Remembering Unix Desktops - and What We Can Learn From Them (theregister.com) 155
"As important as its historically underhanded business dealings were for its success, Microsoft didn't have to cheat to win," argues a new article in the Register.
"The Unix companies were doing a great job of killing themselves off." You see, while there were many attempts to create software development standards for Unix, they were too general to do much good — for example Portable Operating System Interface (POSIX) — or they became mired in the business consortium fights between the Open Systems Foundation and Unix International, which became known as the Unix wars.
While the Unix companies were busy ripping each other to shreds, Microsoft was smiling all the way to the bank. The core problem was that the Unix companies couldn't settle on software standards. Independent Software Vendors (ISV) had to write applications for each Unix platform. Each of these had only a minute desktop market share. It simply made no business sense for programmers to write one version of an application for SCO OpenDesktop (also known as OpenDeathtrap), another for NeXTStep, and still another one for SunOS. Does that sound familiar? That kind of thing is still a problem for the Linux desktop, and it's why I'm a big fan of Linux containerized desktop applications, such as Red Hat's Flatpak and Canonical's Snap.
By the time the two sides finally made peace by joining forces in The Open Group in 1996, it was too late. Unix was crowded out on the conventional desktop, and the workstation became pretty much a Sun Microsystems-only play.
Linux's GPL license created an "enforced" consortia that allowed it to take over, according to the article — and with Linus Torvalds as Linux's single leader, "it avoided the old Unix trap of in-fighting... I've been to many Linux Plumbers meetings. There, I've seen him and the top Linux kernel developers work with each other without any drama. Today's Linux is a group effort... The Linux distributors and developers have learned their Unix history lessons. They've realized that it takes more than open source; it takes open standards and consensus to make a successful desktop operating system.
And the article also points out that one of those early Unix desktops "is still alive, well, and running in about one in four desktops." That operating system, of course, is macOS X, the direct descendent of NeXT's NeXTSTEP. You could argue that macOS, based on the multi-threaded, multi-processing microkernel operating system Mach, BSD Unix, and the open source Darwin, is the most successful of all Unix operating systems.
"The Unix companies were doing a great job of killing themselves off." You see, while there were many attempts to create software development standards for Unix, they were too general to do much good — for example Portable Operating System Interface (POSIX) — or they became mired in the business consortium fights between the Open Systems Foundation and Unix International, which became known as the Unix wars.
While the Unix companies were busy ripping each other to shreds, Microsoft was smiling all the way to the bank. The core problem was that the Unix companies couldn't settle on software standards. Independent Software Vendors (ISV) had to write applications for each Unix platform. Each of these had only a minute desktop market share. It simply made no business sense for programmers to write one version of an application for SCO OpenDesktop (also known as OpenDeathtrap), another for NeXTStep, and still another one for SunOS. Does that sound familiar? That kind of thing is still a problem for the Linux desktop, and it's why I'm a big fan of Linux containerized desktop applications, such as Red Hat's Flatpak and Canonical's Snap.
By the time the two sides finally made peace by joining forces in The Open Group in 1996, it was too late. Unix was crowded out on the conventional desktop, and the workstation became pretty much a Sun Microsystems-only play.
Linux's GPL license created an "enforced" consortia that allowed it to take over, according to the article — and with Linus Torvalds as Linux's single leader, "it avoided the old Unix trap of in-fighting... I've been to many Linux Plumbers meetings. There, I've seen him and the top Linux kernel developers work with each other without any drama. Today's Linux is a group effort... The Linux distributors and developers have learned their Unix history lessons. They've realized that it takes more than open source; it takes open standards and consensus to make a successful desktop operating system.
And the article also points out that one of those early Unix desktops "is still alive, well, and running in about one in four desktops." That operating system, of course, is macOS X, the direct descendent of NeXT's NeXTSTEP. You could argue that macOS, based on the multi-threaded, multi-processing microkernel operating system Mach, BSD Unix, and the open source Darwin, is the most successful of all Unix operating systems.
Of course I remember! (Score:2)
Of course I remember! I am still using what could very well qualify as a "Unix" desktop nowadays. I can't see much differences from my ~1985 SunOS desktop as a matter of fact.
Re:Of course I remember! (Score:5, Interesting)
The very thing in TFA was pointed out nearly twenty years ago with the landmark Who needs an enemy when you can divide and conquer yourself? [usenix.org]
In fact, /. covered at at the time [slashdot.org].
Re: Of course I remember! (Score:2)
Same here, in fact last year I moved away from Mac after probably a decade and a half and went back to Linux. I speced out a machine specifically to have a modern Unix workstation-class feel, and I couldnâ(TM)t be happier.
Re: (Score:2)
Re: Of course I remember! (Score:2)
Iâ(TM)ll keep it simple, and only since you asked, itâ(TM)s a System 76 Thelios Mira. Normally wouldnâ(TM)t blatantly promote a product here. 14th gen i7, nvidia 4060, 64GB. Iâ(TM)m using their Pop!os, so far no reason to switch to anything else.
Re: (Score:2)
For the most part, that old school Sun with SunOS 4.1.4 versus a Linux desktop, so much has changed, but so much has stayed the same.
I do miss the "UNIX Workstation" vibe, because SunOS, IRIX, HP-UX, even AIX (which ironically did come on a laptop) were useful for a lot of things, and as a daily driver OS, had solid security and tools for local backups, provided you had a SCSI tape drive.
XFCE4 on FreeBSD (Score:5, Interesting)
My XFCE4-desktop is awesome, thank you very much. Last uptime was 386 days — and it only went down, because the video card's fan stopped working...
Firefox, Thunderbird, and Libreoffice have to be recompiled on occasion, but that's nothing compared to the forced biweekly reboots my corporate desktop is undergoing — running the OS, that is alleged to have "won"...
Re:XFCE4 on FreeBSD (Score:5, Interesting)
A similar argument to those running a mainstreak linux as well. far far far fewer distruptive reboots and wasted time. macos as well. Windows always seems to want to force a reboot and update when you can tollerate it lease and somehow they haven't fugured out how much users hate that. Sure, you can alter that in an enterprise environment or if your a nerd, but most people just suffer through it.
I'm quite happy with macos or a linux desktop myself, I just want it to work, I don't want to have to fight with it.
Re: XFCE4 on FreeBSD (Score:2)
You do know that as a user (or admin) you can defer updates to a time of your choosing, so they don't interrupt your workflow...
Updates that *require* a reboot is really only an issue if a user chooses not to control their system updates.
Re: (Score:2)
Bill Gate's moment of brilliance was when he realized that "good enough" was really good enough for most customers. Everyone outside of the hardcore Unix and mainframe world was intent on creating the perfect OS, the perfect office suite, the perfect GUI. Microsoft was able to get mostly-usable products out the door and priced appropriately for most businesses that would run on the vast majority of hardware available in configurations that people could figure out without a lot of training. Perhaps if Pet
Re: XFCE4 on FreeBSD (Score:2)
Re: (Score:2)
The 40 minute reboot time can be chosen freely. Have it do that at 3 AM, or tomorrow at 9 PM, you have full control over that. If you're 5 minutes out from a client meeting and the update kicks in at that moment, it's a user problem, sorry.
Changing directories delay: Yes, that's something that happens (rarely, but it does). It could have quite a few reasons, but let's not split hairs about it.
Home directory permission challenge? Not something I am familiar with. Not saying it doesn't exist, just never heard
Re: (Score:2)
Odd. My work Windows box needed a reinstall once in 7 years, because a third party program didn't uninstall correctly and started causing blue screens. I've never reinstalled my home couch laptop in the nine or more years that I've been using poor old thing. What in the world are you doing that you need to reinstall twice a year?
Re: (Score:2)
There was a time with Macs, in the System 7-9 range, where a cooperative-tasking OS was definitely not up to snuff. It was a standard operating procedure to restart when doing a work session, restart afterwards, and restart every few hours, because if one application misses their WaitNextEvent() call, the OS was hung, and you have to reach for the interrupt or reset button. Comp.sys.mac.advocacy discussed this to no end in the early 1990s, and how WorkPlaceOS or even A/UX was going to save the platform.
Re:XFCE4 on FreeBSD (Score:5, Insightful)
reboots have a huge range of annoyance. An occasional reboot for an update isn't so bad, but Windows on a laptop that isn't online for a day or two can force an update within minutes of bootup. When you opened the laptop because you needed it and then are stuck in a 10 minute update it's infuriating. macbooks or linux laptops don't do this to you.
Re: (Score:3)
Your XFCE "desktop environment" is awesome. You probably don't own a UNIX desktop (a workstation from a UNIX vendor) which seems to be the topic of the paper. Otherwise it does not make sense that they omit other famous desktop environments like Gnome and KDE. Not a single linux distro is UNIX certified, see here the current users of the UNIX brand https://www.opengroup.org/open... [opengroup.org]
Re: (Score:2)
+1 informative, I wasn't aware of Apple MacOS being certified.
Re: XFCE4 on FreeBSD (Score:3)
Re: (Score:2)
Your XFCE "desktop environment" is awesome. You probably don't own a UNIX desktop (a workstation from a UNIX vendor) which seems to be the topic of the paper.
I use XFCE as well. I can confirm that it's not a UNIX desktop because it doesn't have any harebrained ideas like "focus follows mouse" or those butt-ugly Motif styling themes.
Re: (Score:3)
Re: (Score:2)
here is how to turn on focus follows mouse in XFCE [xfce.org]
here is a Motif theme for XFCE
Re: (Score:2)
sigh. time to start using preview again.
https://www.xfce-look.org/p/12... [xfce-look.org]
If I used the mobile interface I wouldn't have to wait to post this followup. Whee!
Re:XFCE4 (Score:3)
I use XFCE as well. I can confirm that it's not a UNIX desktop because it doesn't have any harebrained ideas like "focus follows mouse" or those butt-ugly Motif styling themes.
I run xfce nowadays too. I turn off the harebrained "click to focus" that MacOS and Windows used to do to get the focus follows mouse that I had back in the olwm/twm/fvwm2/CDE days. When I was forced to run Windows , I found apps that could mostly turn on focus follows mouse. I never found one for Macintosh System 7. Its nice to have choices.
You also get your choice of butt-ugly styling themes instead of what Steve Jobs or the UI group at Microsoft decided (Windows 8, everything is a tablet?). FVWM was
Re: (Score:2)
K-UX and EulerOS used to be UNIX 03 certified, so Linux can be UNIX certified. K-UX certification lapsed in 2019, and EulerOS did not renew certification in 2022. But they both passed certification.
The main reason nobody bothers with that is that it's very costly, it's time limited, and it only applies to a specific build and set of components. And... nobody cares anymore, outside of a few companies which have regulations that software require specific certifications. But those are rare these days.
Heck, eve
Re: (Score:2)
Dennis Ritchie always consisted Linux to be as much of a unix derivative as bsd or any of the commercial unicies. I'll take his opinion on unix over the pay to play UNIX (TM).
Re: (Score:2)
Oracle makes servers. HPe makes workstations, so does Dell (Precisions), but those are intended for Windows or Linux.
Other than that, there are no real dedicated "UNIX workstations" these days. If I wanted a "Linux Workstation", I'd be more into lack of noise (Noctua fans, etc.) than power, and then have all the heavy lifting done by a server in another room.
Re: (Score:2)
My desktop PC is currently running 24/7 because I have the CCTV camera software installed on it. Currently in a (lengthier and costlier than I would have wanted) process of upgrading my networking and server stacks, which would probably take months and cost in excess of $15K.So, yeah, I am stuck with a 24/7 Windows machine at the moment :)
Re: (Score:2)
Instead of a Windows server for $15k, can't you run that camera off an Allwinner Nano or something for $15? These days camera drivers for Linux are as good if not better than for Windows, and it's not like there's a lack of management software. There's no way your time to figure out how would come to be worth $15k.
Re: (Score:3)
Let me clarify.
The $15K is a ballpark sum for an entire house's worth of physical infrastructure (Optic Fiber, CAT6A, switches and whatnot), all 10g capable, plus technical room racked hardware (servers, routers, switches, a bunch of RPis, UPS with extra battery pack), you know, the works. I am still debating whether to go SFP+ or QSFP28, although the latter is kind of prohibitive. SFP+ I already have in my tech room.
The main server will be Linux-based, an upgrade from my puny Ryzen 2400g machine. I plan to
Re: (Score:2)
My desktop PC is currently running 24/7 because I have the CCTV camera software installed on it.
Then it's not a desktop PC. Words have meanings, and desktop doesn't refer to where you physically put your computer's box.
Re: (Score:2)
Your post meets my expectations of you. Trollish and stupid.
A desktop computer (often abbreviated desktop[1]) is a personal computer designed for regular use at a stationary location on or near a desk (as opposed to a portable computer) due to its size and power requirements.
https://en.wikipedia.org/wiki/... [wikipedia.org]
I use it regularly, it sits near my desk, it's a personal computer.
Whether it runs 24/7 or not is irrelevant.
Re: (Score:2)
You are stupid and disingenuous.
I know.
Go away.
No.
The Common Desktop Environment (CDE) (Score:3)
CDE is a programmers GUI. I could add or change any click behavior and invoke a program (no big deal) or shell script (a big deal). I haven't seen a Desktop like it since I worked on Solaris 10. Do any modern GUIs have this programmability feature built in? Everything now I feel is very restricted in the behaviors allowed.
CDE made Unix workstations so much more productive for me than Gnome or Windows. You just needed some shell scripting skills.
Re:The Common Desktop Environment (CDE) (Score:5, Interesting)
Re: (Score:2)
KDE on Wayland can still do focus follows mouse, but it is quite buggy, sadly. I'm not sure if this is a problem kwin or the Wayland protocol that makes it hard to do it right. I'm hoping that when Mate Desktop finally supports Wayland, I'll find a good implementation of focus follows mouse.
Like you say it's very natural and really speeds up interactions with multiple windows.
Re:The Common Desktop Environment (CDE) (Score:4, Interesting)
Re: The Common Desktop Environment (CDE) (Score:2)
Never made much sense to me. It does sell more computer equipment, though.
Re: (Score:2)
I use and mostly prefer focus follows mouse, click to raise.
MacOS has one neat feature. You can Command-click on any control in a partially-covered window and interact with it without raising it. It's not as good as focus follows mouse, but it is useful. I definitely don't like that macOS always eats the click that raises the window, but I can see why they chose to do that.
Re: (Score:2)
Re: (Score:3)
Personally I found CDE to be incredibly ugly looking. Give me the the Indigo Magic Desktop any day. https://yohanan.org/steve/proj... [yohanan.org]
Re: (Score:2)
It also wasted screen real estate on ornamentation which, as you say, was intolerably ugly. I was running NeXT-look stuff at home on Linux while using CDE on Solaris at work, this wasted far less screen area while also being more intuitive.
Re: (Score:2)
There is a modern clone of sorts, NsCDE. [github.com]
Re: (Score:2)
Yes, you can add scripting to every modern GUI. Even in Windows you can add scripts to the right click context menus, it just takes registry entries (perhaps there is some third party tool to do this from the GUI.)
On a Linux system (and probably most any of the remaining commercial UNIXes as well) you make the right kind of file and drop it in the right directory. AFAICT the right kind of file is usually a .desktop file, and on a KDE 5 desktop the system ones are stored in /usr/share/kservices5, while the u
Re: (Score:2)
CDE is a programmers GUI. I could add or change any click behavior and invoke a program (no big deal) or shell script (a big deal). I haven't seen a Desktop like it since I worked on Solaris 10. Do any modern GUIs have this programmability feature built in? Everything now I feel is very restricted in the behaviors allowed.
CDE made Unix workstations so much more productive for me than Gnome or Windows. You just needed some shell scripting skills.
Most (xfce certainly) can have launchers that you can tie to a program/shell script. Can be in an application menu, on a tool bar. Or on the desktop.
Hardware (Score:5, Insightful)
Re:Hardware (Score:5, Interesting)
Back in the day, every computer platform was proprietary and vertical. It was hardware, OS and applications all part of one ecosystem, it did not matter that many workstations were based on Unix, they were all separate. What you bought was hardware, the software that came with it was not your choice to make.
The PC changed that, there was no "compatibility" industry for anything but the PC and even that was limited.
This article demonstrates no understanding of that history. Unix system vendors didn't somehow miraculously fail to provide a common platform, there never was one. Hell, even MS-DOS failed, if your computer didn't run true IBM PC-DOS customers didn't want it. Ask AT&T. Microsoft was playing the same game as everyone else, a single OS that ran on a single hardware platform. Eventually that changed with NT, but by then the game was over and even then efforts to broaden past Intel failed until recently.
Re: (Score:2)
Re: (Score:2)
Actually, it started with the Compaq Presario, the first Compaq that could run straight MS-DOS instead of needing a special Compaq specific version. That wasn't Compaq's idea, it was because Gates refused to provide any more hardware specific versions of DOS.
Exactly. MS-DOS was often very machine specific and you need to get program that ran on your version. Just as Linus was able to exert control over Linux, Gates was over MS-DOS; he just got there first.
Re: (Score:2)
There was a Tandy PC which had MS-DOS in ROM. That was nice because no matter what, you could always get access to your files and applications, and could always format a disk with something.
I sort of wish for a machine that had an OS, or even a recovery environment, like Windows PE. in ROM, perhaps one that could be upgraded via signed updates. That way, if you need something to recover from, you have that. Windows PE would be useful, especially if machines are in AutoPilot because you can just reimage a
Re: (Score:3)
Also because priests have discovered, very early on, that fear is a more powerful (and durable) motivator than love.
Re: (Score:3)
There is a strong element of fact in your comment, but the situation was a lot more complicated than what you are alluding to.
I'm not going to get out the Unix ancestry chart here, but while every vendor did have "their own" Unix, in practice and by design there were really by far mostly only System V and BSD. Yes, there were other Unixlikes which did exist and weren't based on these, but those existed mostly early on. As Unix became larger and people expected more interoperability between flavors, there we
Re:Hardware (Score:4, Interesting)
you could read Kernighan and Pike's "The Unix Programming Environment" and understand how to use and develop software for any of these platforms because underneath they were mostly the same code, and really did behave in mostly the same way. The UNIX kernel provided enough separation from the hardware to make it standard.
I went through lots of "The Unix Programming Environment" using DOS. People had ported/written unix tools for DOS that were similar enough to the Unix ones. I remember awk, sed, grep, lex, vi (elvis, stevie, calvin), emacs (freemacs, jove, miniemacs?), make, tar, gnuplot, shell. The GNU tools had the GNUish project, but didn't run on an 8088.
Microsoft was very focused on x86, and they even for a while owned Minix, a pretty good Unixlike for standard x86 PCs before they sold it to SCO.
Microsoft created Xenix and sold it to SCO. There was even an 8088 version.
Minix (https://en.wikipedia.org/wiki/Minix) was a very different beast created by Andrew Tanenbaum. It was a microkernel like Mach, created to teach systems programming to students. It cloned the Unix version 7 API and didn't need a Unix license (>$1000 at the time). Its unix tools were similar (or ported versions?) of the DOS tools I had used because their source was available and Minix only did the 64k I & D on 8088. Just like the DOS "small" model. The "proper" unix tools needed more RAM than that.
Anywho, TLDR Unix vendors didn't provide a common hardware platform, they provided a common software platform on which your code would compile and behave like you expected.
Yes. When I was a Unix sysadmin, we'd compile a whole suite of GNU and other tools for SunOS, Solaris, Irix, Ultrix, DEC OSF/1, HP-UX and AIX. There were lots of tiny API differences for each OS. Linux often just implemented more than one API to make porting easier. X11 was one example.
I don't miss chasing down dependencies back then. Linux really did packaging with Slakware, .deb and .rpm that the unixen didn't as much. Debian added dependency with apt and .rpm systems didn't really have it until yum came along.
Re: (Score:2)
Whoops, I meant to write Xenix there. ISTR reading someplace that they bought parts of it in, rather than developing it themselves, beyond licensing UNIX... but I don't remember the details.
Re: Hardware (Score:2)
The driver I saw was that CAD programs ran best on Unix workstations, and until PC Hardware got fast enough and robust enough, there was no other option. Then, suddenly, PC hardware matured, Windows NT came out, and the cost difference between a Sun, HP, or SGI workstation and and up-spec OC workstation was too big to ignore, capable PCs were a fraction of the cost of Unix workstations. To try and stay relevant some folks tried very hard to sell Unix for PCs (I'm looking at you Solaris for x86), they even s
Re: (Score:2)
I think HP was simply called HP-Unix or HP-UX. Maybe you were thinking about IBM AIX? Smitty was really cool, run, run, run, stumble and fail! Try again with changes, run, run, run, success!
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
HP servers used PA-RISC, then they switched to Itanium. What they use now, I have no idea.
Re: (Score:2)
Re: (Score:2)
I ran across an HP-UX system on a 68k system back in the day too.
you could argue... (Score:3, Insightful)
"You could argue that macOS, based on the multi-threaded, multi-processing microkernel operating system Mach, BSD Unix, and the open source Darwin, is the most successful of all Unix operating systems."
You could, but it would be a fanboy argument and meaningless. MacOS is successful because it's MacOS, not because of Unix, and it is only coincidentally Unix under the covers.
And why bring up Linux, its developers or its founder? Linux is a kernel, not a desktop, and Linux as a desktop hasn't standardized or arguably succeeded yet.
'...one of those early Unix desktops "is still alive"...'
no, it's not. NextStep was not "Unix" and Mach/BSD was not a desktop.
Re:you could argue... (Score:5, Insightful)
You could, but it would be a fanboy argument and meaningless. MacOS is successful because it's MacOS, not because of Unix, and it is only coincidentally Unix under the covers.
Strong disagree. While creative types have long favoured Mac, most hard-core developers and power users eschewed it for other platforms — until around 2003/2004, when OS X became mature enough and developers with UNIX-style toolchains moved over in droves.
Go back to relevant /. stories around that timeframe, and you’ll see how common it suddenly become to go to conferences and see 80%+ of developer laptops being PowerBooks (and later MacBook Pro’s). That was virtually unheard of just a year prior — and a lot of devs still prefer it to Windows, because most UNIX-style commands and toolchains “just work”.
Yaz
Re: (Score:2)
And yet they still have, what, 8% market share in the desktop world?
Re: (Score:2)
And that's also as much as they have ever had, and their share of the desktop has been as low as IIRC 1.2% before towards the end of the 68k era when it was obvious that a PC would stomp a Mac hard for a lot less money.
Re: (Score:3)
And yet they still have, what, 8% market share in the desktop world?
Which is far more than SunOS and all traditional Unix workstation vendors combined ever had. Apple brought unix to the consumer desktop.
Re: (Score:2)
and yet my most productive s/w development ever was NeXTstep - not just that it was Unix-y, but it had Interface Builder and Objective-C and Display Postscript.
Re:you could argue... (Score:5, Informative)
'...one of those early Unix desktops "is still alive"...'
no, it's not. NextStep was not "Unix" and Mach/BSD was not a desktop.
Actually, OpenStep was licensed UNIX. So, legally it was UNIX. But that's just legal pedantry. NS and OS were unix in all the ways that mattered. As is MacOS. They ran all the unix tools (some version or another). They did TCP/IP. Hell, they even ran X11 if you wanted to.
My father once asked me if linux was unix. Almost the same answer - legally it was not, but in every way that mattered it was. What's more, it was going to win (the unix wars).
In a sad way, MacOS and linux are the tied losers of the desktop world. I don't even know offhand what "legally branded UNIX" there is any more. Does Oracle sell some flavor of branded UNIX? You can argue that linux is more unix than MacOS - and that's fine - I'm not going to argue it. But I use MacOS because I like unix and I don't like windows and I do like having a nice desktop experience.
Re: (Score:2)
quote, "But I use MacOS because I like unix and I don't like windows and I do like having a nice desktop experience."
I used Mac OS X for many years because I hated Unix (almost as much as Windows) and the Mac did a good job of hiding Unix from me.
Now I use Mint Cinnamon, which likewise does a pretty good job of hiding Linux's Unix-like traits. And it's a desktop OS that still works like a desktop OS, not trying to mimic a phone or a tablet.
Re: you could argue... (Score:5, Interesting)
Re: (Score:2)
Linux runs on your phone, on the servers that run the internet, on supercomputers, .... and sometimes on the desktop, but since the desktop is mostly a browser now, the OS underneath is largely unimportant now
Re: (Score:2)
Well, a large proportion of the developers I worked with over the last 40 years preferred and still use the Mac. In part that's because when they need to, they can pop up terminal.app and do something with the familiar Unix toolset, such as setting up a quick tool stream using pipes.
So I'm sure it makes some people here happy to toss insults at Mac users as all computer-illiterate hipsters. And I'm sure there are a lot of those using Macs. But that's not the reason why I use it, or my developer friends u
Re: (Score:2)
So I'm sure it makes some people here happy to toss insults at Mac users as all computer-illiterate hipsters.
The iPhone fanbois give you a bad name.
Microsoft’s”cheating” wasn&rsquo (Score:5, Informative)
it was about beating OS/2.
Most of the other UNIX’s weren’t designed to run on 80x86 platforms, so they were never in any real contention with Microsoft. AIX, HPUX, Solaris, Dynix (which was Intel based but had a special architecture separate from PCs) — none of these were in the same market as Windows.
No, Microsoft’s target was OS/2 — which had a bigger resource footprint, but was also a vastly superior OS, with real pre-emptive multitasking, a (by 90’s standards) modern high performance file system, the ability to pre-emptively multitask Windows 3.x and DOS apps well before Windows 95, and a superiors desktop environment (a modern Workplace Shell would still absolutely slay). It was here that Microsoft introduced Win32s and kept changing it every few weeks to break OS/2 compatibility for newer Windows apps. It was here that the per-processor agreements were put into place with systems manufacturers to make selling OS/2 on systems more expensive (for those too young to know, in these agreements the manufacturer paid and charged for a Windows license with every system sold — even if it didn’t come with Windows. So if you wanted an OS/2 system you were paying for both OS/2 and a Windows license you didn’t actually get).
UNIX wasn’t even on Microsoft’s radar in the 90’s — it just wasn’t a PC operating system, and was mostly targeted to systems that didn’t compete with Microsoft. If you wanted a UNIX system, you had to buy your hardware from your OS vendor (much like with macOS today) — virtually nobody (except some of the early cool kids running Linux and *BSD) was buying white-box Intel systems and running UNIX — the numbers were too small for Microsoft to care. OS/2 was their real target — and in the end, it worked.
Yaz
Re:Microsoft’s”cheating” wasn&am (Score:2)
OS/2 was really good back when windows 3.11 was the norm. Much better than windows 95 ever turned out. I used to use it to run a dos based Microstation in one session, while running windows in another.
Also from this article, I have to say I really miss my Sun 10Ultra which used to be my desktop back in '98.
Nowadays just use apple, still avoiding MS.
Re:Microsoft’s”cheating” wasn&am (Score:3)
UNIX wasn't even on Microsoft's radar in the 90's
Did you forget Xenix? Microsoft sold the highest-volume Unix for a chunk of the 1980s, and it ran on Zilog Z8000 and Intel 8086 chips before later being ported to 80286 CPUs (among others). Microsoft marketed MS-DOS as its single-task operating system and Xenix for multi-user or multitasking uses. They eventually sold the rights to SCO, who ported out to 80386 CPUs.
Microsoft ditched Xenix to develop OS/2. They absolutely knew how to run Unix on a PC-like system.
Re:Microsoft’s”cheating” wasn&am (Score:2)
Nope - didn’t forget; as you pointed out MS gave up on Xenix in the late 80’s, and I specifically stated (as you quoted) “UNIX wasn’t even on Microsoft’s radar in the 90’s”.
MS did some stuff with UNIX in the 70s and 80s, but by the 90s they were all in on DOS/Windows and Windows NT, with a bit of Mac OS (before it was UNIX based).
Yaz
Re:Microsoft’s”cheating” wasn&am (Score:2)
Most of the other UNIXâ(TM)s werenâ(TM)t designed to run on 80x86 platforms, so they were never in any real contention with Microsoft. AIX, HPUX, Solaris, Dynix (which was Intel based but had a special architecture separate from PCs) â" none of these were in the same market as Windows.
They were not in the same market, no. But there were versions of AIX, HPUX, and Solaris which would run on actual PCs. There was nothing about them that made them not "designed" to run on a PC. In fact there are still enthusiasts who want to keep running Solaris on PCs; My first sysadmin job was at an almost-all-Sun shop and we briefly evaluated Solaris 2.5.1 on x86 but found it to have compatibility and reliability problems so I instead brought in Linux and we used that for X-Terminals. They could also run
Re:Microsoft’s”cheating” wasn&am (Score:2)
Linux is UNIX (Score:3)
And runs on a whole lot of personal computing devices - phones, Chromebooks (as native OS and Crostini), Windows (WSL), NAS boxes, routers... No need to remember when it's everywhere. UNIX started out as an operating system with widely available sources, if not open sources in modern sense. It's attempts to create closed source proprietary ecosystems that largely failed. If the complaint is that interesting new CPUs were abandoned in favor of Intel, there are lots of ARM options these days and RISC V is coming up. I am actually typing it on a mini Elementary OS desktop that I get to avoid doing personal stuff on a work laptop, but desktops are largely out in general, so UNIX desktops are too. Although at work we do most development on Linux workstations.
Re: (Score:3, Informative)
Re: (Score:2)
Linux is unix, according to Dennis Ritchie.
It is not UNIX (tm), but who cares? If it's enough unix for dmr, then it's enough unix for me.
He was also of the opinion that BSD is unix too.
Re: (Score:2)
Who cares, if it looks like Unix, runs Unix code, and works like Unix - it's Unix in all but a now irrelevant Trademark ...
Openstep and now Cocoa for Mac OS and iOS (Score:3)
Openstep was an open standard and reference implementation of NeXTstep.
Openstep was available from Sun, HP, Dec, and NeXT. Openstep also ran on Windows NT. There was/is also Gnustep. The one place Openstep didn't run was the PowerPC processor used in Macs. Sun even purchased Lighthouse Design to get Lighthouse Design's Openstep office application suite and said "All their wood is behind one arrow" meaning Openstep was the future of Sun. The founder and president of Lighthouse Design, Jonathan Schwartz worked is way up the Sun management to become CEO and sold Sun to Oracle.
- https://en.wikipedia.org/wiki/... [wikipedia.org]
- https://en.wikipedia.org/wiki/... [wikipedia.org]
- https://en.wikipedia.org/wiki/... [wikipedia.org]
Meanwhile, in December 1996, Apple purchases NeXT (or NeXT purchased Apple for negative $400M), and the long painful process of removing features and enshitifying the NeXTstep user interface commenced.
Now we have modern Mac OS with Cocoa and iOS with Cocoa Touch. Some features that Apple dropped from NeXTstep have still not reappeared, and many of the new feature are misguided in my opinion.
I started using NeXTstep/Openstep/Rhapsody/YellowBox/Cocoa/Cocoa Touch in Jan. 1989. The technology shipped commercially in Oct. 1988. Windows 3.1 did not ship until 1991, and Windows 3.1 was a shitty shell on top of DOS which wasn't even an operating system. It was a program loader. Much of the NeXTstep technology like Interface Builder remains in Apple's Xcode IDE to this day. The technology has evolved but still fundamentally works the same way to did in 1988. The fact that Cocoa Touch is STILL ahead of Android and far ahead of other desktop app technologies is testament to how far ahead of its time it was in 1988.
Re: (Score:2)
... The one place Openstep didn't run was the PowerPC processor used in Macs...
In public at that time. But later --
https://en.wikipedia.org/wiki/... [wikipedia.org]
"Rhapsody represented a new and exploratory strategy for Apple, more than an operating system, and runs on x86-based PCs and on Power Macintosh."
Re: (Score:2)
I grew up with a NeXT Cube in the house. I still haven't used a cleaner, saner interface. For 25MHz and 16MB, you got a lot more out of it than you'd expect. More than a lot of machines I use today, sometimes.
Where's the goddamn desktops? (Score:2)
Given that title, I expected a discussion about old desktop environments like CDE, IRIX, and so on.
Re: (Score:2)
This could be the year..
Dunno about y'all (Score:2)
but I've been on windowmaker over X11 since the late 90s and I'm as happy as I could be, I have scripts that I haven't touched since then. I learned that I don't need the latest-greatest, just the one that works for me.
Besides, it is always funny to watch the people who try to sit in front of my terminal and do something.
Does it qualify as a proper "unix desktop" or not or am I in violation of specs purity?
Cost? (Score:2)
The issue was that a Unix workstation was too expensive for the mainstream. I had an SGI Indy .. but that was $5000 ($10k today money?) in 1993 and it had a really shitty OS called Irix that was super-hackable. What I mean by hackable is not that you could have fun modding it but that it had security vulnerabilities galore. I'm 100% sure there's hacked videos of me eating Nachos and drinking cans of RC Cola as I sat in front of its webcam (oh yeah, the Indy was the first computer to have a webcam).
Re: Cost? (Score:2)
Yes. Once the 386 arrived, the cost thing was why Linux took off. Anyone could run Linux. It was very expensive to get a unix or xenix license with the hardware to run it.
It wasn't long before the unix computers were undesirable, because they didn't run Linux. Also, the workstations were only 2 to 3 years faster than pc hardware. So you needed an application that required the fastest hardware to justify the 4 to 10 times cost increase, and the ongoing budget to stay in the arms race. Otherwise after a coup
Re: (Score:2)
The issue was that a Unix workstation was too expensive for the mainstream.
Absolutely this.
Back when I was responsible for maintaining several apps/systems distributed across various OSs (HP-UX, SunOS, AIX and others) I became the go-to person in the company for solving Unix system problems. Even though I was an engineer, not an IT/CS person. Finally, one admin asked me (on our internal Usenet system) how I had acquired all my training and experience on Unixes. The admin staff didn't have (and were prohibited from "fiddling around" with $10K or $20K workstations lest they break t
Gnome 3+ did a great job at killing Linux Desktop (Score:2)
Gnome 2 was so close to perfect. I never understood how the Gnome 3+ disaster was possible in the first place. Gnome is now at version 45 and has not resolved a single one of the gigantic usability problems it introduced out of nowhere.
There's a reason why alternatives like the excellent XFCE are so successful today. I think XFCE just needs a better default configuration and a fun tutorial that better shows off its exceptional capabilities.
Re: (Score:2)
Canonical Ubuntu was using Unity Desktop about at the same time Gnome radically changed for 3+ version. I don't know if Canonical played a role in that change.
BSD kickstarter (Score:2)
Thanks to BSD, open source and widely available source and manuals on its Unix many got a solid grounding in file-based OS’n.
Today that fundamental step through is obscured and fragmented by subsequent iterative compartmentalizations and abstractions. Gone are the simple manuals, flags and instead API’s, toolsets and numerous language interfaces.
Columbia taught that market forces, wage pressures and guarantee of getting a job on WallSt were fundamental.
NeXTSTEP taught that less than 3% could man
Why "remembering"? (Score:2)
I use my nice, efficient and well-working FVWM desktop every day. Made a really good config for my needs about 30 years ago (yes, almost first release). Optimized it a bit more when FVWM2 came out, spending half a day. Since then I have what I want in desktop. Compare that to the MS crap where nothing works as expected, you have to adjust to something new and more crappy every few years and they seem to be totally lacking in vision and UI design knowledge and skill. I mean, 30 years after FWVM had its excep
I Installed Novell Unix on a PC : ) (Score:2)
Re: (Score:2)
Everything as a text stream was (and still is) a terrible idea, as it makes everything dependent upon the multitude of ever-changing human languages,
False. That's only true if both the programmer who created the output and the programmer who parsed it are stupid.
Look at the psuedo-files in proc or sys sometime. Many of them don't even have headers. You're expected to know what order the fields come in. Even if they do have headers, you don't need those headers to parse the files programmatically. They are there solely for the benefit of anyone looking at the files without a tool. Meanwhile, they enable that activity.
When you make updates to such a psued
Re: (Score:2)
This is just garbled.
Luke audio, well better than nothing I guess, but I do like having to kill it every so often when it shits itself.
Dbus is an RPC system, it doesn't replace RPC.
Wayland is currently a mess. X's design mantra was "mechanism not policy". Waylands appears to be "neither mechanism nor policy then shout at the users and tell them how they shouldn't want to do whatever it is that doesn't work properly under Wayland, then maybe reimplement it in a semi working, fragmentory sense which only sort
Re: (Score:2)
The commercial Unix guys thought msdos and windows was just toys.
And they were and are right about that. The MS crap is still toy-level and never exceeded that. Insecure, hard to use, inconsistent, cumbersome, unreliable, intransparent, really bad customization, spying on you, etc., etc., ad nauseam.
Unfortunately, the business morons cannot think in terms of quality for shit and hence went with the cheaper option because "it is the same, right?". And they never even looked at Linux or the free xBDS back when because they did not get the FOSS idea and did not try to find
Re: (Score:2)
Look at the cost of compilers. Want a compiler for IRIX? That would be $2000. Compiler for AIX, not cheap. Even header files on a lot of UNIX variants were a chargeable option.
gcc changed all that. Things changed to where the main use for running a vendor's compiler was to make a stage 1 gcc binary, then people used stage 2, or even a stage 3 binary for all their code compilations, as gcc worked well across architectures.
On the other hand, getting a compiler for MS-DOS, something like Turbo Pascal was