Public Discussions > General Discussion

OS/2 Article on Ars Technica


Neil Waldhauer:
There is a long article on OS/2 in Ars Technica

It was a cloudy Seattle day in late 1980, and Bill Gates, the young chairman of a tiny company called Microsoft, had an appointment with IBM that would shape the destiny of the industry for decades to come.

He went into a room full of IBM lawyers, all dressed in immaculately tailored suits. Bill’s suit was rumpled and ill-fitting, but it didn’t matter. He wasn’t here to win a fashion competition.

Over the course of the day, a contract was worked out whereby IBM would purchase, for a one-time fee of about $80,000, perpetual rights to Gates’ MS-DOS operating system for its upcoming PC. IBM also licensed Microsoft’s BASIC programming language, all that company's other languages, and several of its fledging applications. The smart move would have been for Gates to insist on a royalty so that his company would make a small amount of money for every PC that IBM sold.

But Gates wasn’t smart. He was smarter.

In exchange for giving up perpetual royalties on MS-DOS, which would be called IBM PC-DOS, Gates insisted on retaining the rights to sell DOS to other companies. The lawyers looked at each other and smiled. Other companies? Who were they going to be? IBM was the only company making the PC. Other personal computers of the day either came with their own built-in operating system or licensed Digital Research’s CP/M, which was the established standard at the time.

Gates wasn’t thinking of the present, though. “The lesson of the computer industry, in mainframes, was that over time people built compatible machines,” Gates explained in an interview for the 1996 PBS documentary Triumph of the Nerds. As the leading manufacturer of mainframes, IBM experienced this phenomenon, but the company was always able to stay ahead of the pack by releasing new machines and relying on the power of its marketing and sales force to relegate the cloners to also-ran status.

The personal computer market, however, ended up working a little differently. PC Cloners were smaller, faster, and hungrier companies than their mainframe counterparts. They didn’t need as much startup capital to start building their own machines, especially after Phoenix and other companies did legal, clean-room, reverse-engineered implementations of the BIOS (Basic Input/Output System) that was the only proprietary chip in the IBM PC’s architecture. To make a PC clone, all you needed to do was put a Phoenix BIOS chip into your own motherboard design, design and manufacture a case, buy a power supply, keyboard, and floppy drive, and license an operating system. And Bill Gates was ready and willing to license you that operating system.

[follow the link for the rest of the story]

Eugene Tucker:
Thanks Neil, I found this story to be pretty accurate.

Not a bad article, mostly accurate but a few things I'd disagree with.  In particular

I'd strongly disagree with the claim that Windows 95 was faster, smoother, and crashed less then Warp on the same hardware.  I recall having Windows 95 and Warp 3 on the same system with boot manager and I found the opposite to be true - Warp (running on HPFS) ran smoother and almost never crashed, yet Windows 95 would frequently give me a BSOD and / or make me wait for things to happen.  But I will also point out that both systems were picky about hardware drivers.   When I had issues with Warp crashing I'd frequently find it was related to a bad driver.  I recall I had an ATI video card and an ESS-chipset sound card - both of which would be troublesome if using the stock drivers that came with OS/2.  Using the updated drivers from ATI and ESS fixed many problems (not to mention using drivers such as SIO for the serial port) .   But to be fair the stock drivers that came with Win95 for this hardware didn't work well either until the manufacturer's updated drivers were installed.  And Microsoft's stock serial port driver in Wind95 didn't work very well either.  In retrospect both OSes had issues in this area.

In my view the biggest benefit that Windows 95 had was Microsoft's bundling agreements and the per-processor license agreements which eventually became the subject of antitrust actions.  Microsoft presented Win95 as the upgrade path to Dos/Win3.1 while leveraging their OEM contracts to lock vendors into paying for licenses for every computer shipped, even computers shipped with someone else's OS (or no OS at all).  Then to kill additional competition they offered MS Office license add-ons for low costs that other software vendors couldn't compete with.  This more then anything is what killed other OSes from gaining any foothold in the x86 market. 

I'd also disagree with the claim that early versions of NT were stable and fault tolerant and did everything advertised, that isn't what I recall of my early experiences with NT.  I'd suggest that NT didn't really become the product it needed to be with regards to stability and usability until NT version 4, and even that didn't start to shine until after a service pack or two.  Although I will give props to NT 3.5 which was almost there.  Of course this is one of those things which could be debated.

I'd agree that the failure of the PPC and the PPC port of OS/2 did significant damage to OS/2.  IBM put a ton of money into that project and it was a huge failure.

Overall not a bad overview of the history.

Ian Manners:
Based on the 1996 PBS documentary Triumph of the Nerds.

This comes across as something that Microsoft's people would have done, 98% accurate with key parts changed to slant the context in an effort to rewrite history.
ie, Tell the truth but change the pivots :)

I can only speak from an Australian perspective and noticing that the author of the original article is looking at it as a Windows person claiming to be a convert from OS/2. I had no problems running OS/2 v2.11 with 4Mb of RAM either though that was in a PS/2 77i (which I still have). You have to remember that OS/2 with memory problems was something with a lot more Microsoft code in it.

I went to OS/2 because I was sick and tired of Windows blue screening while either ordering goods over the phone from suppliers, or helping customers on the phone so I grabbed the copy of OS/2 v2.11 that Gavin from IBM Australia had given me and loaded it. Back then we installed OS/2 on our office machines (IBM) because we were IBM re sellers but we had no problems putting OS/2 v2.11 on Aptiva's, non IBM/MCA PC's from Compaq, NCR, even clone PC's we had made in house. Once OS/2 Warp 3 Connect was released I never considered changing back to Windows.

OS/2 v2.11 was not perfect but it was sure a big step up from Windows 3, and do not get me started on Windows 95, it was great for making money from customers with the support but though we tried Windows 95 in the office for a couple of weeks we gave up and went back to OS/2 because it simply worked. We did keep Windows 95 on a couple of support PC's, because we needed to support customers with Windows 95.


I also did a lot of Windows 95 and 98 support back in the day, lots of trying to figure out issues that didn't work as expected, random crashes, and more.

OS/2, especially Warp 3 - sometimes would be a challenge to get it initially installed and running.  Once up and running (and after at least Fixpak 5 had been installed) it was solid. I can recall being asked one time how OS/2 was working for me.  My reply was something along the lines of "It works, it is still working.  By the way I have not rebooted due to a crash yet" (note that I did reboot due to having both Warp and Windows available via Boot Manager).

Of course I will qualify the above with the note that Fixpak 17 was a bit of a exception, while it introduced some significant features (such as the initial SIQ fix) it also caused some instabilities that had to be worked around.  Fortunately those issues were fixed in a subsequent fixpak.


[0] Message Index

Go to full version