Did Linux Kill Commercial Unix? (howtogeek.com) 280
When Dave McKay first used computers, punched paper tape was in vogue, "and he has been programming ever since," according to his biography page at How-To Geek. It adds that "His use of computers pre-dates the birth of the PC and the public release of Unix."
Now long-time Slashdot reader sbinning shares McKay's "short history of UNIX and how Linux got its start," which ultimately asks if commercial Unix was killed by Linux: Unix is still out there, running mission-critical systems that are functioning correctly, and operating stably. That'll continue until the support for the applications, operating systems or hardware platform ceases. If something's genuinely mission-critical and it's working, you leave it working. I suspect someone, somewhere, will always be running a commercial UNIX or Unix-like operating system.
But for new installs? There are enough variations of Linux to make the case to go for a commercial Unix very, very difficult.
Now long-time Slashdot reader sbinning shares McKay's "short history of UNIX and how Linux got its start," which ultimately asks if commercial Unix was killed by Linux: Unix is still out there, running mission-critical systems that are functioning correctly, and operating stably. That'll continue until the support for the applications, operating systems or hardware platform ceases. If something's genuinely mission-critical and it's working, you leave it working. I suspect someone, somewhere, will always be running a commercial UNIX or Unix-like operating system.
But for new installs? There are enough variations of Linux to make the case to go for a commercial Unix very, very difficult.
Probably didn't help (Score:3, Informative)
It's not like Linus came up with the idea of free software.
Are we forgetting mac OS? (Score:5, Insightful)
Mac is certainly commercial Unix even if it's based off of BSD.
Re:Are we forgetting mac OS? (Score:5, Interesting)
FreeBSD has been around since 1993. OpenBSD and NetBSD forked soon after.
In the early years, FreeBSD was more popular and more stable than Linux.
BSD lost ground to Linux for several reasons:
1. The AT&T lawsuit.
2. The earliest BSD kernels required hardware floating-point which low-cost 486SX computers didn't have
3. A toxic culture that was openly hostile to newbies.
Re:Are we forgetting mac OS? (Score:5, Insightful)
A toxic culture that was openly hostile to newbies.
That's true of most open-source projects in my experience.
Re: (Score:3)
No, only the projects which cater to experts. I.e. you find the Ubuntu communities far more tollerating than those of Arch or Debian, and the Devuan guys seem to be outright arseholes at times.
Same on the app side. Got a problem with Inkscape, the community is quite helpful. Got a problem with iptables, RTFM & GTFO N00b.
Re: Are we forgetting mac OS? (Score:3)
I got to say, out of all the communities, The Haskell one is by far the friendliest I have met. Probably due to the difficulty of the language requiring a certain minimum IQ. :)
Re: (Score:3)
No.
The kernel in MacOS is the XNU kernel whose core is Mach with 4.3BSD file system drivers and network stack and an Apple developed device driver subsystem written in C++. There was also a thin layer to provide BSD style device files. Apart from Apple's device driver subsystem, it all came from Next.
Re: (Score:3)
Without going deeper into the debate, the 1st mention of both FreeBSD and NetBSD I could find on Usenet was on July 1993.
https://groups.google.com/g/comp.os.386bsd.misc/c/0BOG_ZNBv2Y/m/cyrQD1jQ9XkJ
Re: (Score:3)
Found an even earlier reference to NetBSD on May 19th, 1993.
https://groups.google.com/g/comp.os.386bsd.questions/c/PniapyOpRKQ/m/iOrsSxp0p8oJ
There are earlier references to freeBSD, but these mostly refer generically to a free version of BSD (mostly to distinguish the free 386BSD from the BSDI commercial 386BSD). They do not specifically refer to a project named "FreeBSD", or simply propose this name doe 386BSD.
https://groups.google.com/search/conversations?q=freebsd%20before%3A1993%2F6%2F1
don't be ridiculous (Score:4, Insightful)
Re:don't be ridiculous (Score:4, Interesting)
It's "consumer-grade UNIX" at best. The userland is around 12 years out of date. Not sure why you got downvoted, that's trivial to verify. Just look at the dates on each tool. It varies, but most of it has remained untouched, and compared with a current FreeBSD or Linux system, it's missing all of the bugfixes, new command-line options and features developed since then. Its compatibility and interoperability with other UNIX systems has been compromised as a result. It's all too common to have to special-case MacOS because it's missing trivial options. Like "readlink -f".
It's a shame because you're absolutely correct that in the early days of MacOS X, its UNIX base was a big selling point. But that clearly hasn't been a priority for quite a number of years.
Re: don't be ridiculous (Score:3, Interesting)
User land got abandoned with the introduction of GPLv3. You can't really blame Apple for not wanting to be dragged into that cesspool of a license.
Re: don't be ridiculous (Score:4, Informative)
Err, the userland is from FreeBSD. It's all BSD-licensed. The only bits that weren't were things like the bash shell (now zsh by default) and the GNU toolchain (now LLVM). I think you can make that argument for some specific GPL-licensed parts, but the vast majority is BSD, and for those parts the argument doesn't hold up. They had nothing restricting them from updating this stuff, but they chose to let it stagnate.
Re: (Score:3)
GPL3 only matters if you are using GNU tools rather than FreeBSD. They've just not bothered, opting to allow anyone that cares to update with Homebrew and get the GNU versions.
The real killer (Score:2)
The real killer was however Microsoft pushing Windows for everything and using incompatible apis and even listing posix api calls as unsafe (some are though, like strcpy), but they scared a lot of management people enough.
Then the Unix vendors forgot the origins of Unix and never saw the writing on the wall that if they could offer the OS for free on their hardware they would gain market shares and make a profit on added extras. Instead Linux took that place on servers, but Linux was just the nails in the c
Re: (Score:3)
I think the intellectual property lawsuit in 1992 (UNIX System Laboratories, Inc. v. Berkeley Software Design, Inc. [wikipedia.org]) changed the course of history, because uncertainty in whether open source BSD was legally feasible offset the uncertainty of whether this Frankenstein's monster of a Unix called "Linux" was technically feasible.
I know I was all set to install 386BSD but concerns over the legal future of the project led me to go with Debian 0.9 instead. For many years BSD on Intel was more mature than Linux b
Re: (Score:3)
Yeah, a better headline would be "Commercial Operating Systems Are Dead."
People keep using macOS as an example, but it's free with the hardware and upgrades are free. Is Microsoft still charging for Windows? I know they charge OEMs, but I thought you could just download it for free at this point.
As you mention, if Linux did not become so popular, the BSDs would just occupy that niche. Given enough time, the price of any widely used commodity software approaches zero. There may be exceptions, like games, but
Re: (Score:3, Interesting)
The BSDs were hindered by their licenses. the GPL on the Linux kernel kernel helped protect it from proprietization much more effectively than BSD licenses, as did the GPL for gcc and glibc.
Re:Probably didn't help (Score:5, Informative)
Linux came out, and had a majority of a year's head start before Jolitz's 386BSD was released. Note: 386 BSD is not BSDI's BSD/386, which became BSD/OS. Linux won because it was out earlier, and people could easily port Minix stuff to the OS. In fact, the first filesystems for Linux were Minix, until extfs and xiafs came out. There was even a filesystem, umsdos, which allowed FAT to handle UNIX permissions.
That head start helped, and the fact that Linux didn't care about a FPU. Just load the emulation model in the kernel, and call it a day. The fact that GNU had a wide set of tools, and people started throwing together distributions didn't hurt either.
Another advantage with Linux was that it wasn't being fought over between USL and Berkley's lawyers, so colleges and other places didn't mind using Linux while the fact that BSD's fate was completely unknown until the case was settled in 1994, made people shy away from it at that time.
As for killing other UNIX distributions, there were TONS of those things in the late 1980s. Dell had their own SVR4 variant, and there was XENIX, SunOS/Solaris, IRIX, NeXTSTep, ULTRIX, BSDI BSD/OS, A/UX, AIX, AIX/PS2, HP-UX, and even real time operating systems like QNX were getting into that game. Back then, you paid $500-$1000... just for the operating system. If you wanted to compile code, better reach deeper in that wallet because a compiler would cost you another $1-2k. Then, some UNIX flavors required licensing for more than two users (AIX comes to mind). The average person just didn't have the cash for a UNIX. All the UNIX players didn't really care what was going on, because they were competing with each other, and not seeing the forest for the trees. They didn't think a GPL licensed operating system would take hold, because they really felt that PCs were toys, Linux was a toy OS, and that "people with real compute needs" would always go for IBM iron, Alpha, MIPS, or "big boy" CPUs, as opposed to use something so "pedestrian" as a PC-AT clone which only ran "toy operating systems". They didn't realize that Intel and AMD were gaining on them slowly, but surely.
Microsoft was another factor as well. With all the PCs coming into offices with Windows 3.x, then Windows 95, the need for a UNIX OS in a lot of niche areas vanished. People who used to use FrameMaker on IRIX moved to MS Office. People who did WordPerfect on AIX moved to Microsoft Word.
Had the "big boys" realized that their main competition would be commodity PCs that could be bought cheap and stacked deep (the classic Beowulf cluster), and acted upon it, things might be different. However, they battled for markets which just started shrinking to nothing in the mid to late 1990s, leaving only Apple, IBM and Sun/Oracle by 2010.
As for a commercial UNIX in the 2020s, other than MacOS, it seems to definitely be a niche market, other than either legacy stuff, or embedded operating systems.
Overall, it is sad that so many UNIX variants didn't make it, but on the other hand, not having to pay big bucks for the OS, compiler, libraries, or other things that we take for granted now, sort of makes up for it.
Re: (Score:3)
The fact that GNU had a wide set of tools,
And the GNU tools were (and still are), frankly far far better. They had way more options for getting useful stuff done whereas the BSD people insisted on some misplaced notion of "purity", because apparently convenience and power was for lusers.
They didn't realize that Intel and AMD were gaining on them slowly
Worse they didn't realise Intel and AMD were gaining on them *quickly*. In 1993 PCs were "toys" according to many with dinky little 486s at best with "slow"
Linux, or SCO Group? (Score:5, Insightful)
I can't say for sure whether Linux was the reason that commercial Unixes died, but I have to think that money-grubbing sociopaths like the company recently known as The SCO Group. And Betteridge agrees with me!
Re: (Score:2)
Wow, I really failed at proof-reading that comment before posting it. "money-grubbing sociopaths ... had something to do with it."
That company's earlier name -- Caldera -- was a much better fit, given that almost all they produced was noxious gas and hot air.
Comment removed (Score:5, Informative)
Re: (Score:2)
I can't say for sure whether Linux was the reason that commercial Unixes died, but I have to think that money-grubbing sociopaths like the company recently known as The SCO Group. And Betteridge agrees with me!
I think the root cause of commercial Unixes dying is server variants of Microsoft Windows (and maybe OS/2?). They pushed the idea of ease of use and that they could be administered by college kids rather than bearded unix gurus. How true that was is open to plenty of debate, but it certainly sold to management. That eventually pushed commercial Unixes into a niche (or really a set of niches). Linux ended up pretty much eating those niches, but that was just a matter of eating the leftovers from the main cou
Commercial Unix killed Commercial Unix (Score:3, Interesting)
Have you ever tried to use commercial unix? It doesn't run on PC hardware. If it does, it has almost no drivers. The installers are crap. The package managers are worse.
Trying to get the software you actually want to use on it means boostrapping an entire GCC tool chain from scratch (or using 10 versions out of date vendor-supplied packages that install in weird locations).
The service managers are inscrutable (like worse than systemd inscrutable).
If it breaks there is nothing useful logged anywhere.
It's honestly a thousand times worse than running Windows in nearly every way, and that's saying something. Of course no one uses these horrible horrible operating systems anymroe.
Re: (Score:2)
There are a lot of Linux fanboys out there who can't read documentation and think that all the Unix variants will work just like Linux. They break systems, mangle installs and blame the OS. It's not the OS - it's you. Read the documentation, follow the instructions, and it works.
The only annoyance I've had is trying to get linux-specific software working on Unix using the various compatibility layer software that vendors provide. Most of the time it works, but sometimes you have Linux kernel-specific or
Re: Commercial Unix killed Commercial Unix (Score:2)
Re: (Score:2)
I can confirm. The last commercial Unix I remember using that used PC hardware was in the early 90s(only had a terminal) and it was SC* Un*x(I refuse to type their name for fear of them reviving a lawsuit). After that the only commercial Unixes that I used would have been on a SPARC or an SGI.
Re: Commercial Unix killed Commercial Unix (Score:2)
Re: (Score:3)
Re: (Score:2)
The last commercial Unix I remember using that used PC hardware was in the early 90s
There was Coherent OS on 286/386, UNIX Like and very cheap, I actually was using that until Linux killed it too
No TCP, but was very good for the price and for the time.
Re: (Score:2)
AIX wasn't THAT bad. That weird thing underneath that would revert system changes on reboot if not made with SMIT/SMITTY was a wee touch bothersome.
Re: (Score:2)
Trying to get the software you actually want to use on it means boostrapping an entire GCC tool chain from scratch
Didn't some commercial UNIXs started changing extra for the development set ? I remember people putting gcc on Sun (I think) because they did not want to spend thousands for cc
My first job... (Score:3)
Re: My first job... (Score:5, Insightful)
You also forgot the obvious.
Sco Unix, solaris, And other unixs where expensive. 1,000's of dollars per cpu or other weird metric.
Why pay 10,000's of dollars for an install when Linux had all of that you could use the same techs for the most part and only have to pay hardware and some side support if you needed/wanted it.
Re: (Score:2)
Not quite. Most Unix software licenses were free back when Linux started making inroads into the server market. Unix hardware, however, was very very expensive. The difference in functionality, reliability, scalability, speed and performance per watt was obvious to anyone who knew anything about hardware. But cheaper hardware almost always wins, hence why Intel/AMD servers beat out specialist Unix hardware once reasonable multi-core multi-processor boxes came out.
Re: (Score:2)
Quasi true. I do remember that Sun and SGI wanted a big fat license fee for a real compiler and linker (rather than the crippled linker needed to add a new driver to the kernel). Or you could get a 30 day trial license and bootstrap your very own GNU suite.
The hardware had some catching up to do and it did.
Re: (Score:3)
From fairly early times you could download precompiled binaries for gcc for SunOS. I had gcc on my 3/260, which I upgraded to a 4/260. In both cases I started with a binary and then built the latest version.
But yes, Sun did want money for their compiler. It did produce substantially faster code than gcc on SPARC back then, and most organizations that needed it could well afford it. Where I worked that had it (Silicon Engineering, in Scotts Valley) we had literally just one license. But we used the amd autom
Re: (Score:3)
Hold on: Solaris was sold as CPU license. It came with the computer.
Source: I used to be a Sun VAR.
Re: (Score:2)
The SCO that actually made Unix for servers had almost nothing in common with the bastards you mention. The only shared thread was ownership of some of the intellectual property. The actual businesses, whether in terms of owners, employees or customers, were fundamentally non-overlapping.
If we want to really have fun pointing fingers, I remember Xenix running in a doctor's office in the late 1980s. Microsoft Unix! Now that was a combination for the ages.
Re: (Score:3)
Re: (Score:2)
I also used to service Xenix systems in doctor and dentists offices in the early 1990s, later upgraded to SCO.
The SCO lawsuits have to do with IBM contributing copyrighted and trade secret SCO source code into the Linux kernel. SCO weren't the bad guys everyone says they were--they were just trying to protect their IP that was blatantly stolen by IBM. It's true! Look it up.
Re:My first job... (Score:5, Informative)
Some of us remember what happened in that trial [wikipedia.org]. SCO lost on, or dismissed, every claim they made. Most of the IP they claimed to own were actually IBM inventions, and they had a nutty idea that because IBM put implementations of those inventions into AIX, SCO could control them. SCO also had a theory that the GPL was unconstitutional, because that would have (at least in their twisted thinking) meant that SCO's distribution of Linux didn't constitute distributing it under the GPL.
Yes, the case is still somehow alive in appeals court -- thanks, terribly slow US legal system -- but SCO very plainly were the (criminally?) insane bad guys here.
Re: (Score:3)
Does iOS count? (Score:2)
Re:Does iOS count? (Score:5, Insightful)
It was originally based partially on BSD, right?
I think it is safe to say that the number of people who were considering paying for a commercial Unix package, and then decided to go with iOS instead, is zero.
Re: (Score:3)
Alternately, everyone with an Apple product is in fact a commercial Unix user.
As much I I'm not an Apple fan, credit where it's due. They adopted an OS designed to make computers as reliable as appliances, and used it as the foundation of some very nice computer-powered appliances.
Re: (Score:3)
There is more to it than buying a license, you have to certify your product against the Open Group Unix standard, of which there are several.
This costs a bunch of money, so that is why few Linux distributions bother. The Windows NT Posix subsystem is not even remotely compliant, so it has never been Unix certified.
For a list of Unix-certified OS:es: https://www.opengroup.org/open... [opengroup.org]
Huawei EulerOS 2.0 is certified Unix, and is Linux based. MacOS is regularly certified when a new version comes out.
But some me
All about Price (Score:5, Interesting)
It was the cost, outside of Linux in the early days, you had Xenix for PCs was around 2000 USD. BSD was stuck in court, so every went to Linux and improved it quickly while all commercial UNIX fought among themselves and laughed at Linux, charging Immense amounts for hardware and the OS.
This repeated with Motif, Linux was looking for a good X toolkit and people were begging OSF(?) to release Motif under a decent license, but they went for short term gain too. Look were they are, in reality no one cares about Motif either.
This repeats across many Industries too. So answer, yes.
Re: (Score:2)
It didn't take long for them to get replaced by lesstif. Then that got replaced too, but for different reasons.
Re:All about Price (Score:4, Insightful)
Yes. Once the horses bolted, they closed the barn door.
Why does court matter? (Score:2)
Every once in a while I see a statement like this wonder, why were the proceedings affecting anyone.
I picked up FreeBSD-1.1.5.1 circa 1994 without knowing anything about any legal challenges until years later. It installed on my i486-33 and I haven't looked back since — why did this court you're talking about matter to anyone else?
Re: (Score:2)
why did this court you're talking about matter to anyone else?
People and small Companies were afraid to invest time and resources in BSD in case they lost. Hardware was changing very fast in those times, so with people being leary to invest time Linux was able to advance quickly
Re: (Score:2)
Because the battle started about 1987 with the first "amateur" BSD project... BSD386? 386BSD?
They were so afraid of the looming suit from AT&T they restricted who could play in their sandbox and told Linus to go pound sand... And Linux was born.
Re: (Score:2)
I was barely paying attention to Linux at the time, but perhaps it reduced its level of commercial support? As a home or hobbyist user, nobody cares about you, so you don't really care. If your OS goes away, you eventually get another OS your your next computer.
However, if you're a business - an Amazon or Google might not be keen to build their software stack atop a foundation that might be yanked out from under them or suddenly require them to start paying steep UNIX licensing fees. And I would imagine
If not Linux, it would be Windows (Score:5, Informative)
There was no way commercial UNIX would continue to be successful in their niche markets. Like the Amiga of the old, they had to give way to more generic and more capable alternatives.
Remember SGI? They had the best graphics hardware out there. However when nvidia ate their lunch, nobody would use IRIX anymore (their brand of UNIX). Similarly, Solaris went out of the way when AMD and Intel brought massive cores onto server platforms. They were hardware companies first, and OS second.
The only holdout is IBM, and they have actually been very smart on evolving AIX to meet the modern needs.
Re: (Score:3)
The funny part is that Nvidia started life as a bunch of SGI engineers didn't like where SGI was going (including the deadly embrace of Windows and downsize fever).
Re:If not Linux, it would be Windows (Score:5, Interesting)
No, Nvidia's founders came from LSI, AMD, and Sun. You're thinking of 3dfx.
killed what? (Score:2)
Re: (Score:2)
No, MacOS is not UNIX. It is some conglomeration of a Mach kernel with a BSD wrapper around it and their own layers of stuff on top of that.
The article wasn't so much about Unix being dead, as about commercial Unix being obsoleted by Linux.
Re: (Score:3, Interesting)
The kernel is quite irrelevant, actually — unless you're writing device drivers. Of course MacOS is Unix — "layers of stuff on top" is what you'd say about X11 too.
What killed commercial Unixes is Intel/AMD — offering far more powerful computers than anything "commercial Unix vendors had". By the time Sun et al realized, what's happening, and made their own x86 ports, it was
Re: (Score:3)
"The kernel is quite irrelevant, actually — unless you're writing device drivers."
And who needs device drivers anyway?
Re: (Score:2)
How many people are writing them would've been a relevant question...
Re: killed what? (Score:2)
Re: (Score:2)
Re:killed what? (Score:4, Informative)
Re: (Score:2)
Re: (Score:2)
Re: killed what? (Score:2)
macOS is indeed UNIX. http://flying-geek.blogspot.com/2020/03/apple-makes-great-unix.html
Comment removed (Score:5, Informative)
Re:Linux is Unix (Score:4, Informative)
Killed the need for all our SGI stuff. (Score:2)
I know the archive doesn't go back that far... (Score:4, Funny)
The KO blow to UNIX was when Slashdot invented Red Hat and pitched it to Linus. At that point, we mixed the insurance that Microsoft had at the beginning, with Linus' control of the operating system and what stays and what go. Priced at exactly the cost of Windows, Red Hat Enterprise Linux had everything it needed to impress the bosses, so that's where UNIX got replaced with Linux.
The death of proprietary workstations killed UNIX (Score:5, Informative)
Commercial UNIX flourished in the days of the "workstation", as epitomized by the first SUN workstation running SunOS, when every hardware platform had its own private UNIX OS, with its own unique culture and flavor. When 32-bit hardware made it down to the PC market in the late 1980s, bespoke workstations died a rapid death, taking their "private" UNIX flavors with them.
In the early 1980s, UNIX System V was the effort to get all UNIX systems to pursue greater commonality. But it took several years to be ported to a majority of platforms. On my SUN workstations, I personally grieved the death of "Berkeley" SunOS when "System V" Solaris arrived. But by then the writing was already on the wall, and no workstation renaissance occurred. The PC simply got too powerful too quickly, including multi-socket servers.
Mainframes hung on to their proprietary UNIX flavors for nearly two more decades, eventually running Linux either natively, or over a thin compatibility layer.
However, the hardware aspect is only part of the story: The UNIX System V license was fairly restrictive in many ways, both to ensure sufficient revenue to continue development, and to strictly control the ability of vendors to create incompatible versions. The alternative presented by Linux was like a breath of fresh air to vendors racing to market. Early worries concerning the fragmentation of Linux fell away as the sheer code mass of Linux made contributing back to the tree faster and easier than forking it. This feedback effect "floated all boats", where all vendors benefitted when a new Linux kernel was released. System V was never able to generate this synergistic effect.
Re: (Score:2)
Yes, that's pretty much how I remember it.
Re: (Score:2)
You bring back memories. In the 80s I worked at a small software shop and kept porting our software to different Unix platforms. After I left, I worked as a contractor 'porting specialist', even though my knowledge was mostly of the seat of the pants try this try that variety since I seldom had good docs. I do remember how much I liked the Berkeley Unices, 4.2 and 4.3 and how I disliked SystemV, and with the grand unification it seemed like they threw out all the best parts of 4.2.
To address the question
I had a different take on System V (Score:3)
In the early 1980s, UNIX System V was the effort to get all UNIX systems to pursue greater commonality.
I always thought it was more an intellectual property issue, with the proprietary status of Unix on shaky ground. As I understood it:
1) Copyright (in the US) hadn't been extended to software when it was written, and changing the law to cover software (as was later done) doesn't retroactively cover
2) Patents weren't interpreted to cover software until still later.
3) So trade secret was the only normal IP p
Did Free Unix Kill Commercial Unix? (Score:2)
FreeBSD NetBSD OpenBSD ...
Didn't need Linux when Unix was becoming free anyway.
macOS - certified UNIX (Score:2, Informative)
Re: (Score:2)
Powerful PC hardware killed commercial Unix (Score:5, Insightful)
In the old days, if you wanted a server with a bunch of fast CPUs, you bought a Sun machine, which ran Solaris. If you wanted something with great IO and 3D hardware, you bought an SGI machine, which ran IRIX.
Now you can buy a cheap Dell with a bunch of fast CPUs, great IO, and powerful 3D hardware, and it's not locked in to the hardware vendor's OS. It runs Linux or Windows or any of the BSDs. It'll run OpenSolaris/illumos if you want, but Linux and *BSD are pretty darn good and familiar, and so they're what most people use.
As for the merits of the OS itself, commercial Unix had some neat tricks, especially for storage and virtualization. However, the free Unixes have a much larger variety of packaged software and libraries. I think a lot of people prefer them for that reason.
Re: (Score:3)
This -- I watched it happen from within the electronic design automation (EDA) industry.
In the late 90s/early 00s, if you wanted a serious EDA workstation you were buying or leasing Sun hardware. I was writing placer and router code at a startup on a Sun Ultra 5 on my desk in 1999. A few guys were lucky and had Ultra 10s, and our main server was an Ultra 30. There were a few firms out there running HP-PA and AIX workstations, but Sun was out bread and butter.
A few years later, my Ultra 5 was replaced by a B
The horrible truth (Score:4, Funny)
Linux also killed Minix.
Re: (Score:2)
MINIX was an OS as much as a Rockwell AIM65 was a PC.
It was a teaching/learning test bed and deliberately hobbled to keep it's users/students from being distracted.
Re: (Score:3)
Anyone running an Intel x86 system built in the last couple decades has MINIX running under the covers as the Intel Management Engine.
Correction: Killed proprietary Unix (Score:2)
Linux is now the bedrock of a multi-trillion dollar economy, what's not commercial about that? What Linux killed was proprietary, fragmented Unix. With a better product on every level: reliability, security, hardware support, standards compliance, api stability, scalability, efficiency, features, extensibility. Plus free. Not shedding tears for old school proprietary Unix.
Commodity hardware killed commercial Unix (Score:2)
It had little to do with Linux. It was commodity hardware from Intel that did in commercial Unix. Each major Unix variant was tied to specific, very expensive hardware. You had AIX running on IBM's processor, HPUX on HP's, IRIX on SGI, DECs Unix, Solaris on Sun, etc. All those workstation platforms were expensive. When Intel managed to destroy them on a price/performance level, customers moved in droves. Linux just helped provide an alternative to Windows to ease the transition. It could have easily
Re: (Score:2)
Where were the free BSD's in 1992?
Re: (Score:2)
Nope.... The commodity hardware was there as were the expensive *IXs for it.
The BSD crowd was presumed to have been exposed to actual AT&T UNIX code so they were slowed by the possibility of being shut down by Mother Bell and if not them then her bitch sister UC Berkeley. That made the BSD set way late out of the gate.
All the others had their genesis in original AT&T code/licenses.
Re: (Score:2)
That's kinda what I thought.
Re: (Score:2)
That and NT. AS big problem was that a lot of companies used a Unix variant that went out of business. It was nice to know that you coulld run on a platform that would not go out of business ( Intel) and an OS where the support would stay for a while. I think that as Linux got better that stuff went from NT to Linux. Escpecially when MS started playing pricing games and retiring old systems.
Fun reading everyone's comments (Score:2)
Linux didn't kill UNIX, Intel did..... (Score:5, Interesting)
I remember the day I did an Oracle import on one of the new fancy Intel boxes and went to lunch thinking it would take until the end of the day.
When I came back from lunch the import was done?
I thought, hmmmm there must be a problem with Oracle on RedHat, it couldn't have happened that fast.
But it did, the Intel single threaded performance was kicking the crap out of Sun's Ultrasparc 2 CPU for a fraction of the cost.
I notified Scott McNealy, the CEO of SUN MIcrosystems and he invited me to his executive team meeting via speakerphone to share my findings.
After that the pitchforks came out but it was beginning of game over for Sun who would soon be purchased by Oracle.
We would still be running Solaris, HP, IBM with PPC to this day if it were cost competitive, but the Intel chips seemed to come out of nowhere in the early 2000's and hastened the demise of Solaris, HP/UX, and AIX.
So Intel killed the radio star.
It wasn't that Linux was better, it was that Intel was much cheaper.
We actually took a step back in terms of tools, maturity, and features using Linux when compared to an O/S like Solaris, but cost for hardware and maintenance no longer made sense when you could get a truckload of Intel at a fraction of the cost.
Re: (Score:3)
We told him to quickly find better employment. He thought we were being asses instead of helpful to an otherwise great guy.
Re:Linux didn't kill UNIX, Intel did..... (Score:5, Interesting)
To add to that...
It wasn't just faster and cheaper. It was also much more convenient.
Linux would run on that decommissioned PC you had lying around, hardware you could get from the shop down the street, or, if you wanted fancy, something you could buy online from Dell that would be delivered in a few days.
In _2005_, if you wanted to get anything from Sun "the dot in dot com", you had to _ring a distributor_. If everything went smoothly you might get it the following month!
Another important thing was, well, Java.
On the server, suddenly you didn't have to worry about portable C, makefiles, various compilers, build machines, or any of that. You just copied the jar where it was needed. And it didn't matter one whit if that was on a Linux machine cobbled together from spare parts, or a $1M Sun or IBM. So why would you buy a "cheap" Suns to build and test? One for system testing, just in case. And that was it.
Maybe (Score:3)
I certaiy hope so! (Score:3)
Protection rackets over imaginary monopolies are bad, mmkay.
Richard Stallman and GNU killed UNIX (Score:5, Informative)
People do you not know your Geek history ?
It was Richard Stallman getting pissed off because he wasn't allowed across to proprietary code that led to the GNU Manifesto in 1985.
The GNU Manifesto was an idea to produce a non proprietary UNIX compatible OS system and software.
Richard than formed the GNU foundation (Gnu not Unix) which produced tools like Emacs, GCC and GNU MAKE it also promoted the idea of copyleft and
developed the GNU General Public License. The one thing they did not produce was a Kernel.
In 1991 Linus Torvalds came along with his Linux kernel and the rest is history.
GNU Manifesto: https://en.wikipedia.org/wiki/... [wikipedia.org]
Events leading to GNU: https://en.wikipedia.org/wiki/... [wikipedia.org]
Unix is home, single user linux is no longer unix (Score:3, Insightful)
Re: (Score:2)
And thousands of companies using SuSE/SLES or Debian instead of RHEL/CentOS are surprised at how many people seem to care about the Red Hat and CentOS news [zdnet.com]. I have to deal with it in my work because we support customers on a dozen different distros and versions, including ones past their EOL. (yeah, "free" software isn't a free lunch. And can be more expensive to staff than being a Windows-only shop)
The commercial Unix scene of from the 1980's to 2000's was worse to support. GNU autoconf/automake exists bec