The pivotal importance of IBM's PS/2 range on computers, even today

https://liam-on-linux.dreamwidth.org/87108.html

by AndrewDucker • 4 years ago

63 84 4 years ago

liam-on-linux.dreamwidth.org

h2odragon• 4 years ago

Also the PS/2 keyboard and mouse connectors. Before that mice were speaking to serial ports or required an add in card with a nonstandard port. the people buying PS/2's were often very concerned with mice, they'd seen them on TV and they wanted one.

b06timmer2• 4 years ago

That's why the were called PS/2 mice and keyboards. I still have a couple of serial-to-PS/2 adapters here somewhere.

chasil• 4 years ago

Most mice used an rs232 serial port.

actionfromafar2• 4 years ago

"Bus" mice were popular too. [0] Cheaper to produce than RS2323 mice I think, but required the special interface card or motherboard support.

The bus mice were semi-popular with Amiga owners because they were easily converted to Amigas. RS232 mice could be used with Amigas too, but only worked in Workbench programs which used the operating system drivers. Most Amiga games etc used the mouse hardware registers and the RS232 mouse would be ignored then.

0: https://en.wikipedia.org/wiki/Bus_mouse

f1shy3• 4 years ago

I think there was nothing more standard than RS232. You needed to add a "multi" board (joystick, parallel, rs232), yes, but that was absolutely standard.

rjsw• 4 years ago

The Logitech mouse for the Olivetti M24/AT&T 6300 plugged into the keyboard.

bluedino• 4 years ago

Same physical port. Perfect for plugging into each other and having neither one work. Later on they color coded them (purple and green, some companies used orange?)

I was always jealous of the Mac's ADB. You could daisy chain it, plug any device into any port, it was nice.

Eleison232• 4 years ago

It didn't originate with the PS/2 line, but the Model M keyboard with Selectric roots was standard equipment, an absolute beast, and the best keyboard I've ever had.

midoridensha• 4 years ago

The Model M's buckling spring mechanism has absolutely nothing to do with the mechanics of the Selectric keyboard.

anjel• 4 years ago

Your not wrong, but the design of the model M sought to emulate the feel of the Iconic Selectric, as well as the model F with the solenoids under each key, equally dissimilar to the Selectic's internals.

f1shy3• 4 years ago

I regret so much having dumped that keyboard... best thing ever. Only problem: when typing at night (what I really liked) the whole block was awake! :D

rbanffy3• 4 years ago

It's funny. The Model M was a way to make a cheaper keyboard than its predecessor, the Model F, which, in turn, was IBM's way to make a cheaper, lighter alternative to the beam-spring keyboards they were using in terminals and other "more serious" computers.

Eleison232• 4 years ago

OS/2 Warp was out as I purchased my 386, and I picked up a copy, and I really wanted that to work, but it was already clear that IBM was playing catch-up with Windows because Warp was marketed based on how well it could run Windows stuff.

b06timmer2• 4 years ago

I thought OS/2 Warp was really neat. I worked IT at Bloomingdale's and at the time they were a strictly IBM house. They used IBM computers running OS/2 Warp in their Bridal Registry kiosks. Since this was a kiosk they had touchscreens and they worked very well and had a "sparkle" in the screen, as the early touchscreen CRT's did. If you lost power, or shut them down incorrectly you would get a bunch of DLL errors and have to re-image the PC.

This was the late '90s and early 2000's.

postexitus• 4 years ago

re-image? did that exist then?

karmakaze• 4 years ago

I ran OS/2 1.x (in particular 1.3 was nice) on a 386 clone PC.

Later I ran OS/2 2.x on a PS/2 Model 77 and later Windows NT 3.51

That is why MS had the money to headhunt the MICA team from DEC, headed by Dave Cutler, and give them Portable OS/2 to finish. That became OS/2 NT (because it was developed on Intel's i860 RISC chip, codenamed N-Ten.) That became Windows NT.

Good bit of trivia. I remember it starting out as OS/2 v3 and being marketed as Windows New Technology.

superkuh• 4 years ago

Mirror for those that get blocked by the dreamwidth ddos "protection" like me: https://archive.ph/mLhYa

On the subject, I am typing this response with a PS/2 keyboard from 2003. I use an old computer with old software. Which is why dreamwidth blocked me.

Fatnino2• 4 years ago

I'm typing through a dual ps/2 mouse and keyboard to USB adaptor. Mouse side is empty and keyboard side has a gc56mf to ps/2 adaptor on it. Because my keyboard is from the 80s.

lproven17• 4 years ago

Depends which keyboard I use, but sometimes, same here. My main work PC has a 1991 Model M on it. Best PC keyboard ever made, and I have about 5 of them, carefully kept since people were throwing them away in the 1990s.

I remap Caps Lock to Super, and I'm good to go. I really don't need media-control keys or anything, or a right Super key.

I am trialling an early model of Das Keyboard on my Mac. The volume-control dial is quite nice, because I listen to music and watch films on that. I never use the media keys, and happily, it's old enough that it doesn't have nonsense like light-up keys, multicoloured lighting or any of that ugly frippery.

Most computer lighting and decorations are visible designed to appeal to teenaged boys, with all the taste and discrimination that implies.

aliqot• 4 years ago

http://www.postbulletin.com/magazine/2004/08/index.shtml

"According to Bob Dubke, the second engineer on IBM's 5100 team in Rochester (who now co-owns a locally-based company called eXport Ventures Corp. and also works for Edina Realty), that secret function was his contribution to the design of the computer. The function, which IBM suppressed because of worries about how their competition might use it, was an interface between the assembly code surrounding the computer's ROM exterior, and the 360 emulator hidden beneath it. (IBM declined to comment for this story.) The 5100's emulator gave programmers access to the functions of the monstrous, and much less portable machines, that IBM had produced during the 1960s. An imprint of a hook on the outside of the 5100 symbolized the ability of Dubke's interface to drop into what Titor called "legacy code," and scoop out any necessary operating instructions."

mananaysiempre• 4 years ago

Had no idea that the Steins;Gate John Titor was RL-(hoax-)inspired, huh.

klelatti5• 4 years ago

Interesting comment but the link brings up a 404 now for me.

lproven17• 4 years ago

Try this.

https://patiomensch.wordpress.com/further/john-titor/

Hint: just Google a fragment in quotes, for an exact match.

rasz3• 4 years ago

72pin simm in 1988

Afaik other companies picked up 32bit simms very slowly. SUN in 1991 in SPARCstation IPX. First x86 support showing up in 1992 on 486 boards just before introduction of VLB slots. Apple was year later with LC III/520 and Centris in 1993.

hakfoo2• 4 years ago

I have a 386SX that has a single 72-pin SIMM socket. It xseems to be a 1991 design by the stickers on components.

Even more weirdly, the machine could be equipped with a 286 daughterboard instead of the 386SX. I wonder if it just discarded half of each memory access to fill the 16-bit bus.

rasz3• 4 years ago

brand/model? There was also Compaq DeskPro 386/20N, came out in ~December 1991 https://books.google.com.au/books?id=KT0EAAAAMBAJ&pg=PA106&l...

lproven17• 4 years ago

Sure, yes, because the purpose of 72-pin SIMMs was for 32-bit wide memory access. Before 72-pin SIMMs were 30-pin SIMMs.

Simple arithmetic shows you that you can't route a 32-bit memory bus over 30 pins. You need banks of SIMMs and that means they need to be matched.

But the 386DX was too expensive, because it needed a 32-bit memory bus. So it didn't sell well and was mainly used in PCs used as servers. PC servers didn't really exist yet in 1985.

That's why the 386SX was invented, with a 16-bit memory bus. It meant the motherboard only needed a 16-bit memory bus, meaning it was much cheaper.

But that means it doesn't really need 72-pin RAM.

It wasn't until the 486 (and 68040) went mainstream that ordinary desktop computers needed 32-bit memory buses and therefore 72-pin RAM, but over time, economies of scale drove down the cost of 72-pin RAM so it became cheaper and gradually got adopted, and the 16-bit-bus 386SX (and derivative chips with the 486 instruction set, such as the 486SL and 486SLC) gradually went away.

rvense2• 4 years ago

68020 and 68030 were fully 32 bit, wasn't it? They had the dynamic sizing thing so they could boot of one EPROM, but they had the full bus expose AFAIK.

bombcar• 4 years ago

I recognize the HN comment!

https://news.ycombinator.com/item?id=33007995

lproven17• 4 years ago

Thanks.

Yeah, I probably should have linked it, but then again, I'm not out to name-and-shame anyone or anything.

tbyehl• 4 years ago

What I remember most about the PS/2 is the crazy power switch that internally connected to a metal rod that connected to another switch on the PSU itself.

hulitu• 4 years ago

People were more economical then.

f1shy3• 4 years ago

Oh yes! And the "cachang cachang" sound when turning on/off... was like a nuke switch!

Ekaros• 4 years ago

Ah, the magic way your computer actually uses 0 power when off. Maybe we should go back to those. Imagine how much energy could be saved in the world.

rvense2• 4 years ago

I've never seen a metal rod, but a long plastic arm that goes all the way through the case is common on many types of electronics.

hilbert424• 4 years ago

No doubt the pivotal importance of IBM's PS/2 range and I was instrumental in my organization acquiring many but that wasn't without problems as we already had many ISA machines so PS/2s weren't compatible at a board level. That increased the cost of maintenance amongst other things.

However, when I think back the two most disruptive problems for us were the development of the 80286 and the failure of OS/2. The 80286 was dog of a design not from its silicon per se but rather from a design concept. The switching modes problem, triple-faulting the processor and so on.

I understand the reasoning behind why Intel designed the processor that way as an explanation of why it did so took up most of the time of a long Intel seminar/presentation that I attended just when the chip became available.

It's easy to be wise in hindsight but it would have been better for everyone—except of course Microsoft—if the 80286 had never existed and that the 8086 platform had migrated directly to the 80386.

klelatti5• 4 years ago

Can you expand on the explanation as to why Intel said they designed the 286 as they did?

ForOldHack4• 4 years ago

This is a bit of history, I had to search a while for:

"The designers of the 80286 based it on the Multics segmentation model, but in reality segmentation makes it really hard to implement, because of internal memory fragmentation. A while back I wrote blog post about how an alternative 80286 with segment registers pointing to 256 byte paragraphs giving a true 24-bit virtual address which was then paged into a 24-bit physical address space. 8068 compatibility could then be managed by having the segment registers map to 16 byte paragraphs as before."

But: The screwed up on the switch back to real mode, and on the reservation of the 4th byte of the descriptor tables, which made it so that Xenix-286 cannot run on a 386. ( long discussion... I think that MS purposefully made it incompatible. )

hilbert424• 4 years ago

Too late to reply now, others have already jumped in with splendid answers. :-)

kristianp• 4 years ago

The wording of the need for the '286 isn't quite right. Windows 3.0 supported the 286 too, but also had support for the 386's protected mode.

Related fact, Windows/386 2.0 supported protected mode[1].

[1] https://en.wikipedia.org/wiki/Windows_2.0x

ForOldHack4• 4 years ago

Even without the reference, You were there, and I was there, too. I had to support a client, who needed Win/386 2.0 protected mode for a DOS app, as well as run PageMaker.

lproven17• 4 years ago

I wrote the comment and the blog post. What wording are you referring to?

Yes, Windows supported the 286 too. There was a special Windows 2 version for 286s, and a different edition for 386s; what was novel in Windows 3.0 is that one edition and one install on disk supported all three: 8086, 80286, and 80386, all in one. Windows 3.0 had 3 modes:

- if it had <2MB of RAM or just an 8088/8086/NEC V20 etc. chip, it started in Real Mode.

- if it had 2MB of RAM and a 286, it started in Standard Mode. No 386 features enabled.

- if it had >=2MB of RAM and a 386, it started in 386 Enhanced Mode.

https://komh.github.io/os2books/gg243731/214_L3_386EnhancedM...

https://devblogs.microsoft.com/oldnewthing/20130208-00/?p=53...

I have written about this, too: https://liam-on-linux.livejournal.com/78006.html

It wasn't well-known even then, but Windows 2 and 3 could multitask DOS apps without a 386. V86 mode wasn't needed for the multitasking, but for the memory management. The multitasking worked, the problem was trying to fit several DOS apps into the base 1MB of RAM for them to multitask.

What the 386 delivered was separate 640kB slots for each DOS VM.

cardiffspaceman• 4 years ago

I read about the multitasking of DOS programs in some round-up of DOS multitasking systems (DeskView et al.), and convinced my boss to fund experimentally adding 4 MiB of RAM and a Windows 3.0 license to my 386-based clone. I got a build (cross-assemble) of my code going, and then entered the Brief editor, and pondered what exactly I had enabled. Then an in-house utility was invoked from the makefile to download my code into the target, and my whole system froze. Thus endeth the experiment.

tpmx• 4 years ago

A HN poster questioned the existence of 80286 computers with VGA displays.

And then there were a bunch of Amstrad PCs with 8086 and VGA:

https://www.seasip.info/AmstradXT/models.html#pc2086

https://www.seasip.info/AmstradXT/models.html#pc3086

https://www.seasip.info/AmstradXT/models.html#pc5086

nottorp2• 4 years ago

A HN poster questioned the existence of 80286 computers with VGA displays.

Young uns... I personally had a 286 with VGA back then...

ForOldHack4• 4 years ago

Well, then I might as well point out, that IBM did make a 8-bit adapter for VGA, so I could use my Sony Trinitron with my PC XT 10Mhz clone. Beat the daylights out of my roommates 6Mhz AT, and ran Xenix, which most unfortunately turned me into a Xenix expert of crap ware. After a bad compile, it destroyed both drives file system, so I went back to just Windows 1.0.6, and DOS.

ido2• 4 years ago

My brother in law had a 286 with SVGA around 1990...

2000UltraDeluxe• 4 years ago

Me too. And a Sound Blaster.

thanatos519• 4 years ago

When my IBM PC 5150's CGA monitor lost its vertical sync (and I was on a BBS and vibrated my lips so I could read the text to logout properly ... really!) I got a Cardinal VGA card for the 8-bit bus.

So yeah, VGA was possible all the way back to the original.

chx• 4 years ago

Speaking of pivotal, both IBM and Microsoft poured an ungodly amount of money into translating Windows 95 and OS/2 Warp to a great number of languages, as far as I am aware this didn't happen before. I was a journalist at a computer monthly in Hungary at the time. Now, MS and IBM teams were not talking to each other, in fact Microsoft was working in "ivory tower" mode: the translation happened in house almost in secrecy. IBM was at least partially outsourcing to local firms who were coordinating with us. An awful lot of the very language pertaining to computers were invented or re-invented due to this. Yes, during the 80s of course there were already computer magazines and even the TV was talking about but it was absolutely awful and didn't fit the language well. The effort IBM put in was, well, pivotal. We didn't get paid -- but we had a major advantage of turning the magazine into readable and very soon we have eclipsed everyone else on the market. Of course this had other reasons but the effects of this can not be dismissed.

lproven17• 4 years ago

Relevant:

https://www.os2museum.com/wp/slovenian-os-2-warp-4/

pjmlp3• 4 years ago

PS/2 failure is also one of the reasons why PC compatibles stayed around, its main goal was to regain control of the PC.

lproven17• 4 years ago

The PS/2 was PC compatible.

The term means "can run IBM PC software". The PS/2 could. It was compatible. Same OSes, same apps.

Remember this was the 1980s. There were almost no protect-mode 32-bit OSes around, and the few that there were then had PS/2 versions. DOS and DOS apps ran fine on PS/2s, unmodified.

pjmlp3• 4 years ago

No it wasn't, it had the MCA architecture for the motherboards, which was out of reach to the clone makers.

https://en.wikipedia.org/wiki/Micro_Channel_architecture

It was thought that by creating a new standard, IBM would regain control of standards via the required licensing. As patents can take three years or more to be granted, however, only those relating to ISA could be licensed when Micro Channel was announced. Patents on important Micro Channel features, such as Plug and Play automatic configuration, were not granted to IBM until after PCI had replaced Micro Channel in the marketplace. The overall reception was tepid and the impact of Micro Channel in the worldwide PC market was minor.

https://en.wikipedia.org/wiki/IBM_PS/2

The PS/2 line was created by IBM partly in an attempt to recapture control of the PC market by introducing the advanced yet proprietary Micro Channel architecture (MCA) on higher-end models. These models were in the strange position of being incompatible with the IBM-compatible hardware standards previously established by IBM and adopted in the PC industry. However, IBM's initial PS/2 computers were popular with target market corporate buyers, and by September 1988 IBM reported that it had sold 3 million PS/2 machines. This was only 18 months after the new range had been introduced.

rbanffy3• 4 years ago

Not to understate the importance of the PS/2, but...

To say OS/2 failed because it didn't run more than one DOS app at a time is not really accurate. Well before Windows 3 (or NT) was established, OS/2 had its own app ecosystem and it was very pleasing to use, if you had a fast computer. Early OS/2 was meh, but from Warp and on, it was a very nice desktop OS.

OS/2 died because Microsoft cut deals with all big OEMs where they got Windows almost for free, as long as they didn't offer OS/2. It didn't help that, in order to run Windows software, OS/2 needed an expensive Windows license (and an almost full Windows).

There were other personal computers with GUI's back then as well - the Mac is still with us (well, not really - all current Macs are more NeXT than Mac), but there were Atari's and Amigas (and some Ataris and Amigas could run Unix). By then, any 386 with sufficient resources could run X (Sun had the 386i, running Unix and DOS apps side by side).

In the end, OS/2 failed because Windows had to succeed, and Microsoft made sure it did.

lproven17• 4 years ago

Are you talking about Microsoft/IBM OS/2, i.e. 1.x, or IBM-only OS.2, i.e. >=2.x here?

For 2 and above, the IBM only versions, well, yes, kinda sorta. But that was too late. The game was already over.

OS/2 2.0 came out in April 1992.

Windows 3.0 came out in May 1990, 2 whole years earlier. It already had established an ecosystem before 32-bit OS/2 appeared.

Secondly, OS/2 2 really wanted a 386DX and 4MB of RAM, and a quality PC with quality name-brand parts. I owned it. I ran it on clones. I had to buy a driver for my mouse. From another CONTINENT.

Windows 3.0 ran on any old random junk PC, even on a PC XT class box with EGA. At first only high-end users of high-end executive-class workstations got the fun of 386 Enhanced Mode, but that was all OS/2 2.0 could run on at all.

OS/2 died when OS/2 1.x was a high-end OS with low-end features, and a cheapo low-end 386SX PC with 1 or 2MB of RAM, with MS-DOS and DESQview (not DESQview/X, just plain old text-mode DESQview) could outperform it.

(Remember the 386SX came out in 1988 and was common by the time Windows 3.0 shipped.)

rbanffy3• 4 years ago

Windows 3.0 came out in May 1990, 2 whole years earlier\

Oh yes. IBM, it seems, was late in getting the news Microsoft wasn't as committed to OS/2 as IBM was. I imagine what was said at IBM's top management back then.

What Microsoft saw that IBM didn't was that having a continuous path without breaks was more valuable than having an objectively better product. That's also why we are using PCs today and not Amigas, Ataris, or Suns.

At least we can also use NeXT machines. ;-)

The pivotal importance of IBM's PS/2 range on computers, even today

Related Stories