There would be no need for an hour glass or spinning pinwheel in my OS, because the user would never have to wait for anything. (The hourglass icon was originally an apology for poor hardware performance back in the old days; that we're still seeing it in a world of 2GB of RAM and dual-core processors indicates that something has gone very wrong with operating system development.)
The new OS would target the emerging netbook market too, and the goal would be to create a operating system geared towards portable computers. You might not have realized but the traditional desktop computer is dying. The only people using desktop computers nowadays are gamers and office workers. Nearly everybody uses a laptop nowadays, even if it never actually leaves their house, or even shifts from their desk.
An important point about the new operating system would be its philosophy. The goal would not be to create the best operating system in the world. It would not be to create the most advanced operating system, or the most innovative, or the most technically accomplished. The goal would be to create an intuitive and "good enough" operating system that most people can use without training, or wincing when something doesn't work as they anticipate. The goal would be to ensure the features people expect are present, and that they're at their fingertips.
As exciting as new operating system features are, they just aren't needed or desired by most people. With Windows XP, and Mac OS X 10.4, and recent versions of Ubuntu (I'd cite 8.04 as an example), we have reached a pinnacle of operating system development. Things have got as good as they can get. Any new features from now on will just get in the way.
Of course, all of this is only a thought experiment. If I did win the lottery (and I don't actually play, so that's even more unlikely than usual), I wouldn't be so foolish as to create a new OS.
For starters, I'd probably be sued into oblivion by Apple and Microsoft. Desktop operating systems in particular appear to be a minefield of software patents (although I wonder if I could get around this by basing development here in Europe, where I live, and where software patents simply don't exist).
However, the biggest issue is that operating systems simply don't matter any longer. They're very much a 90s thing. The 90s were about exploring Aliceâ€™s living room. The noughties are about what happens when we step through the looking glass.
What matters now is online, and what you can do therein. If I had any sense, I'd invest my millions in creating online applications, and trying to bring open source and open standards to that particular world--a world that appears, right now, to be almost exclusively proprietary.
Keir Thomas is the award-winning author of several books on Ubuntu, including
Ubuntu Pocket Guide and Reference.