Released April 6, 1992, Windows 3.1 was the successor to the very successful Windows 3.0. It wasn’t great, just like Windows 3.0 wasn’t great. But it was a graphical user interface that ran on very inexpensive ordinary PCs, and enough people considered it good enough that it more or less ensured the future belonged to Windows.
What was wrong with Windows 3.1
My blog post about what was wrong with Windows 3.1 is old enough now that you can consider it retro itself. The ironic thing about Windows 3.1 was that it came out near the dawn of the Internet age and yet was so unprepared for it. It supported high color displays and multimedia, it had scalable fonts, but had no networking capabilities. And it was a 16-bit environment that ran on top of an aging 16-bit operating system, MS-DOS, even though the CPU was 32 bits. Its multitasking was cooperative, rather than pre-emptive. It had no memory protection.
In short, it was consumer grade and full of compromises. If it was all you knew, you didn’t necessarily notice. I knew dozens of people, maybe hundreds even, who used Windows 3.1 every day and thought it was great. But if you’d used something else, like an Amiga or a Unix workstation like a Sun, Windows 3.1 felt rickety and unstable. If you were generally used to running one program at a time, it worked pretty well. But if you tried to keep several large programs in memory at once, it tended to show its weaknesses.
Why Windows 3.1 won
Windows 3.1 ran on inexpensive ordinary PCs that you could buy anywhere. And it came bundled with those PCs, so it wasn’t necessarily even apparent there was any alternative. You’d go to a consumer electronics store, buy a PC, and it booted into Windows. Windows 3.0 had been popular enough to attract software development, so a large enough ecosystem built up around it to catch on.
While Microsoft was anticompetitive toward larger software publishers, they did charge less for development kits and documentation than their rivals did. They also produced a larger variety of software languages for Windows. Microsoft provided C and C++ compilers for serious development, but also Visual Basic, which provided a lower barrier of entry.
Literally any other option available in 1992 would have been better than Windows 3.1. But none of them had been building momentum since 1990. There were people who believed it wasn’t too late in 1992. I was one of them. But the momentum wasn’t letting up in 1992 and no one else picked up any. So maybe the game wasn’t over right when Windows 3.1 was released in April 1992, but it wasn’t long before it was. Between Windows 3.0 and 3.1 sales, Windows had a larger install base than the Commodore 64 by the end of the summer. It sold 3 million copies in just its first six weeks, according to an article in the May 21, 1992 issue of the Washington Post. Windows 1.0 had taken two years to sell half a million copies.
If Windows 3.11 and Windows 95 hadn’t happened, then maybe it would have been possible to turn the tide. Maybe. But Windows 3.11 and Windows 95 did happen, even if Windows 95 was late. Flawed as they were, and Windows 98 after it, the improvements were enough to hold down the fort until Windows XP was ready nearly a decade later.
Over the holidays I was giving relationship advice and used the terms Mr. Right and Mr. Right Now. Windows 3.1 sure wasn’t Mr. Right. But in 1992, it was Mr. Right Now.
David Farquhar is a computer security professional, entrepreneur, and author. He has written professionally about computers since 1991, so he was writing about retro computers when they were still new. He has been working in IT professionally since 1994 and has specialized in vulnerability management since 2013. He holds Security+ and CISSP certifications. Today he blogs five times a week, mostly about retro computers and retro gaming covering the time period from 1975 to 2000.
could you explain why not GEM or os/2 in 1990/92 a success pc?
GEM had a couple of nice apps for it but nowhere near the software library Windows had. Ventura Publisher was the most successful app on GEM and I don’t even know what the second most successful one was. I can name two DTP programs on Windows that were more successful than Ventura: Pagemaker and QuarkXPress. Basically, by 1991 or 92 it didn’t matter what you wanted, there was a Windows app that would do it.
OS/2 had the advantage of being able to run 16-bit Windows software, but it needed twice the memory that Windows 3 or 3.1 did, and not everyone was willing to do that. I ran OS/2 until around 1996 and was happy with it. But I do think a lot of publishers skipped making a 32-bit OS/2 app since you could just run the 16-bit Windows app on it. A bigger 32-bit software library by 1993 or 94 would have given OS/2 a better chance than it got. I really liked OS/2 but every time I say that you’d be amazed how many people tell me I’m nuts. And in the early 90s every time I mentioned OS/2, people would just say “What about Windows NT?” and use that to justify staying on Windows 3.1. No one I knew considered NT until NT4 around 1996 or 97, but the mere existence of NT was enough to keep people from considering OS/2 for some reason I never fully understood.
thanks
I knew a neighbor who had the Samsung Sensor 286 with GEM in 1992 I was impressed
my pc lab had both IBM Aptiva and OS/2
how was NT4
btw I owned a Tandy 1000 and used deskmate
what do you think of BeOS and Next Step on x86
BeOS was interesting but there was so little software for it I didn’t see it as practical. When I wanted a fast computer with no software I could just use my Amiga. I used Nextstep on a 68040 and found it slow, and, again, so little software. If it was slow on the 68040 it wasn’t going to be fast on a 486 either. There was no reason to run Nextstep when I could just run OS/2, which gave me the stability of Nextstep while being way faster and being able to at least run DOS and Windows software.
did you see BeOS demonstrate their abilities ? like running multiple video clips and dropping on a book ?
OS/2 iirc had a buggy
didn’t Nextstep have object oriented environment ?
Yes I saw the demo. It was an interesting OS with no interesting software available for it. As for OS/2 being buggy, I never had problems with it. And yes, Nextestp was object oriented and that’s one of the reasons it was so painfully slow.
i heard some audio interesting software available for it
if successful completion may be nteresting software available for it
i didn’t know Nextestp was object oriented and that’s one of the reasons it was so painfully slow. maybe 486 but perhaps the pentium pro 300mhz?
could Nextestp run on a mac os?
Nextstep became Mac OS X. Getting it to run on Apple hardware took time.
I don’t think Nextstep would have run decently even on a Pentium II at 300 MHz given how painful OS X was on G3 and G4 hardware.
The early 90s were an interesting, transitional period. Windows 3.1 was kind of useless, except to run a very few programs specifically written for it (kind of like you needed Windows 2?(?) to play Balance of Power).
Up through 1995, all our home computers used DOS. We were all LAN partiers. In ’95, I started working computer tech support. The first place I worked at, it was all DOS except for one person in the office who had an experimental advance copy of ’95. In ’96, I did tech support on a Win3 machine, while another used OS/2. I dunno if there was a ’95 machine in the office.
It wasn’t until ’97 that ’95 became the standard, when I worked at Medtronic. At that point, it was the clear winner.
By the way, I just discovered your blog, and I like it quite a lot. I wish the ability to comment was left open longer as there are many articles I would like to comment on. Over at my blog, galacticjourney.org, I leave them open indefinitely, and I’ll get comments even a decade after the fact. 🙂
Thank you. I used to leave all posts open for comments indefinitely, and like you, I’d get comments on decade-old posts. But I found an inordinate number of comments I get after a few weeks were spam and/or people looking for someone to argue with so I had to lock it down. My regulars get a chance to comment, and I don’t have to waste energy and effort dealing with trolls. And it gives me a way to give a perk to my regular readers.
Something I’ve taken to in the last year is republishing posts when I revise them, which opens them up for comment again.
Oh, I could see spam being annoying. Akismet blocks all the spam on our site, but I don’t know if that’s available to you. Sorry about the trolls!
Akismet is a lifesaver, I would have had to shut down years ago, or at least shut off comments, without it. The CPU load of protecting 5,000+ blog posts was the biggest problem I saw. It doesn’t affect the front end all that much but it did slow down the administrative side a lot. Why I have 5,000 blog posts is another question…