back then, but every one enabled more people to run Linux. Would be interesting to try to track it back to see where exactly interest really started taking off.

For me personally it was the Panasonic/Mitsumi interface CD-Rom on the Sound Blaster cards. Vague and possibly incorrect memory suggests 386BSD (and maybe FreeBSD/NetBSD) you could get an expensive SCSI cd-rom working, but not the Panasonic/Mitsumi ones that were vastly cheaper and more available.

Because the internet simply wasn't available in a form useful for downloading a distribution, CD-Roms were the only practical way to install a Unix-like. Cover discs, or Slackware, or Yggdrasil were available at your local book store.

The MPC[0] standard was kind of on it's knees during that period, so it was pretty hard to keep up with the variety of hardware on desks at the time.

This old usenet post from 1992 [1] has contemporary reports from where 386BSD was competitive, but the general consensus even then was "Linux works on more affordable hardware" so the problem must have existed before that post.

edit: [2] This salon article seems to point the finger at the Jolitz's approach to development.

[0]https://en.wikipedia.org/wiki/Multimedia_PC#:~:text=The%20Mu...).

[1]https://www.linux.co.cr/free-unix-os/review/1992/1102.html

[2]https://www.salon.com/2000/05/17/386bsd/

The insanity of how they'd take Binutils/GCC and split it up like a BSD tool, and twist it unrecognizably was just so odd.

I'd been collecting old BSD stuff from the era, and even 4.3BSD/Mach is so weird as well, although Tahoe i386 was doomed for not having source available. A larger issue with the platform as it wasn't open and code wasn't contributed back.

Also, better documentation (the HOWTO series), better distribution (plenty of CDs) and Linux installfests.

The first Unix system I ever used was a 386BSD. The first Unix system I could ever get my hand on by downloading or buying disks was Linux. From the get go, Linux was easy to obtain. I of course preferred the BSDs since that's what I was used to, but just being able to hack on Linux 24/7 from my bedroom without having to share the system with hundred others, and having root access without having to hack someone's else system just allowed me to dive deeper than the BSDs. It was easier to find other folks going through the same and targeting everything first for linux before BSD. It created a tremendous fast feedback loop, easy to obtain, more programs available for it.

I was there for Minix and Linux, I didn't get my hands onto 386BSD at the time (nor knew it existed), only a bit later FreeBSD/NetBSD and eventually OpenBSD. Hence my question: Was 386BSD somehow harder to obtain than Linux?

I guess it was just because at the time, where large transfers over modem were entirely impractical, Linux distributions tended to pop up "everywhere", in book stores as part of books or by themselves, in computer magazines, in rather "micro" distributions on floppy disks etc.? While I guess 386BSD's distribution channels were less fanned out?

I sourced my original Slackware/Redhat/Debian CD image from Infomagic back in the 1990's. This was a common way to get the install media before dialup speeds hit 38400 and beyond since downloading a whole distro was so slow and would require so much disk space.

It looks like they had a seperate BSD image at the time which had the BSD's on it. So probably not much more difficult to obtain than Linux by 1995.

https://archive.org/details/infomagic-BSDisc-1-1-93

I could access it through network via michnet back then. I had no idea where to obtain one, and I think when I did finally find one for sale through some computer magazine, it was in hundreds of dollars (easily $500+) when I barely had pennies to rub together as a high schooler. When I got Linux, I put money together with friends to get a CD. Our major expense was really buying pack of floppies to transfer and install it since we didn't have a CD drive then.

The first Unix system I could ever get my hand on by downloading or buying disks was Linux.

I think this is the reason for the early relative popularity of Linux compared to BSD. We have to remember that in the early 90ies, most people did not have internet at home or at least not broadband internet. I got my first Linux CD-ROM in 1993 and started using Linux in 1994, because a lot of computer magazines came with Slackware Linux. It was also relatively easy to get Linux distributions cheaply on CD-ROMs, e.g. through the InfoMagic Linux Developer's Resource. I only discovered FreeBSD probably through some Linux documentation and for me the only way to get it was to buy a relatively expensive FreeBSD CD set from Walnut Creek (remember, no broadband internet). When I finally got a FreeBSD set in 1996, I was amazed by its consistency, documentation, and quality. But by then, Linux already had all the attention.

Why magazines shipped Linux? I don't know. Maybe some were afraid to ship BSD because of the AT&T lawsuit. But I think Linux was also the better story: 'a student hacker somewhere in Finland writes a kernel and takes over the world/is going to break down the Microsoft empire'.

It's difficult to say. In some ways, the GPL reassured anyone participating in the Linux ecosystem that no one company was going to make a play for the whole thing by hiring up key people and running with it as a proprietary system or some kind of "open core" thing where much of it was free but the fancy bits were proprietary.

There was some crossover between open source *BSD and a closed source OS called BSDi. So this fear was not totally unjustified.

Also people were still burned from the "Unix Wars", where closed-source BSD variants played a major role. So there was the feeling that GPL would discourage incompatible forks.

As if modern Linux distributions are super compatible among themselves.

Aren't they? Not binary compatible, but you should be able to compile more or less the same software on any GNU/Linux unless you're missing dependencies.

Packaging, file locations, init systems, desktop stack in use, ...

The answer might be it was easier to duel-boot DOS with Linux in the early days (something about partitioning). However the takeoff point probably wasn't until a few years later when Linux 2.0 was the first with SMP support. Many other reasons contributed.

You could boot Linux from DOS back then, by using LOADLIN and relying on UMSDOS support which could manage a Linux-compatible filesystem within a DOS directory such as C:\LINUX\. No need to mess with partitioning at all. With some added prompts in AUTOEXEC.BAT, you could basically turn even plain DOS into something as powerful as a modern UEFI boot manager. Of course this stuff became less feasible after NT-based systems took over, making conventional dual boot and separate partitions the best choice overall.

I'm 99% certain you could boot BSD from DOS back then too. 386BSD 1.0 contains a "boot.exe" which works with FreeDOS -as long as you don't load anything else!

https://github.com/386bsd/386bsd/tree/1.0

I think that FreeBSD 2.0 may have used the boot utility as well -but I'm not sure.

I remember when I encountered Linux in the late 90's there was a distro that sat on MS-DOS; I think it may have been "Monkey Linux" ( https://projectdevolve.tripod.com/table/descript.htm )? It wasn't Slackware -but I pretty quickly found slackware and began using "zipslack" (https://en.wikipedia.org/wiki/ZipSlack) before diving in and doing a real Linux install.

So by 97 or so Linux was figuring out more sophisticated ways to co-exist with DOS if not Windows95. Probably had been for a couple years before that -but 97 is when I found Linux so I don't know.

And yes, by the time XP was released Linux was popular enough that people didn't bother making kludges like that any more. It probably didn't help that vfat and fat32 were different than the old umsdos file system so that the drawbacks out-weighed the benefits.

Yeah, BSD was still doing its funky "slice" subpartitioning scheme instead of using MBR partitions directly (and for the next two decades as well). Linux has always just used MBR/EBR partitions (prior to GPT).

Yes, exactly that. I wrote another reply about that. Running 386BSD on the PC I had available meant overwriting the disk, and that was not an option (it ran OS/2 and had to continue doing so for a time). So that's why I ended up with Linux (early 1992). I never had thoughts about GPL vs BSD at the time.

The autocorrect of dual booting to duel booting really encapsulates the spirit of the era. At that poit tbe #1 feature of OS/2 was the boot manager and the ability to boot into logical drives.

For many years the BSD tcp/ip stack was by far the best. Linux lacked viable networks far into its infancy. So I too expected BSD to reign Supreme.

For the life of me I cannot explain it, because the rise of distro carried Linux up the foodchain to mainframes and down to replace vxworks and ultimately dominate phones to spaceships.

Maybe it was people more than software?

Maybe it was Linus Torvalds' famously friendly personality. :)

Devils advocate. He did realize (eventually) how toxic it was and has taken steps to better himself

People are overly sensitive. Linus was cool and very friendly and nice person most of the time on the mailing lists. Very very rarely he writes those furious replies, but honestly, it was very fun. I just don't understand why people are so much butt-hurt.

It's literally a different generation; and people appear to bruise a lot more easily at this point in time.

So yes, your first sentence answers your last one. :)

The little interaction I had with Linus was friendly and technical (the only direct contact I had with him was a short email conversation where I provided some details about tape SCSI APIs. If there was more I don't rememember it). Anyway, I followed the Linux mailing list from when it started and for many years, and I read everything (including all the patches up to around kernel 2.3.51, if memory serves). I never saw anything I reacted negatively too, from his side. Quite the opposite.

Can't edit.. that "too" in the last sentence should be "to" of course. Hopefully nobody got confused.

Linus strongly disagreed with your comment, and even took a leave of absence to fix the problem.

Devil's Devil's Advocate (God's Advocate?): his strictness, while perceived as some as aggressive/toxic, is the main reason why the Linux kernel code is high quality.

Contrast with many other open-source projects which allow the product to suffer rather than offending the contributors.

Yep, I remember comments like "You know Linux is good shit because Linus won't tolerate bad patches."

There was a "mailing list alpha male" culture back then which inspired a lot of people to join up. Maybe some people didn't like it, but in Linux's case, the cult of personality attracted a lot of followers.

Modern Linux dev is mostly a 8x5 job now, so that mercurial king stuff became really tired.

Publicity wins over utility. The flamewars between -baum and Torvalds brought alot of attention to Linux.

So it was partially Torvalds' (and Tanenbaum's) friendly personalities!

I'd forgotten that Linus started out with Minix.

People probably helped - I installed Linux for the first time back in 1994, and that was because a couple of local guys introduced me to it.

I did download slackware floppy sets, over university internet connection, a few times. But it wasn't too long until there started to be Linux magazines available, and they'd often have a CD-ROM attached with a new distribution to try out.

(I actually remember having to buy a CD-ROM for the first time somewhere around 95, because my computer didn't have one and installing with floppies was getting very old by then.)

There are rumours that the mstcpip stack had large parts of the bsd stack in it. I cut my teeth in windows networking using Windows for workgroups, with a novell tcpip stack and a 64kb daemon Internet Internet connection over isdn. Having that network connection meant I could (1993) get some linux/freebsd setups, but I would have to leave the ftp connection running all night. I then diverted off full time into freebsd when I joined Yahoo.

> Was it the specter of lawsuits (which ultimately had no effect?)

I think that this was mostly it.

> Was it GPL vs. permissive licensing (which seems like it would favor BSD in terms of adoption?)

Many folks prefer to adopt & contribute to an OS when they know that they are legally protected from others taking their contributions proprietary.

I think that another factor, sadly more important in the 90s than now, is that the GNU userland was far more user-oriented than the BSD userland. Back then, Unix users actually used the command-line tools, and GNU was in a lot of ways more pleasant (if also, um, a bit scattershot and buggy).

Yeah, the lawsuits hamstrung BSD right as Linux was taking off. It may have been a motivation for creating Linux in the first place.

There's some perspective of Torvalds on 386BSD in this 1993 interview: https://gondwanaland.com/meta/history/interview.html

Applicable part:

"when I started on Linux it [386BSD] wast available (although Bill Jolitz series on it in Dr. Dobbs Journal had started and were interesting), and when 386BSD finally came out, Linux was already in a state where it was so usable that I never really thought about switching. If 386BSD had been available when I started on Linux, Linux would probably never had happened."

I don't think the lawsuit played an issue.

That explains why Linux was started, but not why it won.

Yes, I just replied to "it [the lawsuit] may have been a motivation for creating Linux in the first place".

As for why Linux "won"; I think the low bar to contributing in the earlier days helped a lot; BSD tended to be a lot more bureaucratic. And there was also some amount of drama and splits in the early 90s, which probably didn't help either.

IIRC patent battles in court with AT&T and code removals from the source which kept 386BSD from getting off the ground quickly..

NetBSD and FreeBSD came out in 1993, the same year as Slackware. The BSDs seem like pretty complete Unix systems and are still in use today, but they are also-rans compared to GNU/Linux.

FreeBSD had to reboot the project after some lawsuits -- FreeBSD 2.0 is a different codebase from FreeBSD 1.1

It was a huge setback for every kind of adoption

I didn't actually know that the FreeBSD project had to "reboot" - this is very interesting and adds to the "lawsuit" theory.

i think it was timing personally. i was an old sun guy and really didn’t like the solaris shift. there was about a onetk two year gap where 386bsd was kind of half baked but you could get a .9 slackware install running pretty easy. when freebsd released i switched back. not sure if that’s actually correct, but is the way i remember it

Yeah, the story I read was Jolitz went AWOL, so there was a contentious period when Net/Open/Free split and self-organized and started releasing again.

I was fortunate to have a 386 (bought by my parents for college) with me in the early 90's (there was one other floor mate in college I remember who had a 386 at the time).

I came from BBS's (c128 with modem) and had no notion of BSD vs linux or any of the legal battles of the time.

I can only put it down to "marketing" or some type of "fate" that one day on usenet I learned that I could install "linux" a type of "Unix" (the "same" thing I was using on the workstations in college) on my 386. It was Slackware that I had installed at the time and that's my story.

I eventually did learn about BSD later but it didn't really "click" with me when I tried it. I think maybe Linux was the "zeitgeist" of the time and information on it was just readily more available (at least for me).

I think you've covered most of the points -particularly the lawsuits and the development model (why code where you're not wanted when Linux is just right there ?).

I suspect that BSD may have been a bit more flakey than Linux as well. I've never tried it on bare metal so I may be wrong; but in emulation 386BSD seems very difficult to get working properly.

I wasn't there, but based on that experience I'm guessing that Linux got to a more usable point first?

Another point is that Linux simply ran on a wider variety of commodity hardware earlier than BSD did.

My first exposure to Unix was Sun Solaris at work. I was very interested to install it on my home PC but failed to. During my searches to install Unix, I came across Linux and joined some user groups. At the same time, one of the magazines included an installation CD that basically just worked. Once I resolved the modem issues, there was no looking back. Did try Freebsd later but ended up going back to Linux and have stayed with it since.

PlayStation OS is based on BSD, but as the license goes very few see the outcome of that.

Observing this at the time, it was entirely the lawsuits.

Reply on news.ycombinator.com