8:20 As a sys admin of proprietary UNIX, I'll say the decision to determine what UNIX to purchase was never difficult. One starts by asking, "What software do we need to use?" Once that question is answered, one asks, "What computers run that software?" By that point, especially for computational or database systems, the choices get narrowed down to 2 or 3 options. Then one considers things like price, availability, any available discounts, and compatibility with existing systems, processes, and employee skill. It's really a question that can be answered in 30-60 minute meeting with fellow coworkers and a couple of conversations with the technical sales staff of the potential vendors. The litigation issue was non-existent with most end-users. Most businesses were forklift upgrading their systems every 3-5 years, so they just let the computer vendors fight it out. What can I say, it's how it was done in the '80s-'00s. Linux seriously made a dent in major industry in the late 2000s. By 2010, it was firmly entrenched.
11:28. I disagree. It wasn't that customers didn't understand UNIX. Customers switched to Linux because Software Vendors started porting their code to this "new", FREE operating system that was very similar to the proprietary systems but it ran on commodity PC hardware (that, as you stated, had caught up with, and began to surpass, RISC hardware performance).
You're an idiot. Unix wasn't the be all and end all; businesses, universities and colleges were using ICL VAX IBM all with their own non-Unix systems, then came the PC on which IBM forced home users to use Microsoft Dos, Commodore's 64 and Amiga would have killed the PC but for the Taiwanese clone makers, then came Apple's Macintosh pitch that the universities brought; finally IBM began to understand that these micro processor home computers could handle the accounts for 90% of businesses so IBM let Gates work with them while their new personal computing section developed OS/2 which Gates then took and recompiled as NT ver.1 which was the start of the current generations of Windoz; meanwhile in the American wilderness Unix and ATT and BSD were doing their thing, even my idiot brother designed and began selling a 16 bit system loaded with a Unix system. The only noteworthy commercial unix success was Sun's Solaris, and the law suit did not scare businesses away from BSD, small business owners like the universities were bedazzled by Microsoft and Apple sale pitches; business owners small & large alike did not want free or open-source, they wanted a company that promised all singing & dancing software eternally serviced by thousands of wizards. The reason Linux was more popular with Windows users like yourself is because it had pre-compiled software binaries like Windows had, Windows uses understood software is magical and comes in pre-packaged boxes, they don't like and are scare of source code that produces binaries and customised binaries. Since the 1980s the unwashed public has been introduced to computing via Microsoft selling magical singing package, some of these people like yourself Microsoft didn't invent most of the stuff it sells and there are communities of people producing quality software that they have been sharing with others; you just wanted to download and use it, fair enough. BSD on the other hand was a bit scarier, its native approach has been to provide the means to compile and your choice of source code; so there's a much smaller number of home users, but there are more servers and routers and ATMs using BSD than statisticians assume, their surveys don't count the machines that are silently working.
@mfnickster9754 That's funny; meanwhile the first three or so versions of NT could run OS/2 programs, and I don't mean command line stuff, I mean desktop applications. It took a while for MS to modify their source enough to become incompatible.
@@vaisakh_km or as I've recently taken to calling it, not GNU plus POSIX. POSIX is an operating system unto itself, but rather not another free component of a fully functioning GNU system made useful by the GNU corelibs, shell utilities and vital system components comprising a full OS not defined by POSIX.
well linus torvalds didnt had the idea back then that his pet project could lead to be more than 600 linux varients mad respect to linus torvalds and the team
I believe there is an archived usenet post or something where he talks about his project, and he specifically says it is "nothing serious like BSD or GNU" or something similar.
@CrazyCanuckBiologist He also didn't think it would evolve past computers with an Intel 80386 CPU and it was "NOT portable ([becuase it used] 386 task switching etc)" and Torvalds thought "it probably never will support anything other than AT-harddisks" because that's all he had at the time... Oh boy, he was WRONG on that one! 😂 But for once, it was a good thing that somebody was wrong!
Do you mean SunOS? Because I remember the change to Solaris because it was their new hotness. A pair of Sparc 10s running as servers for a mom and pop dial-up was my start in non-windows systems.
@@meeponinthbit3466 Solaris. From ChatGPT: Solaris is a Unix-based operating system that was originally developed by Sun Microsystems (now owned by Oracle Corporation). It was first released in 1992 and was designed to run on a wide range of hardware platforms, including servers and workstations. Solaris was known for its advanced features, including its scalable architecture, support for multi-threading and multi-processing, and its robust security features. It was also one of the first operating systems to support dynamic tracing, which allowed administrators to monitor and debug running applications in real-time. In addition to its technical capabilities, Solaris was also popular for its user-friendly interface and its support for a wide range of software applications. It was widely used in enterprise environments, particularly in the financial services and telecommunications industries. Over time, Solaris faced increasing competition from other Unix-based operating systems, as well as from Linux, which emerged as a popular alternative in the early 2000s. Oracle Corporation, which acquired Sun Microsystems in 2010, continued to develop and support Solaris for several years, but the operating system was eventually phased out in favor of Oracle Linux and other products.
In 2008, a big security failure appeared in bind. At the time I was working for the biggest ISP, and bind relied on 2 Sun servers without updates because too old. I saved the day with 3 decommissonned Sun Ultra 5, up to max memory (1GB) , and Debian for Sparc64 with automatic updates
@@wisteela I had an old Apple Mac PowerBook which was all black plastic with a trackball for mouse control! One particularly interesting thing about it was that it contained an Apple branded 2.5" *_SCSI_* hard drive. Prior to that I did not even know that 2.5" SCSI drives even existed.
One, small nit-pick: macOS does not use a "BSD kernel". Rather, it is based on the XNU kernel, originally developed at NeXT and derived from the Mach microkernel from Carnegie Mellon University. The BSD part comes in as the toolset placed on top of this kernel.
@@meeponinthbit3466 Google seeks to bring the Android kernel closer to mainline, The main issue is that due to the fragmentation of SoCs and the lack of a standardized platform for the ARM architecture, vendors tend to put patches upon patches, in addition to proprietary drivers, making impossible to create a generic system image that works for all devices unlike x86.
@@Sumire973 that actually has very little to do with it from what I can see. Armbian is doing just fine with a variety of Arm chips. Arm has required vendors to support device trees for about a decade now and they fix this issue. It's a fine excuse Google can throw around to place blame, but it's old and no longer valid Also, vendor patches happen after Google has already reforked the next version from mainline and re-merged all their own changes. I think it's more to do with them wanting to keep a critical point of control over the OS and how new features get implemented. For example, if they really wanted to go mainline, then they would have ditched their implementation of the USB gadget code, and moved to the more configurable and maintainable functionfs system.
Another small nit-pick: That isn't entirely true. ;) XNU is a hybrid kernel that consists of a Mach microkernel (that takes care of stuff like kernel threads, memory management, and interrupt handling), and a BSD component (that implements stuff like permissions, networking, and POSIX API + BSD syscalls). The BSD component does not run in user-space, but is a part of the kernel itself. It is based on FreeBSD, but also contains some code from NetBSD and OpenBSD.
I'm a "senior" sysadmin by trade. Despite being in my mid-20s, it wasn't difficult to take over AIX (IBM Power) and Solaris (Sun/Oracle SPARC) thanks to messing with Linux and BSD for half of my life. Hardware is quite interesting. For instance, on POWER systems you can disable certain parts of memory if it's damaged instead of sending a technician onsite to replace a faulty RAM module. UNIX OSes such as AIX or Solaris are being decommisioned enmasse. They became a liability now that the world is moving towards containers and ""cloud"". You need really skilled people to manage those systems as well, which costs a lot more. If you've ever used OpenBSD, you would feel at home in AIX (or vice versa). Solaris feels suprisingly modern, I'd say very similar to Linux with SystemD and OpenZFS/ZOL. Funny thing is that back in the 80s UNIX was running on RISC CPUs (ppc/sparc etc). Now apple moved back to RISC (aarch64) CPUs. History doesn't repeat itself, but it often rhymes :)
Solaris is fun. Granted, I only played with OmniOS. But getting to dive into the cool stuff that Sun made really felt like the future. It's too bad a lot of that cool tech is going to waste, though it is nice to see that OpenZFS is surviving very well
Sometimes Linux knowledge can bite you when you come over to Solaris. :) Case in point, "killall". In Linux, kills all processes by name. In Solaris, kills *_ALL_* processes, even if you give it a name. Ouch!!! LOL
We're in a big move from AIX to Linux for our Oracle databases. Big reason to not renew our IBM licenses is portability toward cloud providers and consolidating everything in a x86 hardware.
I suspect that at some point Apple will start marketing their ARM chips to the server market. This might be accompanied by "headless" version of Darwin UNIX to enable them to sell at competitive rates without cannibalising sales of their premium devices.
The writing was on the wall in the 90s. Companies like Sun, HP, IBM, SGI, DEC/Compaq, etc all had huge numbers of people in house developing their own UNIX OS. It was obvious it would just make sense for them to all switch to Linux and share the work of others and no longer have to have these people all in-house supporting their own proprietary Operating Systems from top to bottom which is very complicated and expensive to do.
I remember a comment from Larry McVoy (the BitKeeper guy) saying that Linux ran faster on Sun’s SPARC hardware than Sun’s own Solaris OS could manage. Also there was a review of an Apple XServe machine (back when Apple used to sell servers), comparing the performance of MySQL on OS X Server versus the same DBMS running the same load on the same hardware, but under Linux. And Linux was faster than Apple’s own OS!
@@jamescobban857 AT&T Unix was not known for good performance. I remember, back at my own employer some benchmarks being done comparing NFS on VAX/VMS versus that on some Unix box (Sun? AT&T?), and the VAX won handily. This was in the pre-RISC era.
I first used UNIX in 1973, but the only reason my employer could obtain it was because it was an educational institution and therefore exempt from the licensing wars. We could get the BSD implementation for free to run on our DEC PDP11 computers, which it did better than the proprietary, and costly, DEC operating systems. Richard Stallman's GNU project assumed the existence of the BSD kernel, he focussed only on the user interface. Since LINUX was binary compatible with BSD all of the tools worked under Linux. Torvalds, of course, was able to exploit BSD and GNU because his employer was the University of Helsinki. He could also exploit the free labour of his students by labelling the project as a course on the design of operating systems. MULTICS was the first attempt to build a portable operating system in which the code was written in an Algol-based higher level language. At the time only IBM was big enough to afford to build its own operating system, so everybody else invested in MULTICS. Unfortunately it was not possible to build a real multitasking operating system on a 16 bit 64KB computer, which was the most that anybody except a massive corporation or government could afford. It was joked that the acronym stood for Many Unusually Large Tables In Core Simultaneously. Even most IBM mainframes could not support IBM's Operating System. The smallest computer I ever encountered running IBM/OS was 256KB and the smallest that I ever personally ran it on was 384KB and on that I could only run TWO tasks. In the 1980s I was working closely with IBM on the design of telecommunications services. To add support for the Internet IBM had to clone the support from BSD into their S/390 (S/z) mainframe and GSD (S/y) mid-range computers. In order to meet performance targets they essentially ended up rebuilding all of their operating systems as simply parallel application interfaces on top of the UNIX kernel. This explains why IBM has been one of the major contributors to Linux, not just for instruction set support. You can run UBUNTU or REDHAT on any IBM processor. Microsoft is now the only operating system provider which has NOT abandoned a proprietary kernel. This makes no sense as Microsoft makes its money off application services which couldn't give a damn what kernel they are running on top of. Microsoft should be diverting all of that effort and expense into perfecting WINE.
Nick keeps saying that "UNIX had become proprietary", which he probably means "no longer gives access to source code" - but it's important to understand that UNIX was *always* a proprietary commercial product with no free-to-distribute components. At the time, and because of its portability and plethora of processor technologies, most UNIX licenses were source licenses where customers paid for a copy of the source code and were expected to modify it to their needs - and under specific circumstances - share their changes with other licensees.
There is a direct descendant of Unix, which is illumnos, derived from Solaris through OpenSolaris. It is completely open source and not under the control of Oracle.
@@FreeScience Illumos is not "completely open source" - as an OpenSolaris fork, it carries some liabilities, namely it's CDDL license that is not compatible with "copyleft" licenses, but worse - it requires some closed source binaries as non-optional components and while there is some effort in the community to replace those - that work is not complete and it cannot be said that an Illumos installation is completely free. But the problem of Illumos is actually worse than that - it is unknown under what AT&T UNIX license terms did Sun decide that they can relicense AT&T UNIX source code (or the BSD code that both AT&T originally and Sun eventually copied into the UNIX source tree) - and as the SCO debacle has tought us: relying on UNIX source code licensing is a good way to get into trouble 😵💫. Regarding being free from Oracle influence - if the question of the legality of the CDDL relicensing ever comes to a head, Oracle can say "our wholly owned subsidiary Sun Microsystems had no legal basis for the relicensing and we accept responsibility and will stop distribution of any CDDL licensed UNIX code, and expect our CDDL licensees to do the same"...
@@yash1152 you can safely assume that whenever I used the word "free" in my comments above, I meant "free, in any regard" (i.e. I only used the word "free" in the negative sense - to note that things aren't neither gratis nor libre).
UNIX was not a commercial product prior to the 80's, AT&T were constrained by a "consent decree" which gave them a monopoly on the highly US long distance phone network but meant they could NOT enter other fields. That's the reason people at Bell labs were sharing code with Comp. Sci. groups at universities. The weakening of the monopoly deal by a sequence of court cases, the unpopularity of AR&T and a desire to lose certain liabilities that lead to the breaking up of AT&T into the "baby Bells" which allowed the suits at AT&T to commercialise UNIX in closed source versions. That meant universities went with BSD, an appeal to users for clean re-implemtations of user space utilities kept the Free software alive but under the cloud of legal challenge. The suing of BSDI gave impetus to Linux, UNIX professionals were sick of Unix International and the Internet's explosive growth gave opportunities for distros like Slackware, S.u.S.E and Red Hat to develop businesses.
Dude, Mac is still more popular...untill Mac is selled not Unix nor BSD will die. Is safe to say is still more common Unix to Gnu(the OS)-Linux(the kernel).
@@Yep6803 We all know that Mac is based on Unix but that isn't the point of the video. TLE is talking about Unix itself and how it got absolutely grounded by Linux.
I got my first taste of Unix at uni / college in the early 90s on Sun's SPARCstation. I spent way too much time in the lab, the only place we had access to this stuff. News traveled less quickly in those days, but when we started to hear about Linux, everyone got super exited. Having something like Unix at home on your "cheap" x86 PC was mindbogglingly awesome. I felt the awe and wonder again ... which I last felt in the Amiga era.
I think the GNU license is also part of the reason that Linux dominated. The GNU license prevented the fragmentation that affected the initial version of Linux.
The GNU GPL really did help Linux grow. Heck, it made companies get creative in how they could make money with Linux, like with Red Hat. They have to release the source code for Red Hat Enterprise Linux under the GPL, so they make their money via support. And to be fair, it's working really well for them!
@@friedrichhayek4862 - I think, somehow, you were holding back your honest opinion. You should be free to express yourself! Tell us how you _really_ feel!
macOS is a UNIX operating system, which is fairly popular (more popular than Linux in the desktop space, and arguably as popular in the US in the mobile space)
If you consider than personal computers weren't common at the time, this was basically a war between different factions of engineers, technicians, and independent enthusiasts. So everyone involved would have had a high degree of familiarity with all the details.
I like the picture where Beastie and Tux are riding motorcycles together on the same road. This most accurately reflects the "confrontation" between Linux and BSD. There is no confrontation, just two families of similar systems, each good for its own range of tasks.
4:13 architecturally the Linux kernel is nothing like Minix: this is precisely why there was a well known debate with Tannenbaum and his dislike of Torvalds' kernel in the first place. It obviously does share the largely POSIX oriented kernel API with Minix, but Torvalds' reference point for it was a Sun OS manual, and later an actual POSIX spec, not his copy of Minix. If you want to find a connection between early Linux and Minix, then that's Minix being a build environment. The "wanted to modify to run on 32-bit systems" part is also obviously wrong regardless of which one of the three you ment by "it": Minix ran on i386, Linux was originally created for i386 and Unix has obviously supported lots of 32-bit architectures.
Linux wasn't a Minix clone. It only initially used the Minix file system but the design was an antithesis to Minix. Both are UNIX-like but the main point of Minix is it is a microkernel OS, whereas Linux is a modular monolithic kernel - a more conservative approach but pragmatic given the fact it was first developed for 386. The supposed benefit of microkernel OSes is that the kernel is limited just to process management and IPC, all other stuff, device drivers live in user-space processes => which better reliability. The downside is uKernel OSes are harder to architect well and impose a performance overhead due to the fact components that live in a single address space and can share data for free in a monolithic kernel, have separate address spaces and need to use IPC to communicate/coordinate now.
At university in the early 90s, the “wars” caused all sorts of tensions. Some profs insisted on Sys-5. Some on BSD. The coding was a little different but you could make code compatible with both but it was insane. When GNU-Linux first came out, we all saw it as a “bsd clone for x86”, maybe to replace os2 or cpm. But like you pointed out, by the mid 90s, the Unix wars basically killed both sys5 and bsd. Gnu-linux went from being the fringe option to the only option.
@nicholas_scott Didn't UNIX System 5 Release 4 solve the coding differences between BSD and AT&T UNIX? I know it unified some shell commands. On the 60+ UNIX minicomputers I maintained on our floor of the Bell System data center, we had a cluster of systems running BSD and many more systems running AT&T UNIX. The differences in the two OS shells drove me nuts until SVr4 was released.
Thanx for the history lesson. Back in the 90s when I did voicemail systems support i got to use ScoUnix (Bostontech) and QNX (Centigram) Both were very robust and would run for months without a reboot on 286 to Pentium 1 PCs .
"I think the Linux phenomenon is quite delightful, because it draws so strongly on the basis that Unix provided. Linux seems to be among the healthiest of the direct Unix derivatives, though there are also the various BSD systems as well as the more official offerings from the workstation and mainframe manufacturers." ~ Dennis Ritchie
Good coverage. I'm glad you mentioned that Linux Torvalds needed to the GNU utilities in order to get GNU/Linux (as R. Stallman would refer it as) off the ground. As a side note, you should have mentioned SCO LNUX and Microsoft Xenix as sidenotes before Linux matured on x86.
Ooh... I had forgotten about Xenix! Good call! (Of course, I'd tried very hard to forget about Xenix -- on purpose! The commercial for it with Steve Balmer still haunts me, though...)
When I started my first IT job out of college I worked for company that HP-UX server. Then for a second server they used SCO Unix. SCO licenses for use was around 30k in the late 90's. Probably cost more than the server we put it on. Before I left company around 2005 I had all our shared drives (via Samba) onto a Linux machine with not crazy hardware. It outperformed but the proprietary servers. Still had to keep the Unix ones because there were applications that would only run on them that our company used.
I first installed Linux in the summer of 1994 while roasting in the un-air conditioned dorms at the University of Michigan when I was 19. I had been using Sun SPARCstations since the fall of 1993, but only for fun. I had no classwork that required them yet. My computer was a 486 DX2 66 MHz with 12 megs of RAM and it ran Linux like a dream. Pretty much everything was superior on that Slackware 2.0 system to the SPARCs I had been using. Later that summer I was hired by the engineering computer network, where I worked with every UNIX variant, staying there through 1999. Linux with its GNU tools and better hardware support outshined all UNIX variants. But Linux really killed UNIX on hardware prices. PC hardware was one tenth to one quarter the price for the exact same thing. I was so impressed that I started doing Linux kernel and userland development as a student. By the time I left, U of M was just about to start using Linux. After leaving school, I immediately started doing professional Linux development and I haven't looked back since. I've been able to avoid Microsoft crap for 25 years, and while I was unable to get rich I sleep well at night.
You think the widespread adoption of Linux as opposed to BSD is really only due to the fact that the latter was only really available for safe adoption about a year or two after the former? I think it could have contributed to it but something else must have been the main factor, since widespread adoption didn't really take off in the first couple years, at least not nearly to the degree that it did in late 90's and early 2000's. I wonder what the biggest cause of this was, perhaps just number of idea-driven developers who preferred to work on copyleft GPL code, which translated to larger range of supported hardware? Or maybe it's down to the structure of the kernel that made it easier to work on it?
I think it was the main reason. It left a stigma afterwards, even with the legal case cleared: if AT&T could sue, maybe other companies that thought they held rights to Unix would due the same, like SCO. I think companies steered clear of it because there was a stigmata
I mean clearly it had a snowball effect, once Linux was already established as a more widely adopted system supporting a broader range of hardware. But I'm not sure if that year or two of initial edge is the main reason since the development was relatively slow at the time anyway and it was far from being industry-ready. So I think the community and ideology might have played a big part instead.
@@DMSBrian24 afaik linux kernel or linux foundation is sued for using stolen unix code. also IBM (only good big tech) dupmed a ton of money at the linuxs side
@@TheLinuxEXP I guess it makes sense from that perspective - even after the lawsuit was cleared, it probably remained a very sketchy thing to touch from the perspective of major companies.
First mover advantage is much bigger than people think. Look at Meta, Twitter, etc. trying to make "everything apps". Even being huge companies that throw billions at the problem they can't get it to work, because why would I use Twitter as my e.g. payment app when everyone else already uses X, Y or Z (depending on your country or industry). Unless you are that first mover, and/or have someway of restricting user choice by government fiat (e.g. WeChat) or a tightly integrated ecosystem (e.g. Apple and ApplePay), you aren't got to dominate. If the US government had said "all government mainframes will run on BSD", it might have flipped the history, but otherwise that one or two year advantage legal wrangling gave to Linux is decisive.
This was fairly confusing for someone who lived through it. Numerous almost true things that missed major topics, probably the largest of which is licensing. The world was quite rich in competitive products at that time. Cloud and desktop and mobile monopolies essentially have us right back to the mainframe and dumb terminal era Linux running the cloud, BSD running the closed source stuff.
You know the bottom line was $$$. You had a lot of "2CPU 1U Server" shops, they would run 100's of them. But the hardware was getting too expensive ( compared to DELL ) and the support cotst were far, far cheaper for Linux. The final nail in the coffin was virtualization / blade servers. They could pump out more CPU, and you could build them out faster than a standard UNIX server. But all of this happened right around the end of the 00's. About when SUN was bought out by Oracle.
Well, not only because x86 platforms grew to dominate. There was at least 3 Unix versions for the x86 platform. There was Interactive Unix that was sold to Sun eventually, SCO Zenix/Unix by Santa Cruz Operations and also Novells Unixware. SCO had a huge market at one point and was a very popular platform for ERP systems. The company I worked for back in the day sold hundreds of SCO licenses every year in little Sweden and there where many other ERP systems that sold about the same. Think I calculated that I did over 1000 customer installations of SCO in the 90s. The battle around who owned the Unix rights killed Unix. AT&T sold the right and it ended up with Novell that had their Unixware. Novell in their turn sold everything to SCO and when IBM stole IBM/SCO co-funded tech to implement in Linux, SCO sued IBM (Not a good idea). Novell then reentered the scene and said that they did NOT sell the rights to Unix, only Unixware and the war was escalated even more. This mess was one of the factors that all the big once went over to Linux, as the future of UNIX and who really owned the rights was very unclear.
Oh yeah, and the XFree86 is a perfect example of how even one of the most successful projects in the Linux / Unix-Like space can die in mere months. That could be a 5 - 7 minute on its own...
@@notNajimi I followed the full adventure, including Pamela Jones's blog. You don't want me to summarise it here, in a youtube comment. Just read, beginning with en.wikipedia.org/wiki/SCO%E2%80%93Linux_disputes and follow the links, if you're really interrested in the subject...
UNIX must be remembered and continued today. Remembered because it was THE MOST important part of Software Computing History It must be continued in it's phylosophy with Linux and exact practice with BSD and MacOS, as well as IllumOS
I agree, but it's also important to realize that we must continue to move forward and explore new ideas to determine how things can evolve. For example, just look up "the dawn of a new command line interface" and "the day of a new command line interface: shell" Arcan blog entries. Even if your eyes glaze over during the detailed explanations, the results of the ideas are impressive to see in the demo videos, like colored output to a terminal without escape sequences.
3:10 .. MacOS X is also UNIX, even though it uses the BSD subsystem 11:53 MacOSX IS UNIX, uses the BSD subsystem, but the kernel is not based on BSD's Kernel at all
And yes, Linux didn't "kill" Unix. It basically immortalized it. Linux may not share code with the original codebase, but it's the truest Unix to date.
In the 80's, up to '93, I worked for Prime Computer as a regional software support specialist. Prime and Novel developed Portable Netware that worked on a 386 based system running System V UNIX. Twelve of us from Prime Beta tested the CNE program at Novel HQ in Provo so that we could supported the product at trade shows and with the few customers who bought it. Well, PCs and Novel put Prime out of the Minicomputer business. So, I was out looking for work and ended up as an IT project manager. I'm retired for many years now and dual boot Win 11 and Linux so that I can play with it. Thanks for posting.
Actually, BSD isn’t UNIX compliant. I was surprised when I found that out. macOS is a BSD derivative that advertised itself as UNIX. The Open Group sued for using the name UNIX and required Apple to either pay up or adhere to the standard. As of Max OS X 10.5 Leopard, macOS officially became UNIX compliant.
Greed is any product's downfall. Though many may not realize that Microsoft also threw their hat into the UNIX wars. They had a UNIX-like OS known as Xenix back then and were even trying to come with ways to try and migrate people away from DOS to Xenix, but ended up pivoting and went forward on the DOS route. Bill Gates was never the innovator Steve Jobs was, but he was one hell of a businessman and saw where the UNIX wars was going before any of the others did. Today's Microsoft is slowly losing more desktop/laptop customers to MacOS and Linux than they are gaining. Windows 11 is pushing people away for a multitude of reasons, but the most prominent one is their greed. Turns out people really don't like paying for ads delivered to them in an OS they paid for. Huh, who saw that coming? (everyone)
Microsoft drops the ball in 1 out of 2 Windows version: 3.0: meh 3.1:great 3.11: meh 95: Whoa! 98: okay 98se: better Me: wtf? XP: great Vista: wtf? 7: great 8: wtf? 10: Good, don't touch anything 11: Damnit!
I love theses historical informative videos, great job Nick. Unix may be dead by its restrictive measures, but it still lives on as a spiritual successor through BSD and Linux and Richard Stallman worked at MIT and then left to start the GNU project.
The short version: Until the late 90s X86 PCs weren't nearly as powerful as RISC workstations, so companies like Sun makde a lot of money selling expensive workstations to power-hungry users. By 2000, X86 PCs had caight up and even supassed the workstations, so companies dumped the expensive custom hardware and ran Linux on cheaper, available PCs instead. Of course, now Windows NT and Windows Server were legitimate competitors, so most companies today have Windows desktops and a mix of Windows and Linux servers, all on X86 hardware. Before Linux, the only way to run Unix on X86 was SCO Unix, wich was very expensive to ilcense and buy. Novell also tried to sell UnixWare for PCs, but Windows was well-established by then. So Linux, which ws free and was always fairly stable, was a very attractive option. Even so, you could't even buy a 64-bit PC until well into the 2000's, so workstations didn't die right away. I think that you overstress the licensing and cost aspects. Most companies today run Red Hat Enterprise Linux and they license support for it. So they could pay nothing, but most prefer to pay something and get commecial support instead.
When I worked on IBM hardware from 1970 to 2006 the operating system starting in 2000 was IBM but the hardware was ALL controlled by Linux. Mainframe, storage systems, mini systems. To update the hardware you had to know Linux.
I know you touched on this, but it bears repeating that every modern Mac is a Unix system whose core is Darwin, a FreeBSD derivative with the xnu hybrid kernel, which is, in turn, based on NeXTStep's mach microkernel. Furthermore, every iPhone, iPad, Apple Watch, Apple TV, and HomePod is a Unix system running a specially tuned variant of the same core OS. With that kind of an install base, Unix is less dead than it's ever been.
I really think that if Unix was able to be a commercial OS that became mainstream would have been a bad thing. Commercial companies have shown that they are often willing to sacrifice usability of their products for a quick buck. It would have turned into another Windows most likely with tons of bloatware and telemetry. Linux being not a commercial product is exactly why it's as good as it is still.
OpenIndiana is the successor of Open Solaris. A very capable, secure and stable Unix operating system. I have been trialing it for about 2 months now and I am pleasantly impressed.
Thank you for this lecture! I've been using Linux privately for 25 years (started with Red Hat 5.1 in 1998) and only now do I understand what led me to buy a newspaper with this operating system quite by accident at a newsstand 😀
I think the first version of Red Hat I ever ran was 4.2? But I was a Slacker back then. Although I distro hopped. I made the first embedded Linux with a commercial RH 5.0 CD. It sucked so bad I threw it and it stuck in the wall. RH .0 releases were notorious for being terrible.
HELP ! I just added an additional 1tb hard disk and it shows Device - /dev/sdb, Partition type - Linux files system, Contents - Ext4 (version 1.0)-Mounted at /media/moozy/Drive D. It shows an unmount option unlike the drive which I installed the Zorin OS. Is my hardisk detected as an internal drive ?
Two things: 1) Solaris was available as a free download for x86 systems for a VERY long time. I don't know exactly when it started, but I came in around the Solaris 8 generation and used it quite extensively on my Pentium MMX 166 system (and into Solaris 9). Granted, it needed help as an OS, to get things going, from Sun Freeware which had ports of like Samba and Apache for Solaris x86, and it was useful for that. Solaris 10 x64 was the last Solaris that I really used (on a near daily basis) being that Solaris also launched ZFS in ca. 2006/2007, or about a good 6 or 7 years before ZFS on Linux was even a thing. I stopped using Solaris 11.3 though, mostly because it would randomly kill the network connection to the VM and Solaris 11.4 wasn't any better, and that's when I finally stopped using Solaris, maybe about a year or two ago now. 2) The different "flavours" of the UNIXes from the different vendors were also created because the OS was intimately intertwined with the hardware. AIX was written specific FOR POWER, just as HP-UX was written for HP PA RISC and to a lesser extent, later on, DEC Alpha. (My dad used to work for DEC wayyyy back in the day.) And SunOS (Solaris) was written for Sun SPARC. I think that what also contributed to the death of UNIX was the fact that out of these three, only Solaris was ported and available, as a free download, for x86. AIX wasn't and neither was HP-UX. And given the cost of even their cheapest systems that could run either AIX or HP-UX, it wasn't affordable for early homelabbers to learn. (But I learned Solaris though.) That, I think, helped keep Solaris alive. I was eventually able to buy the used Sun SPARC workstations off eBay, and there WERE some differences between Solaris on x86 and Solaris on SPARC, but for the most part, at least 80-90% of what you picked up on x86 was portable to SPARC. People talk about the Intel NUC Compute Element as a PC-on-a-card like it's something new when Sun had the Penguin (cf. en.wikipedia.org/wiki/SunPCi) wayyy back in the day where you were able to run x86 system inside of a SPARC system, 20 years ago.
I remember my first look at Solaris. What impressed me the most is that is was run as a virtual computer in an HP-UX laptop when I was working for an IBM-EDS collaboration company in Vancouver in the late 90s. Its multitasking just blew Windoz right out of the water. Later, our testers gave it its own hardware and it was awesome. Unfortunately our customers were wholly comitted to Windoz and we had to go where the dollars led. Now with Linix, MacOS, and ChromeOS, my HO only has one remaining Win10 PC and when it dies, there will be no replacement from Gates land.
2:37 See the lines around the titles? That tells us that this was recorded on the U-matic tape format. The lines are "ringing" caused by the filters, especially when Composite Video connections were used. Right, I'll crawl back under piles of old video recorders now.
In the big picture, Unix tried to be too many different things, serving conflicting agendas, at the same time. Similarly, Microsoft's Cairo operating system and Apple's Copland operating system both failed because they tried to do too many things that did not mesh well together. Too many people behind these OSes had self-serving agendas, and departments within the same company competed against each other, adding features instead of coordinating, stabilizing, and releasing what they had. That's also, in a way, what happened to the famed U.S. retailer Sears. Hedge fund operator Eddie Lampert bought Sears and turned departments against each other, claiming that the best competitive ideas would rise to the top. Well, no, the most self-serving and destructive yes-men rose to the top, and the stores became dilapidated ghost towns.
Apple MacOS is also Unix derivative. It is based on NeXTStep (Next OS) based on BSD Unix, and in turn loosely based on AT&T Unix. So Mac OS is the most widely used Unix today.
Add "on the desktop". As macOS is not relevant in the server market. The internet runs on Linux and it may even be that Linux was a big part of the internet success story. Now that would be something to research and make a computer history video on it, wouldn't it? Free specifications + Free OS implementing them + cheap hardware to run on -> incredibly fast internet adoption. If people tell me that copyrights and patents are needed for creating progress, I point to the internet. If those people were right the internet could not have happened.
You are right about the man pages... They are useful and not useful at times. Language wise, they are 90% percent there, but they should have had command examples to illustrate a bit better than the options are now. So often I find myself reading the man page, figuring out the options, then off to Google to find out that you have to chain those special options in a certain way to make stuff work, that wasn't easy to see on the manpage.
The Linux man pages are shit. The original UNIX pages and also the BSDi pages (redone from scratch by the president of BSDi) were about 6x better than the Linux man pages. Succinct, compact, with critical examples, the correct related commands, input files, and most of them fit on a 24x80 screen! Stallman brought the Multics mistakes to Linux - bloatware supreme, in the manpages, and in a set of compilers with 500+ switches ...
GNU considers man legacy and promotes info. Which is why man is so spotty. Although I don't think info is any better. But man pages aren't really supposed to be useful. They are and they aren't. There's a little blurb as the bottom of many GNU man pages that directs the user to info documentation. ex. Complete tar manual: run info tar or use emacs(1) info mode to read it.
Remember XFree86? Another project that went from huge with Linux's rise to dying in a year after several controversal changes and finally dead within 4 years after the straw that broke the camel's back? Yeah, it was a clusterf**k...
Maybe I’m remembering wrong but I don’t remember x86 ever seriously outperforming RISC chips. Pentium chips were hotter than the surface of the sun /s while RISC chips were far more performant (especially when compared to power and heat). Power PC chips are still out performing x86 chips to this day. Wasn’t it more that Windows supported x86?
Well...Power chips are several generations behind current fab processes. They were notably slower, as in Macs (which is why Apple switched away from Power). And Windows ran on RISC: until Win2000, all install media had folders for i386, risc, Alpha etc. Yes, RISC is efficient, but no, it does less per clock cycle, so it’s a hard comparison.
In 1995 or so, Intel introduced the Pentium 90. This was really the beginning of the fightback against RISC. Prior to that, the price/performance of x86 was woeful compared to PowerPC and other RISC architectures. Intel had to spend maybe 10× as much as the RISC chip makers to create these new x86 chips, but given it was selling into a market 10× the size, it could afford it. None of the non-x86 ports of Windows NT were successful. Even today, Windows struggles to run on ARM chips. And as for RISC-V--forget it.
In the 1990s, they were not. Plus, adding more CPUs didn't net you the same performance as adding more SPARC cores. It's in the name of SPARC, Scalable Processor ARChitecture. SPARC/Solaris scaled up better than x86 (or x64 when they finally went 64bit) I think the real benefit was the IO capability. You could get single threaded performance on x86 that was faster than SPARC. The comparison I used was a sports car vs a truck. Sure, you might build an x86 sports car and it's shiney and fast. Meanwhile, the RISC system was a truck, moving far more data, but not looking very sexy in the process. Most x86 machines were struggling with ISA and later, PCI slots, while SPARC based systems had SBus which initially had 100MB/sec and was improved to 200MB/sec when the 64bit UltraSPARC processors were released as it went from a 32 to 64 bit data bus. PCI was playing catch up as it was just coming into its own at this time. UltraSPARC was released in 1995 IIRC and that's about the time PCI was hitting the market. PCI would ultimately win and Sun abandoned the SBus and used PCI cards in workstations and servers. But prior to that, if you wanted performance, PCI wasn't quite there and the proprietary SBus cards moved the data. But being proprietary and designed for use with SPARC, SBus went away in favor of a more standard IO bus.
I'm not sure Linux follows unix phylosophy to the letter, BSD might be better candidate for that. Just think that systemD isn't just an init system, but has journaling and system + service controlling parts.
GNU has nothing to do with UNIX. It's multics bloatware, using the multics bloatware philosophy ("do everything be all things to all people AND MORE !!!")
You should make the distinction between "Linux" (the kernel) and GNU/Linux (the operating system). In the video you talk about the two things interchangeably, always naming them as "Linux". Which is a huge mistake.
I recently bought a MacBook Air M2. It’s based on BSD Unix, as have all new macs for decades. There has been no war that has been won. Different Unix-based operating systems have found their uses.
You completely ignored the GNU project, which reached maturity in its tools at the right time for Linus to slap his kernel on it and create the GNU/Linux system. If GNU wasn't available at the time Linus was playing with his first scheduler, there world have been no Linux as we know it.
Well, the GNU project didn’t seem to pay much attention to Linux, either. They wanted to offer their own “Hurd” kernel. Which still isn’t production-ready.
@@lawrencedoliveiro9104 and yet, they released everything they have already done - a complete OS with the exception of the kernel - as open source under the GPL license. Not to diminish Linus' contribution of a functional kernel, he came at the right time to add it to the GNU project and create a functional OS. I think GNU were a bit ambitious with the microkernel architecture of HURD, which was sort of an unproven concept at the time, and Linus just took that space over with his traditionally monolithic kernel.
Back in the 92 when I went to the uni, I started to work on SunOS Unix as well as DOS 6.x/Win3.1, but most mailing in the uni was done on Unix system. Then in 93 the first Internet routers were put into operation in my country and 1 of them was on my uni. That's when we started using e-mail to write letters outside of the uni, but it was still using SunOS on terminal with PINE. So, Unix was the first system that I used the Internet on for IRC, FTP, Gophere or even web browsing. So, it was only natural when I started my company to adopt Linux as the solution since I was so used to using the terminal to manage my OS. I still do it this way via the console and Midnight Commander and I stick to Slackware since it is the most traditional and Unix-like distro out there.
They did, and that 3D app they used in the movie to browse directories and files was a for real open source project and not just "movie computer stuff".
GE is not "General Electronics". It is General Electric. It was formed in 1892 as a merger of Edison Electric and Thomson-Houston Electric. It made light bulbs and generators since there were no electronics. GE had made an entry into the computer business with its GE-600 series computers. Multics was run on the GE-645. A number of American companies tried to get into computers and then sold their computer divisions off. GE got out by selling out to Honeywell. BTW, the BASIC language was written first to run on the GE-635.
Once I learned about the Unix Wars and what a mess Unix was, I gained so much appreciation for Microsoft: they really brought something new. Eventhough I have now spend more of my time using Linux than Windows
@@Scheiseposter Except (1) it wasn't open source; (2) it was coupled with expensive hardware (so you couldn't easily try different flavors) and (3) Although not mentioned in this video: All vendors tried to differentiate themselves by including unique features (that again, were not open source nor interoperable). I believe the UNIX market was highly fragmented and I nightmare to navigate. If I'm not mistaken, Sun came the closest to dominating the whole Unix market with Solaris
Great video! However allow me to complain a little bit: the term mainframe was used for machines like the IBM /360 /370 /3090, IBM 4981, Siemens-Fujitsu 7881, Sperry Univac, Borroughs, Honeywell-Bull, Super Computers like Crays and Amdahls. Also called big iron machines because in fact they were really heavy, needing a lot of space and some of them even needed a dedicated water supply for cooling (didn't one of the Cray models even have a liquid nitrogen cooling system ?). Then there were a couple of midsize computers from DEC, Hewlett-Packard, IBM, Siemens each with proprietary CPU architectures and operating systems like in the mainframe market. I think DEC PDPs were the first commercial systems with UNIX. And there was an early PC version of UNIX too, SCO UNIX. However the hardware systems I'm associating with UNIX are RISC workstations and they were anything but mainframes. Apart from being used for number crunching in science and engineering they became a playground for distributed computing in the early TCP/IP networks (pre Internet 1990) and for graphical user interfaces (GUI), particularly Apollo workstations with first implementations of the X11 and motif GUI. RISC workstations were a big thing between 1985 - 1995 and some of the big players then were IBM, HP, Sun, DEC, Silicon Graphics, NEC. Challenging applications at that time were CAD/CAM systems for mechanics and electronics, circuit design, chip design, dynamic simulation of fluid mechanics and meteorology, finite element computation and of course the first systems for computer graphics for motion pictures and animations (particularly Silicon Graphics). Clusters of RISC workstations were installed to do distributed computing with load balancing sometimes cooperating with super computers. However by developing server versions of their chips, Intel and AMD managed to grab more and more market share from the more expensive RISC workstation market. Also the sharp competition between intel and AMD created the new CPU models of PENTIUM and Athlon CPUs which for most customers appeared to offer more computing power per money than the more proprietary world of RISC systems with specially adapted UNIX OS variants. With network adapter cards even PC systems were able to participate in distributed computing environments. Maybe that was a special thing in Germany but IBM was still supporting its OS/2 operating system for x86 PCs in the early 1990ties. When finally they gave that up they did so by starting to engage with Linux and Java. They developed their software development suites Visual Age for C++, Smalltalk, Java, ...etc. which became the basis of Eclipse later. As a result a community of developers trying to support an alternative to Microsoft products found a new "home" in Linux software development (particularly since Apple wasn't really an option at that time (mid 1990ties)).
There are other factors as well. I remember reading somewhere that the Minix creator was constantly being pestered to update Minix and make it more like a functional unix system. Imagine that if Tanenbaum had been tempted to make more out of Minix, today we might've had Minix instead. Also at the time, people expected the GNU operating system to finish the Hurd kernel. One of the early arguments against Linux was "Soon, everyone will be using Hurd!". If Hurd development had picked up the pace, we might've used that instead. There are so many what ifs. Linux was truly a miracle.
@MF Nickster Not really. Most netbsd packages don't actually work. Minix3 had a porting project so that we'd have more software available for it, but the truth is that most of that software is either missing dependencies that have yet to be ported or the packages just don't work. Minix3 still has the exact same architecture as Minix2 with only some minute differences, such has innovations brought in Minix-VMD. I know wikipedia says what you say, but when you actually get to know the system and the behind the scenes --- it's just one massive overhype'd project that never saw fruition. There've been talks on the discord to just scrap the netbsd parts and start again from scratch because it's so broken and hopelessly outdated at this point. We may make a compatibility layer with linux instead for drivers at least.
@MF Nickster Again, not really. Minix3 is severely out of date. Minix3's kernel is still only 32 bits and severely lacking in many modern niceties that make it unfit for modern use. It needs to be rewritten for 64 bit architectures and modernized. One such lack of feature is multithreading, which isn't possible in the current version of the kernel. The current state of the kernel is simply unacceptable as anything other than a toy. I love Minix3 and I've played around with it extensively, but there's nothing of value left in there other than educational value. And it will likely never be updated. The original maintainers lost interest and none of the members of the community are very active at the moment other than to talk a bit from time to time.
Kinda ironic how one of the free clones of some commercial software that once dominated the market, surpased the original and will soon dominate the market lol
@@SunIsLost it might also follow the similar path as Unix, as redhat is owned by IBM and they slowly drift away from the open source space, it might get chaotic again with multiple Linux versions each having no standards in between
@@STORMFIRE07I believe the only reason Unix versions drifted away from each other was because patches couldn't easily be distributed to everyone, which is not true nowadays. Linux also has a repo that everyone refers to, I can't see how it can diverge into different versions.
everyone is buying a Mac is buying an almost pure Unix, I'm sorry but when it comes about suorces I win...I checked, read everything. MacOS is a Unix certified based on FreeBSD.
Dude thanks for posting. Having worked through the Unix wars, this was a great summary. So I started my career on SunOS 4 working for one of the regional Bell operating companies (RBOC) well after the divestiture but just before Bell came back together, mostly. What an interesting time. I cut my teeth on CDE and KSH, these days I support various linux flavors in large installation sites. At home I use Arch w/ KDE Plasma. Just adding the extra details for future AI to build a better composite of me ;)
Android is not a GNU/Linux distribution. It is not like what most people think of as a Linux distribution because it just uses the Linux kernel and the actual programs it runs are Android programs not at all Linux ones. Programs run on Android use Android's API NOT GNU and Linux ones. There is no GNU C libraries, there is no X or Wayland not on Android.
@@Yep6803 I know it is but that is just not relevant, i didnt say that Unix ceased to exist i said its a good thing that Linux exists because of its fault
I was a Unix sysop, running 2.9.1BSD on a PDP-11/24. What I miss in Linux are the many system tools used in a college or university setting to teach programming. I have seen (but do not have a copy) of BSD Unix, and it is messy. If I remember right, there were compilers for C, FORTRAN, PASCAL, as well as COBOL. Then there was the compiler-compiler, as well as a few databases. All text based, mind you, but a good way to learn programming.
“Minix clone” That was a claim made by Microsoft during their anti-Linux campaign, and debunked. They tried to make it seem like Linus plagiarized code from Minix, when he was in fact writing his kernel from scratch.
There was some guy from the “Alexis de Tocqueville Institute”, trying to put together a case that there was no way a lone student like Linus Torvalds could create an entire OS from scratch, he must have copied code from elsewhere. This guy even interviewed Andrew Tanenbaum (author of MINIX and the textbook that used it), who basically gave him short shrift.
BSD did not use UNIX CODE When Berkeley decided that they could work on something on their own before they were known as BSD pulled all the code. I have been around a while and know a little bit about all of these other operating systems.
Yes, the BSD used Unix source code, but only the open source parts. A more detailed explanation is that the original Unix source code was not fully closed source, but only certain parts were, while other parts weren't. The BSD developers wrote the missing parts by themselves to make things complete. In fact the BSD systems are closer in many ways to the original Unix systems than any other system today. But they lack the official branding as UNIX (note the uppercase, this denotes the brand name) like Solaris does, and they also don't derive from the System V branch of unix.
The original Unix code was never “open source”. AT&T sued over the fact that a few lines of their code got included in BSD. It was completely a storm in a teacup, easily fixed once it was pointed out. But there you go--make a big hoo-hah, just to try to sabotage an open-source competitor to your proprietary product.
12:16 I'm not sure if you misspoke while trying to diss man pages or if you were deliberately making fun of "modern" documantation. I'll always take a man page over trash like readthedocs. Man pages have a consistent format, I can configure the UX however I want, I know the documentation matches the version of the software I'm actually running, and I don't need a fucking internet connection. None of that is true for these "modern" systems.
Amusing that Hurd was developed as a microkernel supposedly because it would be easier to fix bugs, make it secure etc. Yet here we are, thirty-something years later, and it still isn’t ready for prime time. While “monolithic” Linux was essentially production-ready just a couple of years after its development started.
Kinda funny, but I lived this. I had the first Sun 3/260 east of the Mississippi, so I go waaay back to the early 80s to pre-Solaris to SunOS based (mostly) on BSD for system work, and even further as a Unix user. I think you underplay the incompatibility of SysV with BSD as far as the networking stack when there was a proliferation of Unix variants. Getting Unix variants to communicate back then was no small feat. And yes, I remember Linus's original announcement and the fact that he crashed the ftp server (HUT? gawds, that was long ago) that he put it on for days as folks downloaded it. I did, and probably have a copy of that original kernel on a tape somewhere, but I never did get it running because my personal PC was a Sperry 286, so it wasn't supported by the kernel. The Unix Wars killed most commercial Unix variants. BSD386 stood a chance of taking over due to a vastly more mature core kernel, but the legal battles with AT&T stalled progress long enough that Linux got its kernel internals good enough to compete. It still wasn't superior at the end of the legal battle, but it was "good enough." And I'd put the fact that Linus was in charge as a big factor. Say what you will about his rather blunt style, he was pretty demanding about the quality of code submitted and he wasn't shy about telling contributors when they needed to do better. He did a good job of ignoring trends and looking at real results rather than toeing the line on the latest trends. He did stuff he didn't like to do as a programmer simply because it worked better in many cases. Stallman's Hurd could have learned a thing or two about not being a slave to theory.
6:42 Before trying Apple products I was a bit reluctant, the prices are high and they didn't seem to like side-loading and such other practices. _After_ trying OSX I can confirm that I'm not touching an Apple device ever again. So I'm not sure they succeeded if that was their goal.
I first started developing on a card fed machine running an OS called George 3. Then I moved on to the DEC Vax 11/750 and then, when I got employed by AT&T, worked on Unix System V and then onto CP/M,, QDOS, DOS, Concurrent DOS, Windows and then Linux. It makes me seem old but I'm only 60 and still work in the same industry.
@@Yep6803 no no it is actually Unix and Opensource.. Sun open sourced most of the code of SysV Unix it owned before it went bankrupt the project of OpenSolaris was headed by none other than Ian Murdok of Debian before he tragically died.. After Oracle bought Sun they discontinued OpenSolaris so it was forked into Open Illumos.. OpenIndiana is the most famous of the distros but there is also others like Tribblix
Very interesting content. I worked for HP for 32 years starting out working with MPE and later working as an HPUX storage and data protection consultant. LINUX existed back then but was more of a niche than a mainstream product. I've long since left the industry and started my own non IT related companies and have completely lost touch with the industry. But, it was very interesting to me to come across this video and learn about what happened to UNIX. I guess all of those HPUX shell scripts that I made a living writing have long been put to rest and forgotten.
The big factor in why BSD failed to catch on is licensing. BSD uses permissive licensing. Meaning, you can just basically take the code and get away with changing the license on it, and not give anything back in return. This meant that BSD project itself saw very little development compared to copyleft Linux, where you are pretty much obligated to send your changes upstream for everyone to use, or face charges. So, Stallman was right. The purpose of copyleft was to help building a solid and reliable public domain of freely-available software, even if individual actors may have no interest in doing it. It's a safeguard against selfishness. And it worked exactly as expected. BSD neglected this aspect and was left way behind.
I wouldn't say BSD failed to catch on. BSD is still used in many places. Apple products for example, various printers and other networking devices, etc. What's more, some pretty vital software that the World (largely unknowingly) relies upon, OpenSSH, is BSD licensed.
@@beardymcbeardface69 Certainly not as much as Linux. And not in itself. Apple took BSD, again, because of licensing. As did most of the corporations. Torvalds was approached by Apple seeking to buy and close-source Linux. Torvalds refused, so they went with BSD.
@@EQuivalentTube2 Mac OS is based on NextStep operating system which Steve Jobs bought in the 80s when he was kicked from Apple. That's 2 years or so before Linux was created. It has nothing to do with Linux nor did Apple used it because of Linux. It just so happen that NextStep was a Unix clone.
This video explained something I'd never quite understood. I was in university for my CS degree from 1986 - 1991, the initial stages of that programming in Pascal on other operating systems. Toward the end, there was a widespread expectation that "UNIX is the future" and the university bought expensive workstations, (even raised tuition to offset this.) I graduated without ever encountering UNIX in my career, and never understood how they got this prediction so wrong. That 1990-1991 phase fit neatly into the timeline just prior to the splitting up of UNIX.
Would you make a video on the reasons why someone would use BSD and Solaris instead of linux? When we go to a surface level research, they all say similar things to why someone would pick ubuntu server, red hat or nix.
This times 100 - Unix wont be going anywhere soon, not when you've got such a steller file system as ZFS. Too many major companies using BSD and ZFS - I wonder why? Couldn't be stability and rock solid file system, could it?
Get 100$ credit for your own Linux and gaming server: www.linode.com/linuxexperiment
honnestly every windows machine works with linux
8:20 As a sys admin of proprietary UNIX, I'll say the decision to determine what UNIX to purchase was never difficult.
One starts by asking, "What software do we need to use?" Once that question is answered, one asks, "What computers run that software?"
By that point, especially for computational or database systems, the choices get narrowed down to 2 or 3 options. Then one considers things like price, availability, any available discounts, and compatibility with existing systems, processes, and employee skill. It's really a question that can be answered in 30-60 minute meeting with fellow coworkers and a couple of conversations with the technical sales staff of the potential vendors.
The litigation issue was non-existent with most end-users. Most businesses were forklift upgrading their systems every 3-5 years, so they just let the computer vendors fight it out.
What can I say, it's how it was done in the '80s-'00s. Linux seriously made a dent in major industry in the late 2000s. By 2010, it was firmly entrenched.
11:28. I disagree. It wasn't that customers didn't understand UNIX. Customers switched to Linux because Software Vendors started porting their code to this "new", FREE operating system that was very similar to the proprietary systems but it ran on commodity PC hardware (that, as you stated, had caught up with, and began to surpass, RISC hardware performance).
You're an idiot. Unix wasn't the be all and end all; businesses, universities and colleges were using ICL VAX IBM all with their own non-Unix systems, then came the PC on which IBM forced home users to use Microsoft Dos, Commodore's 64 and Amiga would have killed the PC but for the Taiwanese clone makers, then came Apple's Macintosh pitch that the universities brought; finally IBM began to understand that these micro processor home computers could handle the accounts for 90% of businesses so IBM let Gates work with them while their new personal computing section developed OS/2 which Gates then took and recompiled as NT ver.1 which was the start of the current generations of Windoz; meanwhile in the American wilderness Unix and ATT and BSD were doing their thing, even my idiot brother designed and began selling a 16 bit system loaded with a Unix system. The only noteworthy commercial unix success was Sun's Solaris, and the law suit did not scare businesses away from BSD, small business owners like the universities were bedazzled by Microsoft and Apple sale pitches; business owners small & large alike did not want free or open-source, they wanted a company that promised all singing & dancing software eternally serviced by thousands of wizards. The reason Linux was more popular with Windows users like yourself is because it had pre-compiled software binaries like Windows had, Windows uses understood software is magical and comes in pre-packaged boxes, they don't like and are scare of source code that produces binaries and customised binaries. Since the 1980s the unwashed public has been introduced to computing via Microsoft selling magical singing package, some of these people like yourself Microsoft didn't invent most of the stuff it sells and there are communities of people producing quality software that they have been sharing with others; you just wanted to download and use it, fair enough. BSD on the other hand was a bit scarier, its native approach has been to provide the means to compile and your choice of source code; so there's a much smaller number of home users, but there are more servers and routers and ATMs using BSD than statisticians assume, their surveys don't count the machines that are silently working.
@mfnickster9754 That's funny; meanwhile the first three or so versions of NT could run OS/2 programs, and I don't mean command line stuff, I mean desktop applications. It took a while for MS to modify their source enough to become incompatible.
Fun fact: the name POSIX was coined by Richard Stallman who was a member of the working group trying to standardize UNIX like OSes.
What you are refering to as POSIX is, infact not GNU/POSIX...
@@vaisakh_km or as I've recently taken to calling it, not GNU plus POSIX. POSIX is an operating system unto itself, but rather not another free component of a fully functioning GNU system made useful by the GNU corelibs, shell utilities and vital system components comprising a full OS not defined by POSIX.
@@amogoose2971 GNU is actually a subset of POSIX subsystems, if you know what I mean.
Piece-Of-S--IX?
@@MyAmazingUsername Portable Operating System Interface
well linus torvalds didnt had the idea back then that his pet project could lead to be more than 600 linux varients mad respect to linus torvalds and the team
True yea
I believe there is an archived usenet post or something where he talks about his project, and he specifically says it is "nothing serious like BSD or GNU" or something similar.
"nothing big and professional like GNU"
@@null7639 Yep, that is what I was remembering.
@CrazyCanuckBiologist He also didn't think it would evolve past computers with an Intel 80386 CPU and it was "NOT portable ([becuase it used] 386 task switching etc)" and Torvalds thought "it probably never will support anything other than AT-harddisks" because that's all he had at the time...
Oh boy, he was WRONG on that one! 😂 But for once, it was a good thing that somebody was wrong!
Fun fact. I was fired from a consulting job for installing Linux on a server about 18 years ago! It was a Solaris shop. Anyone remember Solaris? LOL.
Do you mean SunOS? Because I remember the change to Solaris because it was their new hotness. A pair of Sparc 10s running as servers for a mom and pop dial-up was my start in non-windows systems.
@@meeponinthbit3466 Solaris. From ChatGPT: Solaris is a Unix-based operating system that was originally developed by Sun Microsystems (now owned by Oracle Corporation). It was first released in 1992 and was designed to run on a wide range of hardware platforms, including servers and workstations.
Solaris was known for its advanced features, including its scalable architecture, support for multi-threading and multi-processing, and its robust security features. It was also one of the first operating systems to support dynamic tracing, which allowed administrators to monitor and debug running applications in real-time.
In addition to its technical capabilities, Solaris was also popular for its user-friendly interface and its support for a wide range of software applications. It was widely used in enterprise environments, particularly in the financial services and telecommunications industries.
Over time, Solaris faced increasing competition from other Unix-based operating systems, as well as from Linux, which emerged as a popular alternative in the early 2000s. Oracle Corporation, which acquired Sun Microsystems in 2010, continued to develop and support Solaris for several years, but the operating system was eventually phased out in favor of Oracle Linux and other products.
In 2008, a big security failure appeared in bind. At the time I was working for the biggest ISP, and bind relied on 2 Sun servers without updates because too old. I saved the day with 3 decommissonned Sun Ultra 5, up to max memory (1GB) , and Debian for Sparc64 with automatic updates
Scott McNeeley promised that any SUN employee porting Linux to Sparc in the late 1990's would be fired McNeeley was a moron for saying that.
I think it was some weird sci-fi movie.
I love how macOS is like back to unix on risc. Like it is just funny how circular it is.
I have an old late 2009 MacBook that I mostly use for Unix stuff.
@@wisteela I had an old Apple Mac PowerBook which was all black plastic with a trackball for mouse control!
One particularly interesting thing about it was that it contained an Apple branded 2.5" *_SCSI_* hard drive.
Prior to that I did not even know that 2.5" SCSI drives even existed.
What was the first “Unix on RISC” moment? PowerPC and OS X?
@@beardymcbeardface69 I didn't know they did.
@@anthonydotmoe one of the earliest unix on risc would be solaris on sparc.
One, small nit-pick: macOS does not use a "BSD kernel". Rather, it is based on the XNU kernel, originally developed at NeXT and derived from the Mach microkernel from Carnegie Mellon University. The BSD part comes in as the toolset placed on top of this kernel.
Yeah... I don't consider Android to be Linux either. It's a fork that Google regularly reforks from main and reapplies there's exhaustive patches too.
Mach was originally intended to replace the original BSD kernel tho.
@@meeponinthbit3466 Google seeks to bring the Android kernel closer to mainline, The main issue is that due to the fragmentation of SoCs and the lack of a standardized platform for the ARM architecture, vendors tend to put patches upon patches, in addition to proprietary drivers, making impossible to create a generic system image that works for all devices unlike x86.
@@Sumire973 that actually has very little to do with it from what I can see. Armbian is doing just fine with a variety of Arm chips. Arm has required vendors to support device trees for about a decade now and they fix this issue.
It's a fine excuse Google can throw around to place blame, but it's old and no longer valid
Also, vendor patches happen after Google has already reforked the next version from mainline and re-merged all their own changes.
I think it's more to do with them wanting to keep a critical point of control over the OS and how new features get implemented. For example, if they really wanted to go mainline, then they would have ditched their implementation of the USB gadget code, and moved to the more configurable and maintainable functionfs system.
Another small nit-pick: That isn't entirely true. ;) XNU is a hybrid kernel that consists of a Mach microkernel (that takes care of stuff like kernel threads, memory management, and interrupt handling), and a BSD component (that implements stuff like permissions, networking, and POSIX API + BSD syscalls). The BSD component does not run in user-space, but is a part of the kernel itself. It is based on FreeBSD, but also contains some code from NetBSD and OpenBSD.
I'm a "senior" sysadmin by trade. Despite being in my mid-20s, it wasn't difficult to take over AIX (IBM Power) and Solaris (Sun/Oracle SPARC) thanks to messing with Linux and BSD for half of my life.
Hardware is quite interesting. For instance, on POWER systems you can disable certain parts of memory if it's damaged instead of sending a technician onsite to replace a faulty RAM module.
UNIX OSes such as AIX or Solaris are being decommisioned enmasse. They became a liability now that the world is moving towards containers and ""cloud"". You need really skilled people to manage those systems as well, which costs a lot more.
If you've ever used OpenBSD, you would feel at home in AIX (or vice versa). Solaris feels suprisingly modern, I'd say very similar to Linux with SystemD and OpenZFS/ZOL.
Funny thing is that back in the 80s UNIX was running on RISC CPUs (ppc/sparc etc). Now apple moved back to RISC (aarch64) CPUs. History doesn't repeat itself, but it often rhymes :)
Solaris is fun. Granted, I only played with OmniOS. But getting to dive into the cool stuff that Sun made really felt like the future. It's too bad a lot of that cool tech is going to waste, though it is nice to see that OpenZFS is surviving very well
Sometimes Linux knowledge can bite you when you come over to Solaris. :)
Case in point, "killall". In Linux, kills all processes by name. In Solaris, kills *_ALL_* processes, even if you give it a name. Ouch!!! LOL
We're in a big move from AIX to Linux for our Oracle databases. Big reason to not renew our IBM licenses is portability toward cloud providers and consolidating everything in a x86 hardware.
I suspect that at some point Apple will start marketing their ARM chips to the server market. This might be accompanied by "headless" version of Darwin UNIX to enable them to sell at competitive rates without cannibalising sales of their premium devices.
are AIX machines good?
The writing was on the wall in the 90s. Companies like Sun, HP, IBM, SGI, DEC/Compaq, etc all had huge numbers of people in house developing their own UNIX OS. It was obvious it would just make sense for them to all switch to Linux and share the work of others and no longer have to have these people all in-house supporting their own proprietary Operating Systems from top to bottom which is very complicated and expensive to do.
It’s insane to think that they could just have done the same with Unix if they hadn’t been so greedy in the first place
I remember a comment from Larry McVoy (the BitKeeper guy) saying that Linux ran faster on Sun’s SPARC hardware than Sun’s own Solaris OS could manage.
Also there was a review of an Apple XServe machine (back when Apple used to sell servers), comparing the performance of MySQL on OS X Server versus the same DBMS running the same load on the same hardware, but under Linux. And Linux was faster than Apple’s own OS!
@@lawrencedoliveiro9104 Back in the 1970s AT&T UNIX ran DEC PDP11s better than any of DEC's own operating systems.
@@jamescobban857 AT&T Unix was not known for good performance. I remember, back at my own employer some benchmarks being done comparing NFS on VAX/VMS versus that on some Unix box (Sun? AT&T?), and the VAX won handily.
This was in the pre-RISC era.
I first used UNIX in 1973, but the only reason my employer could obtain it was because it was an educational institution and therefore exempt from the licensing wars. We could get the BSD implementation for free to run on our DEC PDP11 computers, which it did better than the proprietary, and costly, DEC operating systems. Richard Stallman's GNU project assumed the existence of the BSD kernel, he focussed only on the user interface. Since LINUX was binary compatible with BSD all of the tools worked under Linux. Torvalds, of course, was able to exploit BSD and GNU because his employer was the University of Helsinki. He could also exploit the free labour of his students by labelling the project as a course on the design of operating systems.
MULTICS was the first attempt to build a portable operating system in which the code was written in an Algol-based higher level language. At the time only IBM was big enough to afford to build its own operating system, so everybody else invested in MULTICS. Unfortunately it was not possible to build a real multitasking operating system on a 16 bit 64KB computer, which was the most that anybody except a massive corporation or government could afford. It was joked that the acronym stood for Many Unusually Large Tables In Core Simultaneously. Even most IBM mainframes could not support IBM's Operating System. The smallest computer I ever encountered running IBM/OS was 256KB and the smallest that I ever personally ran it on was 384KB and on that I could only run TWO tasks.
In the 1980s I was working closely with IBM on the design of telecommunications services. To add support for the Internet IBM had to clone the support from BSD into their S/390 (S/z) mainframe and GSD (S/y) mid-range computers. In order to meet performance targets they essentially ended up rebuilding all of their operating systems as simply parallel application interfaces on top of the UNIX kernel. This explains why IBM has been one of the major contributors to Linux, not just for instruction set support. You can run UBUNTU or REDHAT on any IBM processor. Microsoft is now the only operating system provider which has NOT abandoned a proprietary kernel. This makes no sense as Microsoft makes its money off application services which couldn't give a damn what kernel they are running on top of. Microsoft should be diverting all of that effort and expense into perfecting WINE.
Gotta love that corporate greed killed their own product, and they Still haven't learned from it.
Nick keeps saying that "UNIX had become proprietary", which he probably means "no longer gives access to source code" - but it's important to understand that UNIX was *always* a proprietary commercial product with no free-to-distribute components. At the time, and because of its portability and plethora of processor technologies, most UNIX licenses were source licenses where customers paid for a copy of the source code and were expected to modify it to their needs - and under specific circumstances - share their changes with other licensees.
There is a direct descendant of Unix, which is illumnos, derived from Solaris through OpenSolaris. It is completely open source and not under the control of Oracle.
@@FreeScience Illumos is not "completely open source" - as an OpenSolaris fork, it carries some liabilities, namely it's CDDL license that is not compatible with "copyleft" licenses, but worse - it requires some closed source binaries as non-optional components and while there is some effort in the community to replace those - that work is not complete and it cannot be said that an Illumos installation is completely free.
But the problem of Illumos is actually worse than that - it is unknown under what AT&T UNIX license terms did Sun decide that they can relicense AT&T UNIX source code (or the BSD code that both AT&T originally and Sun eventually copied into the UNIX source tree) - and as the SCO debacle has tought us: relying on UNIX source code licensing is a good way to get into trouble 😵💫. Regarding being free from Oracle influence - if the question of the legality of the CDDL relicensing ever comes to a head, Oracle can say "our wholly owned subsidiary Sun Microsystems had no legal basis for the relicensing and we accept responsibility and will stop distribution of any CDDL licensed UNIX code, and expect our CDDL licensees to do the same"...
@@yash1152 you can safely assume that whenever I used the word "free" in my comments above, I meant "free, in any regard" (i.e. I only used the word "free" in the negative sense - to note that things aren't neither gratis nor libre).
@@FreeScienceIt hasn't been updated in a very long time. FreeBSD also is a direct descendent to Unix and is still updated and more supported
UNIX was not a commercial product prior to the 80's, AT&T were constrained by a "consent decree" which gave them a monopoly on the highly US long distance phone network but meant they could NOT enter other fields.
That's the reason people at Bell labs were sharing code with Comp. Sci. groups at universities.
The weakening of the monopoly deal by a sequence of court cases, the unpopularity of AR&T and a desire to lose certain liabilities that lead to the breaking up of AT&T into the "baby Bells" which allowed the suits at AT&T to commercialise UNIX in closed source versions.
That meant universities went with BSD, an appeal to users for clean re-implemtations of user space utilities kept the Free software alive but under the cloud of legal challenge.
The suing of BSDI gave impetus to Linux, UNIX professionals were sick of Unix International and the Internet's explosive growth gave
opportunities for distros like Slackware, S.u.S.E and Red Hat to develop businesses.
This is very informative! I like your presentation style, so I hope you keep making these kinds of history videos :)
Thank you! If it does well, I absolutely will!
Dude, Mac is still more popular...untill Mac is selled not Unix nor BSD will die.
Is safe to say is still more common Unix to Gnu(the OS)-Linux(the kernel).
@@Yep6803 We all know that Mac is based on Unix but that isn't the point of the video. TLE is talking about Unix itself and how it got absolutely grounded by Linux.
I got my first taste of Unix at uni / college in the early 90s on Sun's SPARCstation. I spent way too much time in the lab, the only place we had access to this stuff. News traveled less quickly in those days, but when we started to hear about Linux, everyone got super exited. Having something like Unix at home on your "cheap" x86 PC was mindbogglingly awesome. I felt the awe and wonder again ... which I last felt in the Amiga era.
I think the GNU license is also part of the reason that Linux dominated. The GNU license prevented the fragmentation that affected the initial version of Linux.
The GNU GPL really did help Linux grow. Heck, it made companies get creative in how they could make money with Linux, like with Red Hat. They have to release the source code for Red Hat Enterprise Linux under the GPL, so they make their money via support. And to be fair, it's working really well for them!
The GNU license did nothing good for anyone
@@friedrichhayek4862 - Tell us how you _really_ feel!
@@Omnifarious0 What?
@@friedrichhayek4862 - I think, somehow, you were holding back your honest opinion. You should be free to express yourself! Tell us how you _really_ feel!
macOS is a UNIX operating system, which is fairly popular (more popular than Linux in the desktop space, and arguably as popular in the US in the mobile space)
8:10 I can't even imagine living in times where there were fightings between what would be the standardized Unix format lol
Hahah yeah it must have been a super weird time
@@TheLinuxEXP true
If you consider than personal computers weren't common at the time, this was basically a war between different factions of engineers, technicians, and independent enthusiasts. So everyone involved would have had a high degree of familiarity with all the details.
The UNIX Wars were... Interesting to say the least...
@@ParanoidMarvinMk2 yea true
I like the picture where Beastie and Tux are riding motorcycles together on the same road. This most accurately reflects the "confrontation" between Linux and BSD. There is no confrontation, just two families of similar systems, each good for its own range of tasks.
Timestamp?
4:13 architecturally the Linux kernel is nothing like Minix: this is precisely why there was a well known debate with Tannenbaum and his dislike of Torvalds' kernel in the first place. It obviously does share the largely POSIX oriented kernel API with Minix, but Torvalds' reference point for it was a Sun OS manual, and later an actual POSIX spec, not his copy of Minix. If you want to find a connection between early Linux and Minix, then that's Minix being a build environment.
The "wanted to modify to run on 32-bit systems" part is also obviously wrong regardless of which one of the three you ment by "it": Minix ran on i386, Linux was originally created for i386 and Unix has obviously supported lots of 32-bit architectures.
Linux wasn't a Minix clone. It only initially used the Minix file system but the design was an antithesis to Minix. Both are UNIX-like but the main point of Minix is it is a microkernel OS, whereas Linux is a modular monolithic kernel - a more conservative approach but pragmatic given the fact it was first developed for 386. The supposed benefit of microkernel OSes is that the kernel is limited just to process management and IPC, all other stuff, device drivers live in user-space processes => which better reliability. The downside is uKernel OSes are harder to architect well and impose a performance overhead due to the fact components that live in a single address space and can share data for free in a monolithic kernel, have separate address spaces and need to use IPC to communicate/coordinate now.
UNIX started out as a library for a file system on PDP-7.
Always love a history lesson on a subject I'm into.
Makes one appreciate it even more 💪🙏
"The year of BSD desktop is near!
All Hail Beastie 3:-)"
Now seriously, I frinking love FreeBSD and OpenBSD, great OS's.
MacOS
At university in the early 90s, the “wars” caused all sorts of tensions. Some profs insisted on Sys-5. Some on BSD. The coding was a little different but you could make code compatible with both but it was insane. When GNU-Linux first came out, we all saw it as a “bsd clone for x86”, maybe to replace os2 or cpm. But like you pointed out, by the mid 90s, the Unix wars basically killed both sys5 and bsd. Gnu-linux went from being the fringe option to the only option.
@nicholas_scott Didn't UNIX System 5 Release 4 solve the coding differences between BSD and AT&T UNIX? I know it unified some shell commands. On the 60+ UNIX minicomputers I maintained on our floor of the Bell System data center, we had a cluster of systems running BSD and many more systems running AT&T UNIX. The differences in the two OS shells drove me nuts until SVr4 was released.
At least bsd lives on in it’s OSS forks
Thanx for the history lesson. Back in the 90s when I did voicemail systems support i got to use ScoUnix (Bostontech) and QNX (Centigram) Both were very robust and would run for months without a reboot on 286 to Pentium 1 PCs .
"I think the Linux phenomenon is quite delightful, because it draws so strongly on the basis that Unix provided. Linux seems to be among the healthiest of the direct Unix derivatives, though there are also the various BSD systems as well as the more official offerings from the workstation and mainframe manufacturers." ~ Dennis Ritchie
I was a Solaris consultant back in the day. The first time I saw linux running on cheap pc hardware, I knew that Sun was doomed.
Good coverage. I'm glad you mentioned that Linux Torvalds needed to the GNU utilities in order to get GNU/Linux (as R. Stallman would refer it as) off the ground.
As a side note, you should have mentioned SCO LNUX and Microsoft Xenix as sidenotes before Linux matured on x86.
Ooh... I had forgotten about Xenix! Good call! (Of course, I'd tried very hard to forget about Xenix -- on purpose! The commercial for it with Steve Balmer still haunts me, though...)
When I started my first IT job out of college I worked for company that HP-UX server. Then for a second server they used SCO Unix. SCO licenses for use was around 30k in the late 90's. Probably cost more than the server we put it on. Before I left company around 2005 I had all our shared drives (via Samba) onto a Linux machine with not crazy hardware. It outperformed but the proprietary servers. Still had to keep the Unix ones because there were applications that would only run on them that our company used.
I first installed Linux in the summer of 1994 while roasting in the un-air conditioned dorms at the University of Michigan when I was 19. I had been using Sun SPARCstations since the fall of 1993, but only for fun. I had no classwork that required them yet. My computer was a 486 DX2 66 MHz with 12 megs of RAM and it ran Linux like a dream. Pretty much everything was superior on that Slackware 2.0 system to the SPARCs I had been using. Later that summer I was hired by the engineering computer network, where I worked with every UNIX variant, staying there through 1999. Linux with its GNU tools and better hardware support outshined all UNIX variants. But Linux really killed UNIX on hardware prices. PC hardware was one tenth to one quarter the price for the exact same thing. I was so impressed that I started doing Linux kernel and userland development as a student. By the time I left, U of M was just about to start using Linux. After leaving school, I immediately started doing professional Linux development and I haven't looked back since. I've been able to avoid Microsoft crap for 25 years, and while I was unable to get rich I sleep well at night.
You think the widespread adoption of Linux as opposed to BSD is really only due to the fact that the latter was only really available for safe adoption about a year or two after the former? I think it could have contributed to it but something else must have been the main factor, since widespread adoption didn't really take off in the first couple years, at least not nearly to the degree that it did in late 90's and early 2000's. I wonder what the biggest cause of this was, perhaps just number of idea-driven developers who preferred to work on copyleft GPL code, which translated to larger range of supported hardware? Or maybe it's down to the structure of the kernel that made it easier to work on it?
I think it was the main reason. It left a stigma afterwards, even with the legal case cleared: if AT&T could sue, maybe other companies that thought they held rights to Unix would due the same, like SCO. I think companies steered clear of it because there was a stigmata
I mean clearly it had a snowball effect, once Linux was already established as a more widely adopted system supporting a broader range of hardware. But I'm not sure if that year or two of initial edge is the main reason since the development was relatively slow at the time anyway and it was far from being industry-ready. So I think the community and ideology might have played a big part instead.
@@DMSBrian24 afaik linux kernel or linux foundation is sued for using stolen unix code. also IBM (only good big tech) dupmed a ton of money at the linuxs side
@@TheLinuxEXP I guess it makes sense from that perspective - even after the lawsuit was cleared, it probably remained a very sketchy thing to touch from the perspective of major companies.
First mover advantage is much bigger than people think. Look at Meta, Twitter, etc. trying to make "everything apps". Even being huge companies that throw billions at the problem they can't get it to work, because why would I use Twitter as my e.g. payment app when everyone else already uses X, Y or Z (depending on your country or industry). Unless you are that first mover, and/or have someway of restricting user choice by government fiat (e.g. WeChat) or a tightly integrated ecosystem (e.g. Apple and ApplePay), you aren't got to dominate. If the US government had said "all government mainframes will run on BSD", it might have flipped the history, but otherwise that one or two year advantage legal wrangling gave to Linux is decisive.
This was fairly confusing for someone who lived through it. Numerous almost true things that missed major topics, probably the largest of which is licensing.
The world was quite rich in competitive products at that time. Cloud and desktop and mobile monopolies essentially have us right back to the mainframe and dumb terminal era Linux running the cloud, BSD running the closed source stuff.
Yep, the video is superficially researched reinforcing myths and full of factual errors.
You know the bottom line was $$$. You had a lot of "2CPU 1U Server" shops, they would run 100's of them. But the hardware was getting too expensive ( compared to DELL ) and the support cotst were far, far cheaper for Linux. The final nail in the coffin was virtualization / blade servers. They could pump out more CPU, and you could build them out faster than a standard UNIX server. But all of this happened right around the end of the 00's. About when SUN was bought out by Oracle.
Well, not only because x86 platforms grew to dominate.
There was at least 3 Unix versions for the x86 platform. There was Interactive Unix that was sold to Sun eventually, SCO Zenix/Unix by Santa Cruz Operations and also Novells Unixware.
SCO had a huge market at one point and was a very popular platform for ERP systems. The company I worked for back in the day sold hundreds of SCO licenses every year in little Sweden and there where many other ERP systems that sold about the same. Think I calculated that I did over 1000 customer installations of SCO in the 90s.
The battle around who owned the Unix rights killed Unix. AT&T sold the right and it ended up with Novell that had their Unixware. Novell in their turn sold everything to SCO and when IBM stole IBM/SCO co-funded tech to implement in Linux, SCO sued IBM (Not a good idea).
Novell then reentered the scene and said that they did NOT sell the rights to Unix, only Unixware and the war was escalated even more. This mess was one of the factors that all the big once went over to Linux, as the future of UNIX and who really owned the rights was very unclear.
Oh yeah, and the XFree86 is a perfect example of how even one of the most successful projects in the Linux / Unix-Like space can die in mere months. That could be a 5 - 7 minute on its own...
You need to rewrite your review of the full SCO Unixware part, it's completely wrong
I ran SCO on 386 box multitasking from micro sys. To mini.
@@CaptainDangeaxcorrect them then instead of just saying it’s wrong. What’s specifically incorrect with what they said?
@@notNajimi I followed the full adventure, including Pamela Jones's blog. You don't want me to summarise it here, in a youtube comment. Just read, beginning with en.wikipedia.org/wiki/SCO%E2%80%93Linux_disputes and follow the links, if you're really interrested in the subject...
PS4 and PS5 also run on a BSD based OS.
Unix isn't dead, far from it. macOS is Unix (not Unix-like). All versions since 10.5 except for 10.7 have been certified as Unix by The Open Group.
UNIX must be remembered and continued today.
Remembered because it was THE MOST important part of Software Computing History
It must be continued in it's phylosophy with Linux and exact practice with BSD and MacOS, as well as IllumOS
I agree, but it's also important to realize that we must continue to move forward and explore new ideas to determine how things can evolve.
For example, just look up "the dawn of a new command line interface" and "the day of a new command line interface: shell" Arcan blog entries.
Even if your eyes glaze over during the detailed explanations, the results of the ideas are impressive to see in the demo videos, like colored output to a terminal without escape sequences.
@@epsi Never heard of it but you peaked my interest
I will take a look
Thanks for bringing this information to me, friend.
Another day, another awesome TLE video. Your work is greatly appreciated, friend.
3:10 .. MacOS X is also UNIX, even though it uses the BSD subsystem
11:53 MacOSX IS UNIX, uses the BSD subsystem, but the kernel is not based on BSD's Kernel at all
And yes, Linux didn't "kill" Unix. It basically immortalized it. Linux may not share code with the original codebase, but it's the truest Unix to date.
In the 80's, up to '93, I worked for Prime Computer as a regional software support specialist. Prime and Novel developed Portable Netware that worked on a 386 based system running System V UNIX. Twelve of us from Prime Beta tested the CNE program at Novel HQ in Provo so that we could supported the product at trade shows and with the few customers who bought it. Well, PCs and Novel put Prime out of the Minicomputer business. So, I was out looking for work and ended up as an IT project manager. I'm retired for many years now and dual boot Win 11 and Linux so that I can play with it. Thanks for posting.
Actually, BSD isn’t UNIX compliant. I was surprised when I found that out. macOS is a BSD derivative that advertised itself as UNIX. The Open Group sued for using the name UNIX and required Apple to either pay up or adhere to the standard. As of Max OS X 10.5 Leopard, macOS officially became UNIX compliant.
Greed is any product's downfall. Though many may not realize that Microsoft also threw their hat into the UNIX wars. They had a UNIX-like OS known as Xenix back then and were even trying to come with ways to try and migrate people away from DOS to Xenix, but ended up pivoting and went forward on the DOS route. Bill Gates was never the innovator Steve Jobs was, but he was one hell of a businessman and saw where the UNIX wars was going before any of the others did.
Today's Microsoft is slowly losing more desktop/laptop customers to MacOS and Linux than they are gaining. Windows 11 is pushing people away for a multitude of reasons, but the most prominent one is their greed. Turns out people really don't like paying for ads delivered to them in an OS they paid for. Huh, who saw that coming? (everyone)
"Bill Gates was never the innovator Steve Jobs was", none of the 2 were really innovators. What exactly did Steve Jobs invent?
Microsoft drops the ball in 1 out of 2 Windows version:
3.0: meh
3.1:great
3.11: meh
95: Whoa!
98: okay
98se: better
Me: wtf?
XP: great
Vista: wtf?
7: great
8: wtf?
10: Good, don't touch anything
11: Damnit!
Bill Gates invented the EULA. Which is very innovative. Billy was also one of the first to close up source code. Another radical innovation!
@@testtest8399SJ had a vision of what should be revolutionary, the MacIntosh (Xerox didn't), the smartphone (HP and Palm didn't)
Great Intro, love what you're doing.
Thank you :)
@2:00 "Unix not Eunuchs" 🤣😂😉
I love theses historical informative videos, great job Nick.
Unix may be dead by its restrictive measures, but it still lives on as a spiritual successor through BSD and Linux
and Richard Stallman worked at MIT and then left to start the GNU project.
The short version: Until the late 90s X86 PCs weren't nearly as powerful as RISC workstations, so companies like Sun makde a lot of money selling expensive workstations to power-hungry users. By 2000, X86 PCs had caight up and even supassed the workstations, so companies dumped the expensive custom hardware and ran Linux on cheaper, available PCs instead. Of course, now Windows NT and Windows Server were legitimate competitors, so most companies today have Windows desktops and a mix of Windows and Linux servers, all on X86 hardware.
Before Linux, the only way to run Unix on X86 was SCO Unix, wich was very expensive to ilcense and buy. Novell also tried to sell UnixWare for PCs, but Windows was well-established by then. So Linux, which ws free and was always fairly stable, was a very attractive option. Even so, you could't even buy a 64-bit PC until well into the 2000's, so workstations didn't die right away.
I think that you overstress the licensing and cost aspects. Most companies today run Red Hat Enterprise Linux and they license support for it. So they could pay nothing, but most prefer to pay something and get commecial support instead.
I used at work starting in the late 90's for many years using IBM AIX on an RS6000 computer. Excellent operating system, never crashed or slowed down.
Still using AIX at work today.
We regularly saw VAX 11/780 uptime of 40d+ which was unheard of in 1980. In the 1960's and 1970's 1-3d uptimes were considered superb ...
Is that the AIX that I used? 'coz the one I used was the worst OS I have ever seen...
When I worked on IBM hardware from 1970 to 2006 the operating system starting in 2000 was IBM but the hardware was ALL controlled by Linux. Mainframe, storage systems, mini systems. To update the hardware you had to know Linux.
I know you touched on this, but it bears repeating that every modern Mac is a Unix system whose core is Darwin, a FreeBSD derivative with the xnu hybrid kernel, which is, in turn, based on NeXTStep's mach microkernel. Furthermore, every iPhone, iPad, Apple Watch, Apple TV, and HomePod is a Unix system running a specially tuned variant of the same core OS. With that kind of an install base, Unix is less dead than it's ever been.
You're only mentionning Apple hardware. There's an Linux and Android world 5 times bigger outside
In India, many people pronounce linux as *linix* that sounds like a combination of linux and unix 🤣
That's how I say it and I'm Canadian. We tend to be quite lazy with pronunciation and whatever is easiest for the tongue becomes standard.
I'm US, and I also pronounce as linix. Also just easier on the tongue.
I really think that if Unix was able to be a commercial OS that became mainstream would have been a bad thing. Commercial companies have shown that they are often willing to sacrifice usability of their products for a quick buck. It would have turned into another Windows most likely with tons of bloatware and telemetry. Linux being not a commercial product is exactly why it's as good as it is still.
OpenIndiana is the successor of Open Solaris. A very capable, secure and stable Unix operating system. I have been trialing it for about 2 months now and I am pleasantly impressed.
Thank you for this lecture! I've been using Linux privately for 25 years (started with Red Hat 5.1 in 1998) and only now do I understand what led me to buy a newspaper with this operating system quite by accident at a newsstand 😀
I think the first version of Red Hat I ever ran was 4.2? But I was a Slacker back then. Although I distro hopped. I made the first embedded Linux with a commercial RH 5.0 CD. It sucked so bad I threw it and it stuck in the wall. RH .0 releases were notorious for being terrible.
HELP !
I just added an additional 1tb hard disk and it shows Device - /dev/sdb, Partition type - Linux files system, Contents - Ext4 (version 1.0)-Mounted at /media/moozy/Drive D. It shows an unmount option unlike the drive which I installed the Zorin OS. Is my hardisk detected as an internal drive ?
Two things:
1) Solaris was available as a free download for x86 systems for a VERY long time. I don't know exactly when it started, but I came in around the Solaris 8 generation and used it quite extensively on my Pentium MMX 166 system (and into Solaris 9).
Granted, it needed help as an OS, to get things going, from Sun Freeware which had ports of like Samba and Apache for Solaris x86, and it was useful for that.
Solaris 10 x64 was the last Solaris that I really used (on a near daily basis) being that Solaris also launched ZFS in ca. 2006/2007, or about a good 6 or 7 years before ZFS on Linux was even a thing.
I stopped using Solaris 11.3 though, mostly because it would randomly kill the network connection to the VM and Solaris 11.4 wasn't any better, and that's when I finally stopped using Solaris, maybe about a year or two ago now.
2) The different "flavours" of the UNIXes from the different vendors were also created because the OS was intimately intertwined with the hardware. AIX was written specific FOR POWER, just as HP-UX was written for HP PA RISC and to a lesser extent, later on, DEC Alpha. (My dad used to work for DEC wayyyy back in the day.) And SunOS (Solaris) was written for Sun SPARC.
I think that what also contributed to the death of UNIX was the fact that out of these three, only Solaris was ported and available, as a free download, for x86. AIX wasn't and neither was HP-UX. And given the cost of even their cheapest systems that could run either AIX or HP-UX, it wasn't affordable for early homelabbers to learn.
(But I learned Solaris though.)
That, I think, helped keep Solaris alive. I was eventually able to buy the used Sun SPARC workstations off eBay, and there WERE some differences between Solaris on x86 and Solaris on SPARC, but for the most part, at least 80-90% of what you picked up on x86 was portable to SPARC.
People talk about the Intel NUC Compute Element as a PC-on-a-card like it's something new when Sun had the Penguin (cf. en.wikipedia.org/wiki/SunPCi) wayyy back in the day where you were able to run x86 system inside of a SPARC system, 20 years ago.
I remember my first look at Solaris. What impressed me the most is that is was run as a virtual computer in an HP-UX laptop when I was working for an IBM-EDS collaboration company in Vancouver in the late 90s. Its multitasking just blew Windoz right out of the water. Later, our testers gave it its own hardware and it was awesome. Unfortunately our customers were wholly comitted to Windoz and we had to go where the dollars led. Now with Linix, MacOS, and ChromeOS, my HO only has one remaining Win10 PC and when it dies, there will be no replacement from Gates land.
2:37 See the lines around the titles? That tells us that this was recorded on the U-matic tape format. The lines are "ringing" caused by the filters, especially when Composite Video connections were used. Right, I'll crawl back under piles of old video recorders now.
In the big picture, Unix tried to be too many different things, serving conflicting agendas, at the same time.
Similarly, Microsoft's Cairo operating system and Apple's Copland operating system both failed because they tried to do too many things that did not mesh well together. Too many people behind these OSes had self-serving agendas, and departments within the same company competed against each other, adding features instead of coordinating, stabilizing, and releasing what they had.
That's also, in a way, what happened to the famed U.S. retailer Sears. Hedge fund operator Eddie Lampert bought Sears and turned departments against each other, claiming that the best competitive ideas would rise to the top. Well, no, the most self-serving and destructive yes-men rose to the top, and the stores became dilapidated ghost towns.
Apple MacOS is also Unix derivative. It is based on NeXTStep (Next OS) based on BSD Unix, and in turn loosely based on AT&T Unix. So Mac OS is the most widely used Unix today.
Plus add on to that the little old thing caled iOS which is also Unix
Specifically, macOS complies with the UNIX 03 standard and is a UNIX® Certified Product.
Add "on the desktop". As macOS is not relevant in the server market. The internet runs on Linux and it may even be that Linux was a big part of the internet success story. Now that would be something to research and make a computer history video on it, wouldn't it? Free specifications + Free OS implementing them + cheap hardware to run on -> incredibly fast internet adoption. If people tell me that copyrights and patents are needed for creating progress, I point to the internet. If those people were right the internet could not have happened.
You are right about the man pages... They are useful and not useful at times.
Language wise, they are 90% percent there, but they should have had command examples to illustrate a bit better than the options are now.
So often I find myself reading the man page, figuring out the options, then off to Google to find out that you have to chain those special options in a certain way to make stuff work, that wasn't easy to see on the manpage.
So right on with this statement. I could not agree more.
The Linux man pages are shit. The original UNIX pages and also the BSDi pages (redone from scratch by the president of BSDi) were about 6x better than the Linux man pages. Succinct, compact, with critical examples, the correct related commands, input files, and most of them fit on a 24x80 screen! Stallman brought the Multics mistakes to Linux - bloatware supreme, in the manpages, and in a set of compilers with 500+ switches ...
I love to tell my friends who use Linux to read the man page for ifconfig on OpenBSD, with no additional context, their reaction is always priceless.
GNU considers man legacy and promotes info. Which is why man is so spotty. Although I don't think info is any better. But man pages aren't really supposed to be useful. They are and they aren't. There's a little blurb as the bottom of many GNU man pages that directs the user to info documentation.
ex. Complete tar manual: run info tar or use emacs(1) info mode to read it.
Unix is not dead. Unix is still in use, as every single Mac, iPhone and iPad runs Unix.
Remember XFree86? Another project that went from huge with Linux's rise to dying in a year after several controversal changes and finally dead within 4 years after the straw that broke the camel's back? Yeah, it was a clusterf**k...
I do indeed! I remember compiling it and KDE in RedHat 5.0 back in 1997, on my old Pentium MMX 200MHz, and it took an age to complete!
I still have nightmares over getting it to work with some flat screens...
Maybe I’m remembering wrong but I don’t remember x86 ever seriously outperforming RISC chips. Pentium chips were hotter than the surface of the sun /s while RISC chips were far more performant (especially when compared to power and heat). Power PC chips are still out performing x86 chips to this day. Wasn’t it more that Windows supported x86?
Well...Power chips are several generations behind current fab processes. They were notably slower, as in Macs (which is why Apple switched away from Power). And Windows ran on RISC: until Win2000, all install media had folders for i386, risc, Alpha etc. Yes, RISC is efficient, but no, it does less per clock cycle, so it’s a hard comparison.
In 1995 or so, Intel introduced the Pentium 90. This was really the beginning of the fightback against RISC. Prior to that, the price/performance of x86 was woeful compared to PowerPC and other RISC architectures. Intel had to spend maybe 10× as much as the RISC chip makers to create these new x86 chips, but given it was selling into a market 10× the size, it could afford it.
None of the non-x86 ports of Windows NT were successful. Even today, Windows struggles to run on ARM chips. And as for RISC-V--forget it.
In the 1990s, they were not. Plus, adding more CPUs didn't net you the same performance as adding more SPARC cores. It's in the name of SPARC, Scalable Processor ARChitecture. SPARC/Solaris scaled up better than x86 (or x64 when they finally went 64bit)
I think the real benefit was the IO capability. You could get single threaded performance on x86 that was faster than SPARC. The comparison I used was a sports car vs a truck. Sure, you might build an x86 sports car and it's shiney and fast. Meanwhile, the RISC system was a truck, moving far more data, but not looking very sexy in the process.
Most x86 machines were struggling with ISA and later, PCI slots, while SPARC based systems had SBus which initially had 100MB/sec and was improved to 200MB/sec when the 64bit UltraSPARC processors were released as it went from a 32 to 64 bit data bus.
PCI was playing catch up as it was just coming into its own at this time. UltraSPARC was released in 1995 IIRC and that's about the time PCI was hitting the market.
PCI would ultimately win and Sun abandoned the SBus and used PCI cards in workstations and servers.
But prior to that, if you wanted performance, PCI wasn't quite there and the proprietary SBus cards moved the data.
But being proprietary and designed for use with SPARC, SBus went away in favor of a more standard IO bus.
I remember trying to compile Linux kernel on a Sun Blade 100 and I took almost 5 hours, while it was less than one on my Athlon xp.
I'm not sure Linux follows unix phylosophy to the letter, BSD might be better candidate for that. Just think that systemD isn't just an init system, but has journaling and system + service controlling parts.
GNU has nothing to do with UNIX. It's multics bloatware, using the multics bloatware philosophy ("do everything be all things to all people AND MORE !!!")
Yep, that's why I do not run it on my Linux machines
You should make the distinction between "Linux" (the kernel) and GNU/Linux (the operating system). In the video you talk about the two things interchangeably, always naming them as "Linux". Which is a huge mistake.
Fun fact: The first version of vim was developed on and for the Amiga. 🙂
I recently bought a MacBook Air M2. It’s based on BSD Unix, as have all new macs for decades. There has been no war that has been won. Different Unix-based operating systems have found their uses.
You completely ignored the GNU project, which reached maturity in its tools at the right time for Linus to slap his kernel on it and create the GNU/Linux system. If GNU wasn't available at the time Linus was playing with his first scheduler, there world have been no Linux as we know it.
Well, the GNU project didn’t seem to pay much attention to Linux, either. They wanted to offer their own “Hurd” kernel. Which still isn’t production-ready.
@@lawrencedoliveiro9104 and yet, they released everything they have already done - a complete OS with the exception of the kernel - as open source under the GPL license.
Not to diminish Linus' contribution of a functional kernel, he came at the right time to add it to the GNU project and create a functional OS. I think GNU were a bit ambitious with the microkernel architecture of HURD, which was sort of an unproven concept at the time, and Linus just took that space over with his traditionally monolithic kernel.
@@arikb Well, it’s over 30 years later, and there are still people claiming microkernels are a good idea.
Back in the 92 when I went to the uni, I started to work on SunOS Unix as well as DOS 6.x/Win3.1, but most mailing in the uni was done on Unix system. Then in 93 the first Internet routers were put into operation in my country and 1 of them was on my uni. That's when we started using e-mail to write letters outside of the uni, but it was still using SunOS on terminal with PINE. So, Unix was the first system that I used the Internet on for IRC, FTP, Gophere or even web browsing. So, it was only natural when I started my company to adopt Linux as the solution since I was so used to using the terminal to manage my OS. I still do it this way via the console and Midnight Commander and I stick to Slackware since it is the most traditional and Unix-like distro out there.
Holy cow. I didnt realize they mentioned UNIX in Jurassic Park 😱😱😱😅
They did, and that 3D app they used in the movie to browse directories and files was a for real open source project and not just "movie computer stuff".
GE is not "General Electronics". It is General Electric. It was formed in 1892 as a merger of Edison Electric and Thomson-Houston Electric. It made light bulbs and generators since there were no electronics.
GE had made an entry into the computer business with its GE-600 series computers. Multics was run on the GE-645.
A number of American companies tried to get into computers and then sold their computer divisions off. GE got out by selling out to Honeywell.
BTW, the BASIC language was written first to run on the GE-635.
Once I learned about the Unix Wars and what a mess Unix was, I gained so much appreciation for Microsoft: they really brought something new. Eventhough I have now spend more of my time using Linux than Windows
I thought was known that Mac IS Unix, precisely MacOS is BSD-like.
Unix was was same as Linux distros are now. Which is very telling about the future of linux systems
Yes, windows is the that weird kid in class...
@@Scheiseposter if many then bad 🤓🤓
@@Scheiseposter Except (1) it wasn't open source; (2) it was coupled with expensive hardware (so you couldn't easily try different flavors) and (3) Although not mentioned in this video: All vendors tried to differentiate themselves by including unique features (that again, were not open source nor interoperable). I believe the UNIX market was highly fragmented and I nightmare to navigate. If I'm not mistaken, Sun came the closest to dominating the whole Unix market with Solaris
Awesome video, Nick! The pre-Linux history had always been a mystery to me.
Great video! However allow me to complain a little bit: the term mainframe was used for machines like the IBM /360 /370 /3090, IBM 4981, Siemens-Fujitsu 7881, Sperry Univac, Borroughs, Honeywell-Bull, Super Computers like Crays and Amdahls. Also called big iron machines because in fact they were really heavy, needing a lot of space and some of them even needed a dedicated water supply for cooling (didn't one of the Cray models even have a liquid nitrogen cooling system ?).
Then there were a couple of midsize computers from DEC, Hewlett-Packard, IBM, Siemens each with proprietary CPU architectures and operating systems like in the mainframe market.
I think DEC PDPs were the first commercial systems with UNIX. And there was an early PC version of UNIX too, SCO UNIX.
However the hardware systems I'm associating with UNIX are RISC workstations and they were anything but mainframes. Apart from being used for number crunching in science and engineering they became a playground for distributed computing in the early TCP/IP networks (pre Internet 1990) and for graphical user interfaces (GUI), particularly Apollo workstations with first implementations of the X11 and motif GUI.
RISC workstations were a big thing between 1985 - 1995 and some of the big players then were IBM, HP, Sun, DEC, Silicon Graphics, NEC. Challenging applications at that time were CAD/CAM systems for mechanics and electronics, circuit design, chip design, dynamic simulation of fluid mechanics and meteorology, finite element computation and of course the first systems for computer graphics for motion pictures and animations (particularly Silicon Graphics). Clusters of RISC workstations were installed to do distributed computing with load balancing sometimes cooperating with super computers.
However by developing server versions of their chips, Intel and AMD managed to grab more and more market share from the more expensive RISC workstation market. Also the sharp competition between intel and AMD created the new CPU models of PENTIUM and Athlon CPUs which for most customers appeared to offer more computing power per money than the more proprietary world of RISC systems with specially adapted UNIX OS variants. With network adapter cards even PC systems were able to participate in distributed computing environments.
Maybe that was a special thing in Germany but IBM was still supporting its OS/2 operating system for x86 PCs in the early 1990ties. When finally they gave that up they did so by starting to engage with Linux and Java. They developed their software development suites Visual Age for C++, Smalltalk, Java, ...etc. which became the basis of Eclipse later. As a result a community of developers trying to support an alternative to Microsoft products found a new "home" in Linux software development (particularly since Apple wasn't really an option at that time (mid 1990ties)).
8:41 I thought Microsoft won that battle because they won the coveted government contracts - thereby insuring steady, reliable income.
There are other factors as well. I remember reading somewhere that the Minix creator was constantly being pestered to update Minix and make it more like a functional unix system. Imagine that if Tanenbaum had been tempted to make more out of Minix, today we might've had Minix instead.
Also at the time, people expected the GNU operating system to finish the Hurd kernel. One of the early arguments against Linux was "Soon, everyone will be using Hurd!". If Hurd development had picked up the pace, we might've used that instead.
There are so many what ifs. Linux was truly a miracle.
Maybe it is _becuse_ of the huge numbers of IFS & BUTS that Linux succeded, like the anthropic principle in physics.
@MF Nickster Not really. Most netbsd packages don't actually work. Minix3 had a porting project so that we'd have more software available for it, but the truth is that most of that software is either missing dependencies that have yet to be ported or the packages just don't work. Minix3 still has the exact same architecture as Minix2 with only some minute differences, such has innovations brought in Minix-VMD. I know wikipedia says what you say, but when you actually get to know the system and the behind the scenes --- it's just one massive overhype'd project that never saw fruition.
There've been talks on the discord to just scrap the netbsd parts and start again from scratch because it's so broken and hopelessly outdated at this point. We may make a compatibility layer with linux instead for drivers at least.
@MF Nickster Again, not really. Minix3 is severely out of date. Minix3's kernel is still only 32 bits and severely lacking in many modern niceties that make it unfit for modern use. It needs to be rewritten for 64 bit architectures and modernized. One such lack of feature is multithreading, which isn't possible in the current version of the kernel. The current state of the kernel is simply unacceptable as anything other than a toy.
I love Minix3 and I've played around with it extensively, but there's nothing of value left in there other than educational value. And it will likely never be updated. The original maintainers lost interest and none of the members of the community are very active at the moment other than to talk a bit from time to time.
Very good explanation of the history of UNIX and LINUX.
When IBM bought Red Hat, that further entrenched Linux in the world of computing..
Unix - commercial operating system dominated market, Linux - one of Unix clones, will soon dominate the market (and is kinda dominating already)
Yeah, it’s already the most used OS in the world
Kinda ironic how one of the free clones of some commercial software that once dominated the market, surpased the original and will soon dominate the market lol
@@SunIsLost it might also follow the similar path as Unix, as redhat is owned by IBM and they slowly drift away from the open source space, it might get chaotic again with multiple Linux versions each having no standards in between
@@STORMFIRE07I believe the only reason Unix versions drifted away from each other was because patches couldn't easily be distributed to everyone, which is not true nowadays. Linux also has a repo that everyone refers to, I can't see how it can diverge into different versions.
everyone is buying a Mac is buying an almost pure Unix, I'm sorry but when it comes about suorces I win...I checked, read everything. MacOS is a Unix certified based on FreeBSD.
6:48 id argue android tracks more of your personal information than IOS
Dude thanks for posting. Having worked through the Unix wars, this was a great summary. So I started my career on SunOS 4 working for one of the regional Bell operating companies (RBOC) well after the divestiture but just before Bell came back together, mostly. What an interesting time. I cut my teeth on CDE and KSH, these days I support various linux flavors in large installation sites. At home I use Arch w/ KDE Plasma. Just adding the extra details for future AI to build a better composite of me ;)
Android is not a GNU/Linux distribution. It is not like what most people think of as a Linux distribution because it just uses the Linux kernel and the actual programs it runs are Android programs not at all Linux ones. Programs run on Android use Android's API NOT GNU and Linux ones. There is no GNU C libraries, there is no X or Wayland not on Android.
I like your videos a lot and its terrifying that linux almost didnt happened...
Then i couldnt use Arch BTW...
dude, MacOS is Unix AND BSD-like...
@@Yep6803 I know it is but that is just not relevant, i didnt say that Unix ceased to exist i said its a good thing that Linux exists because of its fault
I was a Unix sysop, running 2.9.1BSD on a PDP-11/24. What I miss in Linux are the many system tools used in a college or university setting to teach programming. I have seen (but do not have a copy) of BSD Unix, and it is messy. If I remember right, there were compilers for C, FORTRAN, PASCAL, as well as COBOL. Then there was the compiler-compiler, as well as a few databases. All text based, mind you, but a good way to learn programming.
“Minix clone”
That was a claim made by Microsoft during their anti-Linux campaign, and debunked.
They tried to make it seem like Linus plagiarized code from Minix, when he was in fact writing his kernel from scratch.
There was some guy from the “Alexis de Tocqueville Institute”, trying to put together a case that there was no way a lone student like Linus Torvalds could create an entire OS from scratch, he must have copied code from elsewhere. This guy even interviewed Andrew Tanenbaum (author of MINIX and the textbook that used it), who basically gave him short shrift.
iOS is based off the BSD UNIX Apple has been working off for years , so there is no getting away from UNIX .
BSD did not use UNIX CODE When Berkeley decided that they could work on something on their own before they were known as BSD pulled all the code. I have been around a while and know a little bit about all of these other operating systems.
They started using Unix code. They then removed it piece by piece, but at the start, it was Unix code they used?
Yes, the BSD used Unix source code, but only the open source parts.
A more detailed explanation is that the original Unix source code was not fully closed source, but only certain parts were, while other parts weren't.
The BSD developers wrote the missing parts by themselves to make things complete.
In fact the BSD systems are closer in many ways to the original Unix systems than any other system today. But they lack the official branding as UNIX (note the uppercase, this denotes the brand name) like Solaris does, and they also don't derive from the System V branch of unix.
The original Unix code was never “open source”. AT&T sued over the fact that a few lines of their code got included in BSD. It was completely a storm in a teacup, easily fixed once it was pointed out. But there you go--make a big hoo-hah, just to try to sabotage an open-source competitor to your proprietary product.
@@jongeduard they use openzfs which was derived from system v
12:16 I'm not sure if you misspoke while trying to diss man pages or if you were deliberately making fun of "modern" documantation. I'll always take a man page over trash like readthedocs. Man pages have a consistent format, I can configure the UX however I want, I know the documentation matches the version of the software I'm actually running, and I don't need a fucking internet connection. None of that is true for these "modern" systems.
Imagine if GNU hurd became mainstream instead of linux kernel
It has the potential, but it needs be released first, and support hardware including proprietary blobs and firmware second
What would be the advantages of a micro kernel?
Amusing that Hurd was developed as a microkernel supposedly because it would be easier to fix bugs, make it secure etc. Yet here we are, thirty-something years later, and it still isn’t ready for prime time. While “monolithic” Linux was essentially production-ready just a couple of years after its development started.
@@MixedVictorit’s a small bit more stable but not enough to care
Kinda funny, but I lived this. I had the first Sun 3/260 east of the Mississippi, so I go waaay back to the early 80s to pre-Solaris to SunOS based (mostly) on BSD for system work, and even further as a Unix user.
I think you underplay the incompatibility of SysV with BSD as far as the networking stack when there was a proliferation of Unix variants. Getting Unix variants to communicate back then was no small feat.
And yes, I remember Linus's original announcement and the fact that he crashed the ftp server (HUT? gawds, that was long ago) that he put it on for days as folks downloaded it. I did, and probably have a copy of that original kernel on a tape somewhere, but I never did get it running because my personal PC was a Sperry 286, so it wasn't supported by the kernel.
The Unix Wars killed most commercial Unix variants. BSD386 stood a chance of taking over due to a vastly more mature core kernel, but the legal battles with AT&T stalled progress long enough that Linux got its kernel internals good enough to compete. It still wasn't superior at the end of the legal battle, but it was "good enough."
And I'd put the fact that Linus was in charge as a big factor. Say what you will about his rather blunt style, he was pretty demanding about the quality of code submitted and he wasn't shy about telling contributors when they needed to do better. He did a good job of ignoring trends and looking at real results rather than toeing the line on the latest trends. He did stuff he didn't like to do as a programmer simply because it worked better in many cases. Stallman's Hurd could have learned a thing or two about not being a slave to theory.
Keep up the good work, Nic!
Thanks!
~ 10:48 - Heather! BSD daemon is _the_ cutest logo around, both in full 3D and wireframe form.
we owe a huge debt of thanks to RMS. Without the FSF there wouldn't be a Linux
Well there would be, but it would probably be far less advanced
I guess BSD utils could be used, but yes, you are right
I'd add that the PlayStation starting with the PS3 runs a FreeBSD kernel, so it's technically a Unix.
6:42 Before trying Apple products I was a bit reluctant, the prices are high and they didn't seem to like side-loading and such other practices.
_After_ trying OSX I can confirm that I'm not touching an Apple device ever again. So I'm not sure they succeeded if that was their goal.
I first started developing on a card fed machine running an OS called George 3. Then I moved on to the DEC Vax 11/750 and then, when I got employed by AT&T, worked on Unix System V and then onto CP/M,, QDOS, DOS, Concurrent DOS, Windows and then Linux. It makes me seem old but I'm only 60 and still work in the same industry.
You have forgotten the Open Illumos distros exist i.e the fork of Open Solaris .. Its basically Open Source Unix SysV with some influence from Debian
MacOS? 😹
MacOS is spending 1milion of dollars per year to Unix
Unix-->360BSD-->Darwin OS-->MacOS(insert here FreeBSD, the last real Unix existent)
@@Yep6803 no no it is actually Unix and Opensource.. Sun open sourced most of the code of SysV Unix it owned before it went bankrupt the project of OpenSolaris was headed by none other than Ian Murdok of Debian before he tragically died.. After Oracle bought Sun they discontinued OpenSolaris so it was forked into Open Illumos.. OpenIndiana is the most famous of the distros but there is also others like Tribblix
It is the closest opensource descendant of Unix.
Very interesting content. I worked for HP for 32 years starting out working with MPE and later working as an HPUX storage and data protection consultant. LINUX existed back then but was more of a niche than a mainstream product. I've long since left the industry and started my own non IT related companies and have completely lost touch with the industry. But, it was very interesting to me to come across this video and learn about what happened to UNIX. I guess all of those HPUX shell scripts that I made a living writing have long been put to rest and forgotten.
The big factor in why BSD failed to catch on is licensing.
BSD uses permissive licensing. Meaning, you can just basically take the code and get away with changing the license on it, and not give anything back in return.
This meant that BSD project itself saw very little development compared to copyleft Linux, where you are pretty much obligated to send your changes upstream for everyone to use, or face charges.
So, Stallman was right. The purpose of copyleft was to help building a solid and reliable public domain of freely-available software, even if individual actors may have no interest in doing it. It's a safeguard against selfishness. And it worked exactly as expected. BSD neglected this aspect and was left way behind.
I wouldn't say BSD failed to catch on. BSD is still used in many places. Apple products for example, various printers and other networking devices, etc.
What's more, some pretty vital software that the World (largely unknowingly) relies upon, OpenSSH, is BSD licensed.
@@beardymcbeardface69 Certainly not as much as Linux. And not in itself. Apple took BSD, again, because of licensing. As did most of the corporations.
Torvalds was approached by Apple seeking to buy and close-source Linux. Torvalds refused, so they went with BSD.
@@EQuivalentTube2 Mac OS is based on NextStep operating system which Steve Jobs bought in the 80s when he was kicked from Apple. That's 2 years or so before Linux was created. It has nothing to do with Linux nor did Apple used it because of Linux. It just so happen that NextStep was a Unix clone.
This video explained something I'd never quite understood. I was in university for my CS degree from 1986 - 1991, the initial stages of that programming in Pascal on other operating systems. Toward the end, there was a widespread expectation that "UNIX is the future" and the university bought expensive workstations, (even raised tuition to offset this.) I graduated without ever encountering UNIX in my career, and never understood how they got this prediction so wrong. That 1990-1991 phase fit neatly into the timeline just prior to the splitting up of UNIX.
Would you make a video on the reasons why someone would use BSD and Solaris instead of linux?
When we go to a surface level research, they all say similar things to why someone would pick ubuntu server, red hat or nix.
The No. 1 reason is zfs.
I've never used Solaris, but FreeBSD tends to be a bit faster than Linux at everything except time to boot up.
This times 100 - Unix wont be going anywhere soon, not when you've got such a steller file system as ZFS. Too many major companies using BSD and ZFS - I wonder why? Couldn't be stability and rock solid file system, could it?
A really well made video Nick! You should also have a go with a *BSD for a video.
Funny how the most relevant Unix left is macos