ArsTechnica recently re-published an article from 2013 that covered "the triumph and tragedy of OS/2", IBM's long-forgotten consumer desktop operating system. At six pages, it's certainly one of the longer pieces not written by John Siracusa and covers a time in my childhood when there appeared to be a great deal more drama and excitement in the world of software1. Microsoft was a scrappy startup and IBM was the monolith that made covetable notebooks, powerful mainframes, and really expensive software. While I never used OS/2 Warp in any real capacity, the operating system left quite a mark on me.
In the mid-90s, a friend of mine and I operated a little business where we would help people with their computer issues. The venture was called SultanSoft. We had delusions of grandeur and, when not helping someone with a DOS installation or upgrading some hardware, we could be found sitting in my room in front of my first computer, building applications that very few people would ever use. One of the most frustrating jobs we had was trying to get someone's relatively new computer to run OS/2 Warp with some semblance of performance. Windows 3.1 ran just fine, but OS/2 was required.
Everything about the job was complicated. Installing the operating system took far longer than expected. Configuring the "lightning fast" 28.8kbps modem was a pain. Discovering that the client bought the version of OS/2 that didn't ship with the Windows emulator — which he needed — was disappointing. But the most frustrating thing of all was the lag. Everything just took so long on IBM's operating system that we would often wonder if there was something wrong with the hardware. A company like IBM would never release bad software … right?
We tried everything we could think of to get the machine to perform better. At one point, we even tried upgrading the machine, an Aptiva with 2MB of memory, to a ridiculous 4MB RAM! The 2-megabyte, 30-pin SIMM had to be ordered directly from IBM via an authorised service centre and cost just under $70 before tax. When we put the memory into the machine and booted it up for the first time, we thought for sure the 66MHz 386 machine would take off like a rocket.
But it didn't. In fact, we couldn't see any real performance difference despite doubling the RAM. The machine used SCSI hard drives, which were considered the fastest consumer drives back then. It had a dedicated video card with a good amount of memory. Heck, it had a CD-ROM! The machine should have been amazing, but it wasn't. In the end, we were asked to replace OS/2 with Windows 3.11. We completed the task the same day and, because the computer had a whopping 4MB of memory, the system just flew. The IBM Aptiva could handle everything that was thrown at it.
After this unfortunate interaction with the operating system, I never tried to use it again. The world was already rallying behind Windows and I wouldn't even see an OS/2 interface until 2009 when I would visit a company in rural Japan that used the system to control their factory machinery.
Most people in their 40s and older will probably have no recollection of IBM's operating system, which is unfortunate. There were a lot of innovative features that were eventually taken and incorporated into other systems, such as multiple desktops, virtual environments, and the like. Fortunately good ideas have a way of sticking around2
Well … from my vantage point, as a teen with a stack of computer magazines and an insatiable appetite for anything and everything related to software and hardware.
Bad ideas also have a way of sticking around, but they generally require a great deal more work to maintain.