OS/2, Microsoft and Open Source

There seem to me to be important parallels between IBM’s OS/2 and open source. To explain why I want to start by recounting my take on the history of OS/2.

In the early 1980s engineers at IBM realized that the DOS memory addressing architecture would need more than band-aids like HIMEM.SYS and the idea for OS/2 was born. OS2 was to be a new desktop operating system designed from the ground up. I remember going to the OS/2 announcement meeting in 1987; although some applications (Lotus 123 for example) were written with an API graphical interface, and IBM had announced that OS/2 would have a API GUI, OS/2 would remain command line driven until October 1988.

The delay in providing the GUI is in part attributable to a difference of opinion as to what mattered most in a desktop OS. OS/2 was a joint project between IBM and Microsoft. IBM’s relationship with Microsoft dates back to 1980 and the first IBM PC; IBM needed and OS and Bill Gates promised to provide one - which he did and we had MS-DOS and IBM-DOS (which was effectively MS-DOS with an IBM logo and a nice linen-covered beige box).

The OS/2 project ran into difficulties in the late 80s as differences in design philosophy and conceptions of the most appropriate business model for a desktop OS began to surface. IBM has staffed the OS/2 project with developers from the MVS labs. MVS was IBM’s flagship mainframe OS and was built on two central pillars: reliability and recoverability. IBM’s engineers took this notion of what a good operating system should do to the OS/2 project. Because IBM saw the PC as a tool for small businesses rather than consumers or end users, reliability and recoverability were seen as key to productivity and to the marketing of the PS/2 and OS/2 products.

Microsoft had a different view. Gates had a vision of a PC on every desk. He saw the consumer and end user market as his top priority. Consequently, Microsoft focused its resources on developing a simple graphical user interface. Two projects were underway at Microsoft; the GUI for OS/2 and Microsoft’s own GUI, called Windows. These differences in approach led to a parting of the ways; Microsoft left the development of OS/2’s GUI to IBM, concentrating its efforts on Windows. When Windows 3.1 was released in 1990, it was in many ways more user-friendly than OS/2; it was easier to set up and peripherals were easier to attach. However it was far less stable and had zero recoverability.

With the parting of the ways, IBM and Microsoft each followed their own strategies. Although IBM worked hard to make OS/2 easier to setup and Microsoft to make Windows more reliable, the foundations had already been laid and the rest, as they say, is history. Gate’s business model turned out to be more appropriate (if profitability is any measure). IBM’s last release of OS/2 (Warp 4) was announced in 1996 and support for OS/2 was discontinued on December 31, 2006, just as Windows Vista was being rolled out towards general release in January 2007.

So what has all this got to do with open source?

Open source developers choose what they work on. There is high status attached to contributing to the OS kernel, so we have in GNU/Linux a very robust OS. However, I still have a hard time getting Firefox browser plug-ins to load under Linux and as a consequence, there is a fair amount of web media content for which I have to resort to Firefox for Windows ( I’m running Fedora 8, Fedora 9 and Windows 2000 under VMWare workstation on Fedora 8 ).

I’m worried that if an open source OS like GNU/Linux is going to move out of the technical user market and into the non-technical desktop market, more work needs to be done on these rather mundane kinds of things that are important to the technical illiterati but not so much for coders and the like. Otherwise, just as Windows beat out OS/2 because it was easier for non-technical users to set up, maintain and enhance with add-on software, so Vista will maintain it’s lead over both desktop distributions of GNU/Linux and OSX (which is hamstrung by Apple’s insistence on tying it to its own hardware, a policy that nearly sank the company in the late 90s). And that, in my view, would be a shame.

A Great Day for America

Barak Obama’s victory last night seems to send two messages; first that race is no longer the impediment to equal opportunity and success it was only a few decades ago, and second that ordinary folk are less sure about the wisdom of unregulated markets than Chicago economists think they ought to be.

While his victory is monumental, any euphoria should be short lived, tempered by the realization that the country faces serious problems that will require bipartisanship and strong leadership to surmount: an economic recession; an underfunded social security; lack of access to health care; terrorism and religious fundamentalism; nuclear weapons in Iran, North Korea, and Pakistan. And those are just the issues that confront us today.

John McCain, in a gracious concession speech, called for all Americans to support the president elect; let’s hope that those whose candidate didn’t make it to the Oval Office will head his call.