Linux on the Desktop: "Better" Isn't Good Enough

Advocates of Linux on the desktop talk about Linux replacing Microsoft Windows, but of course, that's not the way the computer industry works.

Users don't replace what they already have. PCs, minicomputers and Unix servers didn't replace mainframes, even though the vendors of those systems kept insisting that was going to happen any day now.

Instead of replacing old technology, new technology gets deployed for new applications.

Microsoft Windows will remain the dominant operating system for the personal computer as long as the dominant PC applications are the ones we're seeing today: word processing, e-mail, spreadsheets, web browsing, and accessing enterprise applications through a client-server or browser interface.

Simply doing those things better and cheaper will not be sufficient to steal market share from Windows. IBM tried that strategy with OS/2. It was marketed as "better Windows than Windows," and it was. But it didn't get popular enough to survive.

Linux is likely to carve out a good-sized chunk of the market for client operating systems — and notice, by the way, I'm carefully not saying "PC," or "desktop" systems — but it will have to do so by hosting a new kind of application, one that is not available on Windows. This kind of application is known in computer marketing circles as a "killer app," a term used to describe an application that suddenly makes a new technology compelling. The spreadsheet was a killer app for the PC, the web was the killer app for the Internet.

What will be the killer app for desktop Linux? I don't know, precisely. This kind of thing is difficult to predict. The successful few who correctly predict the emergence of a killer app will become very wealthy indeed.

But I can tell you some broad areas where you might look for the Linux killer app:

- Modular computing: Companies like Antelope Technologies and OQO are developing handheld computers that you can plug into a dock and use as a personal computer. Plug the pocket-sized module in to a dock and you can use it with a full-sized monitor, keyboard and mouse.

A company called Synosphere is working on similar technology called a Blue Dock, that can connect an existing Palm or Pocket PC system to a monitor, keyboard and mouse. It is designed to allow users to use their handheld computer as their main PC.

These technologies solve problems that are familiar to any user of handheld computers. Handheld users are perpetually struggling to get data and applications synchronized between the desktop and handheld. It's quite frustrating to be out in the field, start up your handheld computer, and realize the information you're looking for is on the PC in your office - you forgot to download it.

Moreover, handheld computers are expensive — a top-of-the-line handheld can cost $600 or more, more than half the price of a low-end notebook computer. So why not buy a handheld that can also function as a low-powered desktop system, and get the best of both worlds?

Right now, modular computing technologies are focusing on existing, proprietary operating systems, but these markets are immature. Linux already has strong handheld computing technology, in the Sharp Zaurus, and it's getting stronger. This week, we bring you a story about Lycoris, which introduced a version of the KDE desktop for handheld devices.

The Tablet PC is another form factor that looks intriguing for the next generation of personal computing. Right now, Microsoft is pioneering that technology, but a smart Linux vendor could easily steal Microsoft's small lead in this technology.

And we're seeing new technologies for e-ink and roll-up displays that will lead to still more, new and exciting form factors over the course of the next decade. Linux could become the dominant platform for any or all of these devices.

- Smart appliances. Discussion of smart appliances have focused on refrigerators that know when their contents are about to run out and go bad. I already have devices to tell me that, these devices are called my "eyes" and "nose." Be nice if my refrigerator would know when parts are about to wear out, though, and could call for help over the Internet, and then nice Mr. Refrigerator Repair Man would phone me up and see when it would be convenient to fix the failing refrigerator before it broke down while we were away for a week and I came home two days before my wife and found that the house smelled like "Night of the Living Dead" looked.

The preceding is, of course, only a hypothetical example.

(By the way, if you are throwing away potentially-spoiled bottles of salad dressing and ketchup and mustard and things, you will have the option of either throwing them into a plain plastic garbage bag, or a hard-sided and sturdy garbage receptacle. Which option you choose depends on whether you want to have a leaky garbage bag spreading salad dressing all over your kitchen floor. I just thought you might like to know that.)

- Peer-to-peer media. Right now, entertainment companies are hung up on distributing content with the boat anchor of digital rights management strapped to its ankles. Copy protection didn't work when software companies tried it in the late 1980s, it's not going to work any better now. Consumers don't want it. To paraphrase writer, blogger, and net-personality Cory Doctorow: Nobody wakes up in the morning and says, "Today, I want to buy something that enables me to do less with my music and movies."

The successful media companies will be the ones that figure out a way to make money off of unprotected peer-to-peer file sharing, rather than struggling to suppress it. The behavior we call piracy today, and which we prosecute in the court, will in the future be a legitimate consumer activity, which people pay to do.

We have all the technology ingredients in place already for the peer-to-peer media revolution: we have high-speed data connections into homes and businesses, we have PCs, we have MP3 players, we have personal video recorders. The world is waiting for a business model that can make it all work.

Linux is already poised to become a key technology in this new medium. Heard of TiVo? It runs Linux.

- New user interface: When the windows-mouse-keyboard interface was invented, Lyndon Johnson was still in office and the Beatles were still together. I expect there's something better. I have no idea what it is — it's not voice, it's not 3D virtual-reality gimmicks — but I expect there's something better. When it's invented, I expect it will run on Linux before it runs on Windows.

- Linux might replace Windows after all.

I could be wrong in my prediction that Linux won't replace Windows. When I told my co-editor Scot Finnie about the rule that new technology doesn't replace old technology, he responded that Windows 3.x did, indeed, replace DOS. I responded that Windows 3.x was simply the next version of DOS, from the same vendor with a GUI on top, but maybe he's right and it was a case that the new technology did replace the old. And maybe it will happen again.

Windows is more vulnerable now than it has been at any time in the past decade. Users are balking at the relatively high license fees being charged by Microsoft. The price of hardware has plummeted over the past decade, but software costs have set a lower limit preventing pricing from falling much further. The price of a PC is pretty low — but it could be a lot lower if each new PC didn't require a significant payment to Microsoft.

Meanwhile, the virus plague on Windows is getting worse, not better. Sure, Microsoft is doing a good job making its software more secure, but the virus-writers are staying ahead by becoming more prolific. The Internet has been in a constant onslaught from MyDoom for the past couple of weeks — until now, we could count on e-mail viruses peaking and subsiding in a few days, but MyDoom just keeps on coming.

So we may see an opportunity for Linux on the desktop after all, simply by providing a cheaper and less vulnerable alternative to Windows.