I miss OS/2 a lot. For what it was at the time (intel, not ppc) it worked really well. When I was at Netscape, my build machine was OS/2 so I could do windows builds and still actually work. Machines then were much less capable than now, but I rarely had any bogging down of the system.
I absolutely loved OS/2. It was an absolutely phenomenal operating system. I really wish that IBM had put more effort behind it. Today, if MS and IBM could actually cooperate with one another, it'd be great to get it open sourced.
Sadly, by the time OS/2 was really competitive, MS had taken the market, and there was little reason for most users to go buy another operating system when Win3 or Win95 came on their machines, and NT was shipping on workstations.
I wish I could find it but there was a article written by someone who I think had some connection to sales about how OS/2 was effectively "sold wrong" by IBM who just didn't seem to know how to sell software like OS/2.
I remember posting that article on a company intranet once and a bunch of former IBM sales folks (who now worked at the same company I did) chimed in to echo the idea that IBM had a neat product with OS/2 but as an organization had no idea what to do with it.
Not to say it would have overtaken Windows, but it also struggled because it was sold by a company who didn't know what to do with it.
No marketing campaign or OEM sales strategy would have saved OS/2.
All the things that made it better than Windows 3.0/3.1 also meant it was totally unsuited to fill the disruptive low-end of the PC market that ate Windows up.
The resulting Microsoft platform dominance led to ecosystem dominance led to software library dominance and that was the ballgame.
I remember in the warp days they were throwing free cdroms at anyone who asked, full licensed. I got several at trade shows. They were trying to sell it for sure.
"IBM who just didn't seem to know how to sell software like OS/2"
Clearly, IBM was never that great at marketing much of anything IMO. However, many at the time also believed that in addition to not being that good at marketing, IBM (collectively) wasn't really overly interested in marketing OS/2.
Yeah it would have been amazing to not have MS as the dominant desktop OS vendor in the world.
But Warp in particular was just sooooo heavy. I used it in the day, I even got a free cdrom somewhere. But it was just terribly sluggish. And I was a computer science student so I already had more resources than most. They really screwed that up so bad.
That's not my recollection of OS/2 3.0... It was snappy on contemporary hardware and 16MB memory.
But OS/2 was always sensitive to available RAM, and IBM liked to understate its memory requirements. (They pretended that OS/2 2.0 could run on 4MB because they had promised it years earlier. But it was really unusable on only four megs.) Maybe that was the issue?
I did my first internship at Boca Raton in the OS/2 device driver support group. They announced OS/2 PPC while I was there, and also BeOS was dropped around the same time. Suffice it to say it was an exciting time for PPC hardware that I could never afford on my own (Windows 95 also came out that year, it was all so nuts).
In the OS/2 2.x days my primary means of interacting with other OS/2 folks was the Canopus (Wil Zachman) forum on Compuserve (75060,264 checking in). Weird that I still remember my Comuserve user id!
At that time, there was a somewhat common saying "Annoy IBM, support OS/2!". This was essentially mocking the ineptness (disdain?) of IBM marketing for OS/2.
(75060,264 checking in). Weird that I still remember my Comuserve user id!
Over 40 years later, and I still remember mine, too. I have no idea why. Something about those octets that sticks in your brain like phone numbers used to.
I remember at the time there was also going to be the wonderful new kernel that would allow OS/2 and MacOS to coexist on the same machine. As someone who had a Mac and an OS/2 machine side-by-side on his desk, this seemed like it could be a wonderful thing, but alas, it was never to come to be.
I was just a kid during the 1990s when all of this was happening, but a few years ago I remember reading about an IBM project named GUTS where one kernel would run multiple OS "personalities":
Microsoft technically delivered something very close to OS/2’s “Personalities” in Windows NT 4. They called it "Environment subsystems". Each subsystem could run applications written for different operating systems, the 3 available ones were Win32, OS/2 and POSIX. Then there was the "Integral subsystem", which operated system-specific functions on behalf of environment subsystems.
But every subsystem other than Win32 was kneecapped mostly due to politics and market positioning.
In late 90s Microsoft bought a company which had developed a more enhanced Unix subsystem and rebranded it as Interix and marketed as Windows Subsytem for Unix (SFU).
I believe the original WSL was a resurrection of SFU before WSL2 pivoted to a VM-based approach.
No, the original WSL was a weird new thing where NT kernel-level driver actually serviced Linux system calls.
IIRC, Interix still used same approach as original posix subsystem (and Windows and OS/2 subsystems) of providing the interface as DLL that ultimately your application would be linked against.
To the nostalgics among us: what made OS/2 special? 32bit support?
I recall trying OS/2 2.0 or 2.1 back in the day, coming from a DOS/Win3.11 setup. It seemed to have the same basic features as DOS/Windows but wasn't properly compatible with my existing software. Admittedly, this was before I knew anything about programming. I discovered Linux not much later. It wasn't compatible with anything either, but seemed like a totally different and much more compelling proposition.
> To the nostalgics among us: what made OS/2 special?
I started out with OS/2 v1.1. It had threads, DLLs, multi-tasking, much larger memory space, and from v1.2 a somewhat decent filesystem. Coming from DOS 3.2/Win 2.0 this was an incredible leap, in particular the SDK was amazing compared to the ragtag assembly of info I was used to. The _delta_ between two systems haven't been this large ever since, and I think that is what contributes to the "magic" feeling.
The OS/2 WARP Presentation Manager was a better "desktop" paradigm than Win3.11. It supported more customization and stranger "objects" you could store on your desktop. It felt a bit more coherent and a lot more powerful than the Win3.11 Program Manager.
I was mostly a kid with a huge stack of PC games I'd play, and OS/2 was a better launcher for many (but not all) of them than DOS/Windows. I was "dual booting" OS/2 WARP and DOS/Windows, but because of my gaming habits it was more like quintuple booting because I had a long boot menu with I want to say 4 to 7 different combinations of AUTOEXEC.BAT/CONFIG.SYS settings depending on type of game I intended to play or if I was going to use a Windows app or something else, then one OS/2 WARP boot option.
A bunch of Windows apps (many of which ran better, even) and even some games I recall I started launching from OS/2 WARP instead of DOS/Windows, making the first boot choice of the day a lot easier. (Though I don't remember being able to delete most of the other combinations, still had to reboot for certain games and Windows apps that needed more RAM than what OS/2 left for applications. OS/2's biggest problem at the time was a huge RAM footprint compared to Win3.11, much less DOS micro-tuned with AUTOEXEC.BAT/CONFIG.SYS low footprints for specific driver combos.)
Multitasking, SOM (contrary to COM, it does implementation inheratance across languages, multiple inheritance and has meta-classes), object based desktop, Smalltalk for business application development (basically a similar role as VB and .NET have gotten latter on on Windows), Visual Age for C++ had a Smalltalk like experience (although ported to Windows as well).
However this also meant a more beefy hardware than the DOS/Windows 3.x combo.
Back in 1995 it was, to my knowledge, the only OS capable of sharing CD-ROM's on the network. Even MS-DOS and Windows 3.11 machines could access it.
It was also capable of sharing Mainframe printers using a real null-printer-driver, which was not possible on Windows NT3.51 or NT4.0. Windows always messed with the Mainframe codes that it could not understand.
It was also easy to set up OS/2 as a gateway between different network hardware and protocols (Token Ring to Ethernet, or NetBios to IPX/SPX, ...)
Probably a lot of things. Often software is simplified, at the time because of limited hardware and probably other software. Nowadays it's often a deliberate product decision but it seemed for OS/2 no such limits existed. E.g. you could right-click on a program, get the properties, run multiple applications. It even had a Windows emulation so stable that it was never matched by WINE. Of course there was only 16 bit Windows support but still...
Of course it had limitations of its own, I don't think you could any DOS/4GW games. Linux Installation seems simple compared to installing OS/2. I had to go through some sort of pre-installation guide which was printed on a separate paper and not part of the official manual. Also dual boot was meant literally: you booted into OS/2 and then you could "exit" into Windows. Back in DOS/Windows there was a command to do this the other way around. One time I didn't do this for half a year and was really anxious if my setup would make it...
At the time, formatting a floppy disk was a single task thing.
Downloading a file via Zmodem was mostly a single task thing.
The Windows of the day could to the latter, but not the former.
OS/2 on my 8mb 386sx could do both AND have a clock up, and play solitaire, and have another terminal window open.
It took a bit to get there, and there was swapping while everything loaded, but it was true pre-emptive multitasking, while still maintaining the highly time critical I/O stuff that Windows couldn't touch.
OS/2 performed better than windows generally and was more stable.
Back in that time period tech specs, and tech details really dominated a lot of "computers" discussion. I feel like that has kinda changed as far as the larger world goes (even if on HN tech specs are still relevant). Does an every day user want to use it? was less of a question for enthusiasts.
People today don't realize how much time was spent formatting floppy disks, and how slow the process could be. So slow that eventually companies started selling pre-formatted disks and charging extra for them.
OS/2 could give you back hours of productivity each month simply because you could do something else while formatting a disk.
The problem was Windows handling of interrupts. Until, i think, Win2k, generating a lot of interrupts (floppy, parallel port, high network activity) will slow Windows to a crawl.
Computers were far more crashy in those days, but OS/2 crashed far less often than Windows or even DOS did. And sometimes when a program crashed on OS/2, it only killed itself; it didn't take down the whole machine, so you had a chance to save your work in other programs before rebooting.
It also either was, or felt like it was, very very fast. Windows felt like a laggy VNC connection. GEM and the rest weren't much better speed-wise than GEOS on a Commodore 64.
Yes, stability! It was very nice when doing DOS or Windows 3.x development.
The only stability gotcha was when some OS/2 PM application hung the input queue and then the whole of PM became unresponsive. The base OS continued running fine but PM was then unusable.
Wow, PPC? I once had a Norstar NAM for voicemail, and did connect a monitor and keyboard and witness OS/2 on that. I think the NAM was an x86-based system. But I know the Meridian PBXs themselves were completely distinct from Norstar that I used. Never knew Nortel used PPC. Very neat!
Though my time with them came only as non-IP phone systems started to be considered obsolete, I am still a huge fan of the rock-solid stability and realtime speed of those digital systems. Not to mention their lack of a need for subscription services to operate.
Probably some but those old towers were not VoIP capable and the handsets are no longer sold. I doubt there's many around.
The voicemail card was basically an 8088 and i think still used floppies IIRC. I didn't work with them directly but I used symposium heavily as an admin (symposium was the callcenter solution)
After returning a 'Reader Service Card' I found in Info World magazine, I started getting catalogs from Indelible Blue. They were _the_ big reseller of OS/2 software. Lots of good stuff in there, plus they had articles evangelizing OS/2 and ReXX scripting.
I had a brush with the PS/2 and OS/2 back in the day. IBM offered the PS/2-based Personal/370 on the platform with a card that could execute S/370 SW. My team built the I/O channel card for connecting 3480 tapes and the like. It was funny seeing the fat bus and tag cables attached to a little box through an adapter pigtail.
I’m always curious how these projects come about and survive: why go to all of the effort to port for a dead-end product line? As technically sweet as it is? I imagine they would’ve found a decent market if they’d ported to Power Mac.
(Also, was the x86 emulation implemented in-house? I wouldn’t be surprised if some niche small company had a x86 emulator for PPC product that they could be paid to port.)
The plan was for all operating systems on top of IBM's POWER/PPC hardware to be rehosted as "personalities" on top of the Workplace OS microkernel, but in the end, OS/2 was the only personality that saw any real work.
The Workplace OS would also have been used on Apple hardware as part of the abortive Taligent project.
(It also would have been used on x86 and other platforms, but they started with PPC)
In reality, in some ways we are there now. I'm wondering if we can say that the "workplace OS" can simply viewed as a hypervisor and the "personalities" that run on top it are simply VMs (perhaps being paravirtualized).
I think oddities like this were a consequence of a hardware world that was rocketing along the heart of Moore’s Law, alongside a software world that hadn’t matured past multi-year product cycles.
When OS/2 for PowerPC was set in motion, that Intel would “Make CISC
Great Again” with the Pentium was far from clear.
I remember that the "general consensus" was that RISC was gonna win, it was just a matter of when (and when it could be affordable). What was NOT certain was which RISC architecture would come out ahead, so there was a bunch of porting to "remove the risk" - later they would unport most everything and "remove the RISC".
Pentium shook that tree a bit, and Pentium II really razzle-dazzled it.
Well, the thing is that RISC did win. It is just that the RISC that won is the one that Intel baked into their x86 chips.
The Pentium introduced the idea of micro op codes though the Pentium Pro was the first chip to really run with it. The CISC x86 instructions were converted into simpler instructions internally. These micro op codes could be pipe-lined, executed in parallel, and executed out-of-order.
If the Pentium II really razzle-dazzled, it did it with RISC architecture at its core. The CISC instruction decoder added a bit of die size but that did not matter much and Intel had leading-edge manufacturing tech.
The internal parallelism was also put to good use by adding SIMD instructions (MMX). These first appeared in the Pentium MMX and Pentium II but the Pentium III did it much better and of course Intel has continued to add more powerful SIMD stuff over time.
RISC did not win only inside Intel chips of course. Every successful ISA since the 90's has been RISC including ARM and RISC-V. But even RISC chips feature some complex instructions these days.
I'd argue (to some extent) that the proliferation of SIMD instructions demonstrates that RISC did lose, not just the practical war, but also the conceptual one. i.e. we creates many many similiar instructions today, which seems to go against the ethos of RISC.
I'm not sure I agree with "dead end" outside of the benefit of hindsight, or maybe don't get the point you're making. Neither the PowerPC nor OS/2 were dead-end in 1995, and competition in the OS space was still happening. Why wouldn't IBM want to have PowerPC survive, let alone thrive, with OS options? And surely they'd have loved something to take on Microsoft at this point in history.
One of the original design requirements for NT was that it be portable between different CPU architectures, it was one of the driving forces behind its creation.
So much so in fact, Microsoft developed NT 3.1 first on non-x86 architectures (i860 and MIPS), then later ported to x86, to ensure no x86 specific code made it in.
"Windows NT 3.1 was released for Intel x86 PC compatible and PC-98 platforms, and for DEC Alpha and ARC-compliant MIPS platforms. Windows NT 3.51 added support for the PowerPC processor in 1995"...
NT is a pretty interesting bit of PC history, I can highly recommend the book "Show Stopper!" by G. Pascal Zachary that recounts its development, and also dives a bit into why making the OS portable across CPU architectures was so important to the team at the time.
Something I didn't realize until recently was that the original MIPS version of Windows NT was Big Endian. I'd always heard it said that WinNT was strictly, 100%, absolutely always little endian, and the fact that every CPU that got a port (or was going to get a port) was either little or bi endian confirmed this.
Well, it is true, but Windows did run BE on the original MIPS R3000 platform. And only on the R3K[0]. The CPU architecture flag is still defined on modern Windows as IMAGE_FILE_MACHINE_R3000BE. There's an early test build of Win3.1 + GDI somewhere that runs on this platform.
The actual first release of WinNT 3.1 only supported MIPS R4000 and higher, I think. In little endian mode.
[0] I know the Xbox used a modified NT kernel, I've seen claims that the Xbox 360 also was, which would make it the second NT system to run big endian. Not familiar enough with sources better than wikipedia to trust that it actually was.
I believe you're correct - Xbox runs a modified Windows 20000, and subsequent versions I'm not certain on but I know it basically stands up a modified Hyper-V and the parent partition is the interface OS with games booting whatever kernel they were built for inside a VM
One of my first job out of school was as a sales support for the then bleeding edge NT 3.1 MIPS box for a company in Canada. Fond memories of loading stacks of 1.44 floppy disks for NT 3.1 and mangling ARC paths (Advanced RISC Computing, boot firmware). This was pre-internet and documentation was often hard to come by, incomplete etc.
I remember demoing the machines to astonished clients by running a stupid number of Clock apps on the desktop without a hitch.
My first real job out of school was supporting Windows NT on Dec Alpha for a company in Canada.
Things were so weird and wonderful back then. You could get GCC from Microsoft for Windows NT 3.1 for Alpha (crazy). And when Windows NT 4.0 came out there was the FX32 subsystem that ran X86 apps on Alpha (very similar to Apple Rosetta but much earlier).
I did not realize Canada was such a hotbed of Windows NT RISC.
Interesting historical note: the main reason PuTTy exists is because its author was given a Windows NT on alpha workstation and there was no native terminal emulator for it that he needed to connect to other equipment. IIRC, PuTTy still supported alpha into the 2000s until the build machine he had failed.
Yes and no. It was a permanent delete, no visible GUI with deleted files. But there was an actual UNDELETE command that could be used to recover files (or not)
I remember working with a guy who was the internal IBM rep (we were in corporate tech support in the late 1990’s and we were a BIG IBM reseller, so my friend got lots of bennies at work) and he had a PPC 605 (not the 615 IIRC), but he ran AIX on it (we were in the AIX and Unix support team).
> What could have been. If the respective parties... on the PPC 615, OS/2
There was never a chance at that time because x86 chips were produced in such volumes that PowerPC chips couldn't compete price-wise. Also, OS/2 became an instant outsider once Windows 95 was released. Two underdogs don't make a winner. The article says it all:
"The OS was clearly unfinished and not entirely stable. Worst of all, there were about zero applications. Because OS/2 PPC was never truly in use, PowerPC versions of OS/2 applications were never sold."
One of the big reasons people were upgrading to OS/2 was people that wanted a stabler/cleaner Windows 3.1. Most of the hottest apps for OS/2 were Windows apps. IBM started on the back foot in that competition. Windows 95 clipped their heels because it had the stability/cleanliness for cheaper and less of a RAM footprint.
Additionally, while this is US-centric, there were still many households in the mid-1990s whose first computers were PCs running Windows 95, just in time for the World Wide Web to be widely available, which created demand for personal computers. Additionally, this was during the time when Apple was struggling; its Performa lineup geared toward home users was not in the best of shape in 1995 (https://en.wikipedia.org/wiki/Power_Macintosh_5200_LC). By the time Steve Jobs returned and Apple released the first iMac (1998), it was just about time for Windows 98.
Being at the right age when Windows 95 came out, I didn't really know that there was a "Windows" prior to 95. My dad's computer ran DOS and used something called Powermenu as an organiser for executing programs. I think I had to run Wolfenstein in a tiny window for it to be fast enough to be playable, and may have, at one point, deleted one of the required DOS system files in order to try to tweak the life out of it to try to get it playable full screen. I think that was a 286. More years ago than I care to admit.
I was paying very close attention to this world at the time, and I am confident that 90%+ of every PC that could run Win95 at the time of its release had been sold by an OEM as booting straight into the bundled copy of Win3.1.
Yes, lots of people had older PCs and got Win95 with their hardware upgrade cycle.
They were never in any danger of upgrading to OS/2 or anything else: Microsoft owned the OEM sales channel and the software ecosystem for the PC for a solid 5+ years before Windows 95.
What numbers? As much as I hated Windows 3.x (which is why I upgraded from DOS to Linux, not to Windows 95, and never looked back) it did not occur to me that many by 1995 did not have some Windows 3.x installed, as it was required for so much software (even some games).
Still don't support the claim that people were mainly updating from DOS without Windows 3.x. Anecdotally I still think almost everyone using DOS by 1995 had Windows 3.x installed as well. Not necessarily a copy of Windows that was the result of Microsoft selling a copy of course.
Apple somehow managed to claw it's way to releavance from a weaker position in 1998 (with PoserPC!) So if they had their act together they could have done better in the early 90s.
hey squandered their early lead in the US among consumers and education and also ignored the international market.
Not gonna lie Wintel was a formidable force. Microsoft was ruthless in cornering the market.
But technically, OS/2 and MacOS gave Windows a run for it's money, arguably superior on some respects, and you could say the same for PowerPC and Intel.
‘SteveB went on the road to see the top weeklies, industry analysts and business press this week to give our systems strategy. The meetings included demos of Windows 3.1 (pen and multimedia included), Windows NT, OS/2 2.0 including a performance comparison to Windows and a “bad app” that corrupted other applications and crashed the system. It was a very valuable trip and needs to be repeated by other MS executives throughout the next month so we hit all the publications and analysts.’
‘The demos of OS/2 were excellent. Crashing the system had the intended effect – to FUD OS/2 2.0. People paid attention to this demo and were often surprised to our favor. Steve positioned it as -- OS/2 is not "bad" but that from a performance and "robustness" standpoint, it is NOT better than Windows’
"I have written a PM app that hangs the system (sometimes quite graphically)."
I don’t see how SteveB’s “bad app” was out of bounds. IBM was telling these analysts a bad app couldn’t bring down OS/2. As someone who used (and liked!) OS/2 at the time, this was indeed false: IIRC OS/2 in this era had a single-threaded UI loop, and a misbehaving app could (and occasionally did) block input for all apps and corrupt the UI.
You could argue about the difference between this and a Windows “UAE”, but to the user the experience was a broken, unusable OS/2.
I miss OS/2 a lot. For what it was at the time (intel, not ppc) it worked really well. When I was at Netscape, my build machine was OS/2 so I could do windows builds and still actually work. Machines then were much less capable than now, but I rarely had any bogging down of the system.
I absolutely loved OS/2. It was an absolutely phenomenal operating system. I really wish that IBM had put more effort behind it. Today, if MS and IBM could actually cooperate with one another, it'd be great to get it open sourced.
Sadly, by the time OS/2 was really competitive, MS had taken the market, and there was little reason for most users to go buy another operating system when Win3 or Win95 came on their machines, and NT was shipping on workstations.
I wish I could find it but there was a article written by someone who I think had some connection to sales about how OS/2 was effectively "sold wrong" by IBM who just didn't seem to know how to sell software like OS/2.
I remember posting that article on a company intranet once and a bunch of former IBM sales folks (who now worked at the same company I did) chimed in to echo the idea that IBM had a neat product with OS/2 but as an organization had no idea what to do with it.
Not to say it would have overtaken Windows, but it also struggled because it was sold by a company who didn't know what to do with it.
No marketing campaign or OEM sales strategy would have saved OS/2.
All the things that made it better than Windows 3.0/3.1 also meant it was totally unsuited to fill the disruptive low-end of the PC market that ate Windows up.
The resulting Microsoft platform dominance led to ecosystem dominance led to software library dominance and that was the ballgame.
I remember in the warp days they were throwing free cdroms at anyone who asked, full licensed. I got several at trade shows. They were trying to sell it for sure.
"IBM who just didn't seem to know how to sell software like OS/2"
Clearly, IBM was never that great at marketing much of anything IMO. However, many at the time also believed that in addition to not being that good at marketing, IBM (collectively) wasn't really overly interested in marketing OS/2.
My understanding is that by Windows 95, IBM's PC division preferred to sell computers with Windows to OS/2.
Yeah it would have been amazing to not have MS as the dominant desktop OS vendor in the world.
But Warp in particular was just sooooo heavy. I used it in the day, I even got a free cdrom somewhere. But it was just terribly sluggish. And I was a computer science student so I already had more resources than most. They really screwed that up so bad.
They should have called it OS/2 Wait
That's not my recollection of OS/2 3.0... It was snappy on contemporary hardware and 16MB memory.
But OS/2 was always sensitive to available RAM, and IBM liked to understate its memory requirements. (They pretended that OS/2 2.0 could run on 4MB because they had promised it years earlier. But it was really unusable on only four megs.) Maybe that was the issue?
I did my first internship at Boca Raton in the OS/2 device driver support group. They announced OS/2 PPC while I was there, and also BeOS was dropped around the same time. Suffice it to say it was an exciting time for PPC hardware that I could never afford on my own (Windows 95 also came out that year, it was all so nuts).
In the OS/2 2.x days my primary means of interacting with other OS/2 folks was the Canopus (Wil Zachman) forum on Compuserve (75060,264 checking in). Weird that I still remember my Comuserve user id!
At that time, there was a somewhat common saying "Annoy IBM, support OS/2!". This was essentially mocking the ineptness (disdain?) of IBM marketing for OS/2.
(75060,264 checking in). Weird that I still remember my Comuserve user id!
Over 40 years later, and I still remember mine, too. I have no idea why. Something about those octets that sticks in your brain like phone numbers used to.
I remember at the time there was also going to be the wonderful new kernel that would allow OS/2 and MacOS to coexist on the same machine. As someone who had a Mac and an OS/2 machine side-by-side on his desk, this seemed like it could be a wonderful thing, but alas, it was never to come to be.
I was just a kid during the 1990s when all of this was happening, but a few years ago I remember reading about an IBM project named GUTS where one kernel would run multiple OS "personalities":
https://en.wikipedia.org/wiki/Workplace_OS
The 1990s were quite a time for personal and workstation computing.
This was the same design goal that Windows NT had. In fact, it launched with Win32 (Windows), OS/2, and POSIX (UNIX).
I think the OS/2 subsystem was 16-bit OS/2 1.x so nobody cared and the POSIX subsystem was just compliant enough to win government contracts.
This design is why we have the "Windows Subsystem for Linux" (a name everybody hates) because "Windows Subsystems" were already a thing in Windows.
Docker, Distrobox, and even Flatpak are one kernel with multiple "personalities" but they are all still Linux I guess.
You can also argue have this on our desktops today with things like KVM in Linux and Hyper-V in Windows.
Back then if POSIX support was actually serious I would never have bothered to play with Linux, by buying the first edition of Linux Unleashed book.
I also bet that many others would not have cared, and used UNIX/NT personality if that was the case.
Microsoft technically delivered something very close to OS/2’s “Personalities” in Windows NT 4. They called it "Environment subsystems". Each subsystem could run applications written for different operating systems, the 3 available ones were Win32, OS/2 and POSIX. Then there was the "Integral subsystem", which operated system-specific functions on behalf of environment subsystems.
But every subsystem other than Win32 was kneecapped mostly due to politics and market positioning.
In late 90s Microsoft bought a company which had developed a more enhanced Unix subsystem and rebranded it as Interix and marketed as Windows Subsytem for Unix (SFU).
I believe the original WSL was a resurrection of SFU before WSL2 pivoted to a VM-based approach.
Nope, WSL 1.0 was based on pico-process, see Drawbridge project.
https://www.microsoft.com/en-us/research/project/drawbridge/
https://learn.microsoft.com/en-us/archive/blogs/wsl/pico-pro...
Thank you for the correction
No, the original WSL was a weird new thing where NT kernel-level driver actually serviced Linux system calls.
IIRC, Interix still used same approach as original posix subsystem (and Windows and OS/2 subsystems) of providing the interface as DLL that ultimately your application would be linked against.
Thank you for the correction
To the nostalgics among us: what made OS/2 special? 32bit support?
I recall trying OS/2 2.0 or 2.1 back in the day, coming from a DOS/Win3.11 setup. It seemed to have the same basic features as DOS/Windows but wasn't properly compatible with my existing software. Admittedly, this was before I knew anything about programming. I discovered Linux not much later. It wasn't compatible with anything either, but seemed like a totally different and much more compelling proposition.
> To the nostalgics among us: what made OS/2 special?
I started out with OS/2 v1.1. It had threads, DLLs, multi-tasking, much larger memory space, and from v1.2 a somewhat decent filesystem. Coming from DOS 3.2/Win 2.0 this was an incredible leap, in particular the SDK was amazing compared to the ragtag assembly of info I was used to. The _delta_ between two systems haven't been this large ever since, and I think that is what contributes to the "magic" feeling.
The OS/2 WARP Presentation Manager was a better "desktop" paradigm than Win3.11. It supported more customization and stranger "objects" you could store on your desktop. It felt a bit more coherent and a lot more powerful than the Win3.11 Program Manager.
I was mostly a kid with a huge stack of PC games I'd play, and OS/2 was a better launcher for many (but not all) of them than DOS/Windows. I was "dual booting" OS/2 WARP and DOS/Windows, but because of my gaming habits it was more like quintuple booting because I had a long boot menu with I want to say 4 to 7 different combinations of AUTOEXEC.BAT/CONFIG.SYS settings depending on type of game I intended to play or if I was going to use a Windows app or something else, then one OS/2 WARP boot option.
A bunch of Windows apps (many of which ran better, even) and even some games I recall I started launching from OS/2 WARP instead of DOS/Windows, making the first boot choice of the day a lot easier. (Though I don't remember being able to delete most of the other combinations, still had to reboot for certain games and Windows apps that needed more RAM than what OS/2 left for applications. OS/2's biggest problem at the time was a huge RAM footprint compared to Win3.11, much less DOS micro-tuned with AUTOEXEC.BAT/CONFIG.SYS low footprints for specific driver combos.)
Multitasking, SOM (contrary to COM, it does implementation inheratance across languages, multiple inheritance and has meta-classes), object based desktop, Smalltalk for business application development (basically a similar role as VB and .NET have gotten latter on on Windows), Visual Age for C++ had a Smalltalk like experience (although ported to Windows as well).
However this also meant a more beefy hardware than the DOS/Windows 3.x combo.
Back in 1995 it was, to my knowledge, the only OS capable of sharing CD-ROM's on the network. Even MS-DOS and Windows 3.11 machines could access it.
It was also capable of sharing Mainframe printers using a real null-printer-driver, which was not possible on Windows NT3.51 or NT4.0. Windows always messed with the Mainframe codes that it could not understand.
It was also easy to set up OS/2 as a gateway between different network hardware and protocols (Token Ring to Ethernet, or NetBios to IPX/SPX, ...)
It had REXX!
REXX was very powerful albeit a little quirky.
Novel Netware.
Probably a lot of things. Often software is simplified, at the time because of limited hardware and probably other software. Nowadays it's often a deliberate product decision but it seemed for OS/2 no such limits existed. E.g. you could right-click on a program, get the properties, run multiple applications. It even had a Windows emulation so stable that it was never matched by WINE. Of course there was only 16 bit Windows support but still...
Of course it had limitations of its own, I don't think you could any DOS/4GW games. Linux Installation seems simple compared to installing OS/2. I had to go through some sort of pre-installation guide which was printed on a separate paper and not part of the official manual. Also dual boot was meant literally: you booted into OS/2 and then you could "exit" into Windows. Back in DOS/Windows there was a command to do this the other way around. One time I didn't do this for half a year and was really anxious if my setup would make it...
At the time, formatting a floppy disk was a single task thing.
Downloading a file via Zmodem was mostly a single task thing.
The Windows of the day could to the latter, but not the former.
OS/2 on my 8mb 386sx could do both AND have a clock up, and play solitaire, and have another terminal window open.
It took a bit to get there, and there was swapping while everything loaded, but it was true pre-emptive multitasking, while still maintaining the highly time critical I/O stuff that Windows couldn't touch.
OS/2 performed better than windows generally and was more stable.
Back in that time period tech specs, and tech details really dominated a lot of "computers" discussion. I feel like that has kinda changed as far as the larger world goes (even if on HN tech specs are still relevant). Does an every day user want to use it? was less of a question for enthusiasts.
You could print while playing Doom.
Or while formatting a floppy disk!
That was huge.
People today don't realize how much time was spent formatting floppy disks, and how slow the process could be. So slow that eventually companies started selling pre-formatted disks and charging extra for them.
OS/2 could give you back hours of productivity each month simply because you could do something else while formatting a disk.
The problem was Windows handling of interrupts. Until, i think, Win2k, generating a lot of interrupts (floppy, parallel port, high network activity) will slow Windows to a crawl.
Stability.
Computers were far more crashy in those days, but OS/2 crashed far less often than Windows or even DOS did. And sometimes when a program crashed on OS/2, it only killed itself; it didn't take down the whole machine, so you had a chance to save your work in other programs before rebooting.
It also either was, or felt like it was, very very fast. Windows felt like a laggy VNC connection. GEM and the rest weren't much better speed-wise than GEOS on a Commodore 64.
Yes, stability! It was very nice when doing DOS or Windows 3.x development.
The only stability gotcha was when some OS/2 PM application hung the input queue and then the whole of PM became unresponsive. The base OS continued running fine but PM was then unusable.
Is the Windows in your comparison inclusive of NT, or only non-NT windows?
Nortel Meridian PBX systems ran OS/2 Warp on a PowerPC processor.. those systems were rarely rebooted. I bet there are many still running.
Wow, PPC? I once had a Norstar NAM for voicemail, and did connect a monitor and keyboard and witness OS/2 on that. I think the NAM was an x86-based system. But I know the Meridian PBXs themselves were completely distinct from Norstar that I used. Never knew Nortel used PPC. Very neat!
Though my time with them came only as non-IP phone systems started to be considered obsolete, I am still a huge fan of the rock-solid stability and realtime speed of those digital systems. Not to mention their lack of a need for subscription services to operate.
Some ran the AMD K6-2 as well...
Probably some but those old towers were not VoIP capable and the handsets are no longer sold. I doubt there's many around.
The voicemail card was basically an 8088 and i think still used floppies IIRC. I didn't work with them directly but I used symposium heavily as an admin (symposium was the callcenter solution)
I just saw one last week in a building we're looking to purchase. Still in use by the current occupant.
After returning a 'Reader Service Card' I found in Info World magazine, I started getting catalogs from Indelible Blue. They were _the_ big reseller of OS/2 software. Lots of good stuff in there, plus they had articles evangelizing OS/2 and ReXX scripting.
I had a brush with the PS/2 and OS/2 back in the day. IBM offered the PS/2-based Personal/370 on the platform with a card that could execute S/370 SW. My team built the I/O channel card for connecting 3480 tapes and the like. It was funny seeing the fat bus and tag cables attached to a little box through an adapter pigtail.
I’m always curious how these projects come about and survive: why go to all of the effort to port for a dead-end product line? As technically sweet as it is? I imagine they would’ve found a decent market if they’d ported to Power Mac.
(Also, was the x86 emulation implemented in-house? I wouldn’t be surprised if some niche small company had a x86 emulator for PPC product that they could be paid to port.)
The plan was for all operating systems on top of IBM's POWER/PPC hardware to be rehosted as "personalities" on top of the Workplace OS microkernel, but in the end, OS/2 was the only personality that saw any real work.
The Workplace OS would also have been used on Apple hardware as part of the abortive Taligent project.
(It also would have been used on x86 and other platforms, but they started with PPC)
In reality, in some ways we are there now. I'm wondering if we can say that the "workplace OS" can simply viewed as a hypervisor and the "personalities" that run on top it are simply VMs (perhaps being paravirtualized).
I think oddities like this were a consequence of a hardware world that was rocketing along the heart of Moore’s Law, alongside a software world that hadn’t matured past multi-year product cycles.
When OS/2 for PowerPC was set in motion, that Intel would “Make CISC Great Again” with the Pentium was far from clear.
I remember that the "general consensus" was that RISC was gonna win, it was just a matter of when (and when it could be affordable). What was NOT certain was which RISC architecture would come out ahead, so there was a bunch of porting to "remove the risk" - later they would unport most everything and "remove the RISC".
Pentium shook that tree a bit, and Pentium II really razzle-dazzled it.
Well, the thing is that RISC did win. It is just that the RISC that won is the one that Intel baked into their x86 chips.
The Pentium introduced the idea of micro op codes though the Pentium Pro was the first chip to really run with it. The CISC x86 instructions were converted into simpler instructions internally. These micro op codes could be pipe-lined, executed in parallel, and executed out-of-order.
If the Pentium II really razzle-dazzled, it did it with RISC architecture at its core. The CISC instruction decoder added a bit of die size but that did not matter much and Intel had leading-edge manufacturing tech.
The internal parallelism was also put to good use by adding SIMD instructions (MMX). These first appeared in the Pentium MMX and Pentium II but the Pentium III did it much better and of course Intel has continued to add more powerful SIMD stuff over time.
RISC did not win only inside Intel chips of course. Every successful ISA since the 90's has been RISC including ARM and RISC-V. But even RISC chips feature some complex instructions these days.
I'd argue (to some extent) that the proliferation of SIMD instructions demonstrates that RISC did lose, not just the practical war, but also the conceptual one. i.e. we creates many many similiar instructions today, which seems to go against the ethos of RISC.
I'm not sure I agree with "dead end" outside of the benefit of hindsight, or maybe don't get the point you're making. Neither the PowerPC nor OS/2 were dead-end in 1995, and competition in the OS space was still happening. Why wouldn't IBM want to have PowerPC survive, let alone thrive, with OS options? And surely they'd have loved something to take on Microsoft at this point in history.
I believe the Support Elements for some IBM zSeries mainframes were ThinkPad laptops with PowerPC CPUs running OS/2.
There was definitely VirtualPC for PowerPC Macs, I used it to run TurboTax way back in the day.
Didn’t know that OS/2 had a PowerPC port, but more surprisingly, Windows NT also had a PowerPC port. Never heard of those.
One of the original design requirements for NT was that it be portable between different CPU architectures, it was one of the driving forces behind its creation.
So much so in fact, Microsoft developed NT 3.1 first on non-x86 architectures (i860 and MIPS), then later ported to x86, to ensure no x86 specific code made it in.
NT supported quite a few architectures:
> https://en.wikipedia.org/wiki/Windows_NT#Supported_platforms
"Windows NT 3.1 was released for Intel x86 PC compatible and PC-98 platforms, and for DEC Alpha and ARC-compliant MIPS platforms. Windows NT 3.51 added support for the PowerPC processor in 1995"...
NT is a pretty interesting bit of PC history, I can highly recommend the book "Show Stopper!" by G. Pascal Zachary that recounts its development, and also dives a bit into why making the OS portable across CPU architectures was so important to the team at the time.
Something I didn't realize until recently was that the original MIPS version of Windows NT was Big Endian. I'd always heard it said that WinNT was strictly, 100%, absolutely always little endian, and the fact that every CPU that got a port (or was going to get a port) was either little or bi endian confirmed this.
Well, it is true, but Windows did run BE on the original MIPS R3000 platform. And only on the R3K[0]. The CPU architecture flag is still defined on modern Windows as IMAGE_FILE_MACHINE_R3000BE. There's an early test build of Win3.1 + GDI somewhere that runs on this platform.
The actual first release of WinNT 3.1 only supported MIPS R4000 and higher, I think. In little endian mode.
[0] I know the Xbox used a modified NT kernel, I've seen claims that the Xbox 360 also was, which would make it the second NT system to run big endian. Not familiar enough with sources better than wikipedia to trust that it actually was.
I believe you're correct - Xbox runs a modified Windows 20000, and subsequent versions I'm not certain on but I know it basically stands up a modified Hyper-V and the parent partition is the interface OS with games booting whatever kernel they were built for inside a VM
One of my first job out of school was as a sales support for the then bleeding edge NT 3.1 MIPS box for a company in Canada. Fond memories of loading stacks of 1.44 floppy disks for NT 3.1 and mangling ARC paths (Advanced RISC Computing, boot firmware). This was pre-internet and documentation was often hard to come by, incomplete etc.
I remember demoing the machines to astonished clients by running a stupid number of Clock apps on the desktop without a hitch.
Fun times.
My first real job out of school was supporting Windows NT on Dec Alpha for a company in Canada.
Things were so weird and wonderful back then. You could get GCC from Microsoft for Windows NT 3.1 for Alpha (crazy). And when Windows NT 4.0 came out there was the FX32 subsystem that ran X86 apps on Alpha (very similar to Apple Rosetta but much earlier).
I did not realize Canada was such a hotbed of Windows NT RISC.
Interesting historical note: the main reason PuTTy exists is because its author was given a Windows NT on alpha workstation and there was no native terminal emulator for it that he needed to connect to other equipment. IIRC, PuTTy still supported alpha into the 2000s until the build machine he had failed.
I know, I was a Windows engineer, I knew it had been ported to many architectures, but somehow I missed PowerPC :)
It gets weirder.
Nintendo GameCube and Wii are also PowerPC based. And somebody managed to have them run Windows NT: https://github.com/Wack0/entii-for-workcubes
It was also on mips and alpha. There was an intergraph port as well that never went out
NT was great on Dec Alphas.
Solaris (2.5.1 at least) had a PowerPC port as well.
Huh I wonder if this would run on my G4 Mac mini. Or my G4 iMac. I still have both. Would be cool to try.
Never used OS/2 Warp. Is the Document shredder icon the recycle bin ? Love it.
Yes and no. It was a permanent delete, no visible GUI with deleted files. But there was an actual UNDELETE command that could be used to recover files (or not)
Shredder https://files.catbox.moe/ciheyq.png
Undelete https://files.catbox.moe/ny7xbs.png
What could have been. If the respective parties had just gotten their acts together on the PPC 615, OS/2, WordPerfect, and Lotus.
I remember working with a guy who was the internal IBM rep (we were in corporate tech support in the late 1990’s and we were a BIG IBM reseller, so my friend got lots of bennies at work) and he had a PPC 605 (not the 615 IIRC), but he ran AIX on it (we were in the AIX and Unix support team).
Man I was jealous!
> What could have been. If the respective parties... on the PPC 615, OS/2
There was never a chance at that time because x86 chips were produced in such volumes that PowerPC chips couldn't compete price-wise. Also, OS/2 became an instant outsider once Windows 95 was released. Two underdogs don't make a winner. The article says it all:
"The OS was clearly unfinished and not entirely stable. Worst of all, there were about zero applications. Because OS/2 PPC was never truly in use, PowerPC versions of OS/2 applications were never sold."
Was there any act that would have overcome the synergy of Intel’s commodity hardware economics and Microsoft’s ecosystem dominance?
Yes, getting stuff together and getting it out there.
Windows 95 ate the world because the world was mainly still DOS; look at the numbers. It wasn't people upgrading from Win 3.1.
One of the big reasons people were upgrading to OS/2 was people that wanted a stabler/cleaner Windows 3.1. Most of the hottest apps for OS/2 were Windows apps. IBM started on the back foot in that competition. Windows 95 clipped their heels because it had the stability/cleanliness for cheaper and less of a RAM footprint.
Additionally, while this is US-centric, there were still many households in the mid-1990s whose first computers were PCs running Windows 95, just in time for the World Wide Web to be widely available, which created demand for personal computers. Additionally, this was during the time when Apple was struggling; its Performa lineup geared toward home users was not in the best of shape in 1995 (https://en.wikipedia.org/wiki/Power_Macintosh_5200_LC). By the time Steve Jobs returned and Apple released the first iMac (1998), it was just about time for Windows 98.
Being at the right age when Windows 95 came out, I didn't really know that there was a "Windows" prior to 95. My dad's computer ran DOS and used something called Powermenu as an organiser for executing programs. I think I had to run Wolfenstein in a tiny window for it to be fast enough to be playable, and may have, at one point, deleted one of the required DOS system files in order to try to tweak the life out of it to try to get it playable full screen. I think that was a 286. More years ago than I care to admit.
What numbers do you have in mind?
I was paying very close attention to this world at the time, and I am confident that 90%+ of every PC that could run Win95 at the time of its release had been sold by an OEM as booting straight into the bundled copy of Win3.1.
Yes, lots of people had older PCs and got Win95 with their hardware upgrade cycle.
They were never in any danger of upgrading to OS/2 or anything else: Microsoft owned the OEM sales channel and the software ecosystem for the PC for a solid 5+ years before Windows 95.
What numbers? As much as I hated Windows 3.x (which is why I upgraded from DOS to Linux, not to Windows 95, and never looked back) it did not occur to me that many by 1995 did not have some Windows 3.x installed, as it was required for so much software (even some games).
Windows 3.1 sold 3 million copies in the first three months, Windows 95 moved ten million copies in the first year.
Everyone I knew went from either no PC at all, or an older DOS 386-era machine to a Windows 95 computer.
3M copies in 3 months is more than 10M in a year (if you assume sustained sales).
I don't get what the numbers are supposed to imply.
Windows 3.1: ~3 million in first six weeks, ~>3 million in first three months, ~25 million in first year.
Windows 95: ~1 million in first 4 days, ~7 million in first five weeks, ~40 million in first year.
These figures represent Microsoft’s own sales figures.
Still don't support the claim that people were mainly updating from DOS without Windows 3.x. Anecdotally I still think almost everyone using DOS by 1995 had Windows 3.x installed as well. Not necessarily a copy of Windows that was the result of Microsoft selling a copy of course.
Yeah those don't match parent's comment.
Hey give Windows 3.11 FOR WORKGROUPS some respect ;)
Apple somehow managed to claw it's way to releavance from a weaker position in 1998 (with PoserPC!) So if they had their act together they could have done better in the early 90s.
hey squandered their early lead in the US among consumers and education and also ignored the international market.
Not gonna lie Wintel was a formidable force. Microsoft was ruthless in cornering the market.
But technically, OS/2 and MacOS gave Windows a run for it's money, arguably superior on some respects, and you could say the same for PowerPC and Intel.
‘SteveB went on the road to see the top weeklies, industry analysts and business press this week to give our systems strategy. The meetings included demos of Windows 3.1 (pen and multimedia included), Windows NT, OS/2 2.0 including a performance comparison to Windows and a “bad app” that corrupted other applications and crashed the system. It was a very valuable trip and needs to be repeated by other MS executives throughout the next month so we hit all the publications and analysts.’
‘The demos of OS/2 were excellent. Crashing the system had the intended effect – to FUD OS/2 2.0. People paid attention to this demo and were often surprised to our favor. Steve positioned it as -- OS/2 is not "bad" but that from a performance and "robustness" standpoint, it is NOT better than Windows’
"I have written a PM app that hangs the system (sometimes quite graphically)."
http://iowa.gotthefacts.org/011107/PX_0860.pdf http://iowa.gotthefacts.org/011107/PX_0797.pdf
Great links!
I don’t see how SteveB’s “bad app” was out of bounds. IBM was telling these analysts a bad app couldn’t bring down OS/2. As someone who used (and liked!) OS/2 at the time, this was indeed false: IIRC OS/2 in this era had a single-threaded UI loop, and a misbehaving app could (and occasionally did) block input for all apps and corrupt the UI.
You could argue about the difference between this and a Windows “UAE”, but to the user the experience was a broken, unusable OS/2.
DOS ain’t done ‘till Lotus won’t run.