SoylentNews is people
SoylentNews
It was twenty years ago yesterday (August 24, 2015) that Windows 95 was introduced, says El Reg.
Windows 95 was a great success, despite not being the most stable of operating systems. Microsoft's own Windows NT 3.1, released two years earlier, was built on stronger foundations, but high system requirements and lack of compatibility with many DOS applications and games made it unsuitable for consumers. Windows 95 was better in both respects, running in as little as 4MB of RAM – though painfully, with 8MB a more realistic minimum – and retaining DOS complete with 16-bit device driver support.
At the time, most PCs ran Windows 3.1 or 3.11 (Windows for Workgroups), and IBM was pushing OS/2 as a "better Windows than Windows". Windows 95 was a considerable improvement on Windows 3.x, with pre-emptive multitasking, mostly 32-bit code, and plug and play hardware detection. There was also new support for "portable computers", with a battery indicator on the taskbar and the ability to suspend the system without turning it off completely.
Perhaps what I'm going to say will be controversial, but I'm of the opinion that Windows 95 is the greatest software engineering feat ever, given the challenge Microsoft faced at that time. Unlike Apple, which continues to make its own computers, Microsoft did not and, therefore, had to do a vast amount of testing in order to ensure that Windows 95 would work on most existing 32-bit Intel machines.
Original Submission
(Score: 2, Interesting) by SomeGuy on Tuesday August 25 2015, @04:01PM
From a technical perspective, yes OS/2 was "superior". It (2.x and later anyway) was a true 32-bit operating system like Unix, Linux, and Windows NT.
But from a practical standpoint, it didn't run enough of the software people needed. Or if it could run their software, it required extra hassle.
I remember around 1995 or early 96 dealing with an office IBM P90 computer that originally came pre-loaded with OS/2. The office IT had wiped that and installed Windows 3.1 so the users could use the software they NEEDED. And the office IT had not even bothered to install the machines VGA or sound drivers (which I had to clean up), and yet this was still "better" than OS/2 because they could do the work they were required to do. The users were even more happy when we finally got Windows 95 on it.
(Score: 5, Interesting) by Thexalon on Tuesday August 25 2015, @04:59PM
It's the oldest IT problem in the book: Technical perfection versus "How hard is it to do my (non-technical) job?"
From the point of view of users, the occasional crash and stupid modal confirmation dialog boxes were relatively small prices to pay for something they could actually work with. Whereas if you'd given them an OS/2, Mac, or Unix box, good luck with that.
But from the point of view of the tech people, Win95 was a nightmare, because it could so easily run things it shouldn't and could not be easily fixed when something went wrong (which it would).
I strongly disagree with the notion that it was the greatest software engineering feat ever, though: The greatest software engineering feat I can think of is the creation of the C programming language, because 43 years later we're still using it or one of its descendants for the vast majority of critical software across all major platforms. It's basically impossible to find a computing device that is *not* running something written in C. No other software tool has come even close to matching that level of widespread use and longevity.
The only thing that stops a bad guy with a compiler is a good guy with a compiler.
(Score: 0) by Anonymous Coward on Tuesday August 25 2015, @07:23PM
If you don't *start* with a security model, what you end up with is a mess.
-- gewg_
(Score: 1, Insightful) by Anonymous Coward on Tuesday August 25 2015, @07:28PM
And even then you are not guaranteed success.
(Score: 2, Interesting) by Anonymous Coward on Tuesday August 25 2015, @05:03PM
>But from a practical standpoint, it didn't run enough of the software people needed.
>Or if it could run their software, it required extra hassle.
This is exactly the situation that Linux is in today. Wine will run some Windows software, but maybe not.
BTW - I still think that OS/2 had the best user interface of any operating system I've used. Presentation Manager made it easy to modify features so that it worked for me rather than for some marketing team in another part of the world.
I never used Win95 - I went from Win3.1 to OS/2 to NT. Now I only use Linux.
(Score: 2, Informative) by Anonymous Coward on Tuesday August 25 2015, @05:24PM
From a publisher POV OS/2 was a non starter.
The kit to develop was *expensive*. Think 5 digits per developer expensive vs MS 500-2k per dev. Then on top of that it 'ran everything else'. Which means I could target win95/3.1 and hit 99% of the market AND OS2.
I worked in a couple of companies and remember the meetings of 'do we develop a special version for OS2'. The answer was always no as there was a very small market for it. Though the market that was there was willing to pony up big bucks.
Early versions of OS2 were also notoriously picky on what they could be installed on (much like NT3). I underclocked many computers to get that sucker just to install. And doing the 20 disk shuffle on an under clocked computer was not fun.
(Score: 2, Informative) by canopic jug on Tuesday August 25 2015, @05:45PM
But from a practical standpoint, it didn't run enough of the software people needed.
Back when the courts had teeth to enforce antitrust remedies, IBM was prohibited from making both the hardware and software. IBM had a deal with M$ where IBM and M$ would collaborate on the OS and that M$ would be ready with user software. M$ told IBM at the last minute that they were reneging on the deal and had used the time to produce software for their OS instead. There was not time for IBM to dig up an alternate producer for user software. Then to complicate the future of OS/2, some of the copyrights were shared with M$ so that added its own set of problems. So it wasn't just a 'mysterious' shortage of applications that hindered the technologically superior OS/2, it was getting stabbed in the back by Bill which created much of that barrier.
Then there was the vaporware issue. M$ gave the impression, and the trade magazines helped spread the rumor, that M$ was almost ready with something comparable and that everyone should hold off. That also hurt OS/2 since M$ didn't have anything for years and it kept people from buying OS/2. Also, though I'm not sure of the timing, M$ was probably still at the time getting the illegal per-processor fees for all microcomputers sold.
Money is not free speech. Elections should not be auctions.
(Score: 1, Informative) by Anonymous Coward on Tuesday August 25 2015, @08:29PM
Back when the courts had teeth to enforce antitrust remedies
They have always had them; since Reagan, however, they have chosen to use them sparingly.
The revolving door doesn't work for DoJ employees if they make life difficult for potential future employers.
IBM was prohibited from making both the hardware and software
No. The fact that IBM eventually produced and marketed OS/2 by themselves shows this to be false.
IBM was forbidden from -excluding- competitors to their own software.
M$ appeared to be a convenience to Big Blue. Heh.
Going back further:
M$'s DOS 1.0 was offered at $60; IBM priced Digital Research's CP/M-86 at $240.
For the average consumer, the effect was the same as if IBM had produced its own software:
From the start, the IBM PC had a single supplier of its OS.
This was further solidified by whitebox clone vendors and M$'s extremely non-linear bulk pricing.
IBM had a deal with M$
...where IBM handed over the keys to its kingdom.
The employees of Boca Raton facility that produced the PC had no clue about the existing personal computer industry.
It seems pretty clear they didn't read the publications pertinent to that field and weren't aware of who all was producing stuff--specifically software--within that nascent market.
...and IBM's final brilliant move was to build PS/2 boxes incorporating their not-backwards-compatible MicroChannel Architecture bus.
Everybody else in the PC-compatible supply chain went with EISA (Extended ISA, a $0 "intellectual property" alternative that -was- backwards-compatible) and offered lower cost stuff that consumers swooped on.
PCI, another standard, came along soon afterwards.
MCA was a loser.
Having never even tried to get into the desktop software business, it wasn't so long before IBM was out of the desktop hardware business too.
M$ told IBM at the last minute that they were reneging on the deal
Finding a M$ partner that does NOT have a knife in its back requires quite an effort.
and had used the time to produce software for their OS instead
...with "their OS" being a GUI that rode on top of DOS for several years.
-- gewg_