INTRODUCTION IN OCCASIONAL PROGRESS: There always seem to be more interesting things to do than write introductions.


Despite my having long, long, long ago moved on from my early work with DOS and then with the Windows that runs on DOS, some revisiting started in mid-2021 and had taken over my new work by the end of the year.

For decades, looking at any of this again was not even a small temptation to resist. As I noted for Matters Arising From DOS Internals, I had as early as 1996 abandoned any updating of that book’s text or code either beyond Windows 95 or to take advantage of improvements in Microsoft’s 16-bit compiler. I continued to prefer the old Windows in ordinary use for a few years yet, but the future for programming lay plainly with Windows NT. For my studying of operating systems, Windows 98 got at best half-hearted attention. I never installed Windows Me even just to look at. In 2003 I discarded my last computer that ran any version of the old Windows and I don’t recall ever looking back. I kept tools for 16-bit programming for longer, installing them on a new computer in 2007 because, believe it or not, they still found use in commercial work (for code that loads from the MBR and executes before the Windows loader). I didn’t bother with them when upgrading that computer’s hard disk in 2013.

There things stood until mid-2021 when I was alerted to a generally well-made and entertaining video titled Windows’ Hidden Self Destruct Code by Nostalgia Nerd. I’d have said I’m not much given to nostalgia (despite being still a fan of pop music from the 1960s), but the video did get me reflecting on my earliest work that ever got controversial. The video’s subject dates from pre-release builds of Windows 3.1 in late 1991 but was first brought to wide attention by Andrew Schulman’s article Examining the Windows AARD Detection Code in Dr. Dobb’s Journal, September 1993. I surely wasn’t the first to notice the code’s effects but I believe I was the first to learn what it does and how. It’s no secret that I was never fully on-board with much that others wrote about it then or since. The video reminded me that there’s unfinished business on this point, should I want to take it up. I honestly believe I could have left it alone, except for a clarification here and there of what I had long ago settled on as all I needed to say about the AARD code. But then I started reading what Wikipedia has to say, not just about the AARD code but more widely on the early history of DOS and Windows and more generally of personal computing and its rapid development to mass-market use.

I have quite taken fright at how the history of DOS is distorted for having been left to writers who see it only through the prism of other operating systems. Some of this is natural enough in that practical interest in DOS continues because DOS’s vastly superior replacements indulge DOS for compatibility, e.g., to retrieve data from floppy disks that were formatted for DOS. But much too much is written because the most active interest in DOS these several decades after DOS ruled the world is a hobbyist enthusiasm for open-source projects that developed from DOS’s lesser competitors and imitators. It’s a little as if I’ve woken after decades to find that Wikipedia’s pages on milk, cheese and dairy farming got written by fans of almond milk.

The history of the Windows that runs on DOS inevitably picks up many of the problems I see in the history of DOS. And then it had its own. See this form of Windows as the bridge that the mass market was ushered across on its way to modern computing, and it’s striking that the importance of this operating system to history far exceeds the quality of technical understanding as far as might be gauged from commentary on the Internet.

What I can do about any of my concern for the quality of the historical record, I don’t yet know. It will be a while before I’ve worked it through far enough to have developed any sense even of how to organise what I might do.