Opinion: The top 10 operating system stinkers

Enough of the good old days! Let's talk about the bad old days of OSs instead.

I love old technology as much as the next techno-geezer, but come on, it wasn't all wonder and goodness. After we're done reminiscing about the good old days of operating systems, let's reflect on the bad old days of operating systems as well. After all, the bad times are still with us -- even in 2009, there are still some wretched operating systems out there.

In historical order, from oldest to newest, here's my own personal list of the top (bottom?) 10 OS stinkers.

OS/360, 1964

No, no, I'm not talking about the later versions of OS/360 that some of us used on IBM 360 mainframes back in the late '60s and early '70s. For its day, it was fine. Indeed, my very first operating system was an OS/360 descendant with TSO (Time Sharing Option) running on top of it.

What I'm talking about is the very first version of OS/360 -- the one that led its project manager, Fred Brooks, to write The Mythical Man-Month, his classic book on how software development fails. That first version of OS/360, to paraphrase Brooks, came in late, had flaws in its control programs, required more memory than planned, was over budget by several times the original estimate, and, oh yeah, it was slow too.

On the other hand, we did get a classic book on how not to develop software, which included such nuggets as "Adding manpower to a late software project makes it later." Brooks likes to describe it as a software developer's Bible, because "everybody reads it, but nobody does anything about it." As the rest of this tale shall reveal, he was right.

ITS (Incompatible Timesharing System), late 1960s

What can one say about an operating system written in DEC PDP-6 and PDP-10 assembly language that supported one mono-case, six-character filename ... per directory? (Yes, you read that right: Each file resided in its own separate directory.) And security was nil -- for example, no passwords were required, and you could log into anyone's active session and do pretty much anything you wanted with it.

What's amazing is that despite being an incredible pain to use and with no security whatsoever, ITS actually was an important operating system in its day. While it was eventually forced out by the rise of Unix, many programs still in use today, such as the Emacs editor and the Lisp language, got their start on ITS.

For more on ITS and the early days of computer hackers, check out Steve Levy's classic book, Hackers. You'll find it entertaining and amusing, and you'll be very glad you didn't have to use ITS.

GNU Hurd, launched in 1983, still incomplete

Ever wonder why some people refer to Linux as GNU/Linux? The official explanation is that Linux is merely an OS kernel that relies on GNU software to make a complete operating system. GNU was announced in 1983 as a future replacement operating system for Unix, to be made up entirely of free software.

But after more than 25 years in development, GNU remains incomplete: Its kernel, Hurd, has never really made it out of the starting blocks. (I'll refer to the complete OS as "GNU Hurd" to avoid confusion with other GNU software.) Almost no one has actually been able to use the OS; it's really more a set of ideas than an operating system.

And that's why I'm naming GNU Hurd as one of my top 10 stinker operating systems -- because after a quarter century, it has still failed to deliver on its promise of an entirely free Unix replacement. By incorporating ideas and software from GNU (and other sources such as Minix and BSD Unix), on the other hand, Linux has stepped in to pick up GNU Hurd's slack, providing an advanced operating system that is ready to use right now, in numerous distributions.

I, for one, am not willing to wait another 25 years for a chimera. Could we please just drop the dream of the GNU Hurd OS as an idea whose time will never come?

1 2 3 Page 1
Page 1 of 3
It’s time to break the ChatGPT habit
Shop Tech Products at Amazon