The death of the desktop OS greatly exaggerated

For almost as long as people have said this year will be the year of the Linux desktop, people have decried the end of the desktop OS itself.
Sorry Google, not so fast.
The desktop OS isn't going anywhere.

The argument goes something like this:
1. More and more people are using their web browser for applications.
2. The web is producing progressively richer and richer applications.
3. Thus eventually the desktop will be pointless
4. Thus people will stop using it.

According to the proponents, we're all soon going to be using dumb terminals, getting our data and applications though the cloud. The personal desktop is a thing of the past. Armageddon for private computing. We will be beholden to our service providers.

This is all failing to recognize one basic fact. Moore's Law outpaces Neilson's Law. You may be asking yourself what these laws are, or if your tech savvy maybe just Neilson's Law.

Here they are:
Moore's Law - Computer power will double every 18 months. (60% per year)
Neilson's Law - Network bandwidth will double every 21 months. (50% per year)

Note: These laws are based upon statistical analysis of trends. They have been very accurate so far.

What does that mean to you as the end user? What do these numbers mean? Suppose you took out a loan for one thousand dollar at 50% interest, and then lent that thousand to someone at 60% interest. 10 years later the debtor decided he's ready to pay you back, so you pay off your creditor. How much money do you have? $51,000!
(Note that this is way better than 10% over 10 years which only yields $2,600)

That should give you a sense of how much faster your computer power grows relative to your bandwidth.

This alone should put the entire idea that the webtop will replace the desktop to bed for good. Not to say there wont be a place for webtops; or as they used to be called "thin clients". A business can get a whole host of cheap slow computers, and they can talk to a server for their computing needs. But for most people the desktop will reign supreme.

I think people imagine a webtop only world largely because they can't imagine what they would do with all the extra computing power. What could they possibly want that isn't available on the web.? In other words the premise is based upon a failure in imagination.

To address this I will give my predictions for where computing will lead.

Next 10 years: (Desktop 109 time more powerful, internet 57 time more powerful)
Over the next ten years the desktop operating system is going to rely less on the browser than before. Nearly every application is going to start shipping "web enabled". Look at the new office product, look at the new KDE plasmoids, look at the next version of Ubuntu. Everything from the Kernel to the graphics on your screen is going to be designed with the concept in mind that your computer is going to be connected to a web service. Some of these you'll pay for because it's just that good. Some of these will support themselves with advertisments, because its the only way to compete. Some of these will be community driven totally free services. And some of these will simply be communicating with your neighboring laptops in the coffee shop.
Over the next ten years the desktop will become semantic. Providing you with powerrful tools to get information instantly from your computer. This will synchronize with the web. Quoting the bible will be keystrokes away, files will be easy to find, applications will launch with keystrokes. Forget about waiting for applications to launch though, most of them will be ready and waiting as services/daemons on your PC.
Over the next ten years rich computing applications will find their way into every electronic device. Your television, your phone, your alarm clock. They'll all be internet ready, they'll all start learning how to communicate with eachother.
Over the next ten years more of the applications that have been made by software giants will be available from hobbyists, because 100 times the computing power in an office application just doesn't help that much. Developer will be able to produce better, more reliable, higher quality equivalents to many applications. Not that the OpenOffice.org or MS Word are going anywhere, but you'll hardly notice if your editing you document from a cafe terminal, or your home PC.
Software giants will need to turn to more innovative products to survive. Applications that really change the way we view what computers can do for us.

Next 20 years: (Desktop 12,000 time more powerful, internet 3,000 time more powerful)
The next twenty years will see an inversion of the desktop server model. Peoples PC's will begin to offer services to servers. Your desktop applications will start offering to give processing power, and semantic poewr to servers around the world. The content you make on your computer will become a tile in a giant mosaic. Some of these services you provide will be the cost of the now "free" software. Some of these services will pay you. You'll have greater flexibility in what you can provide to people. Privacy concerns will become an international issue, intelectual property laws will radically change. You will also gain the benifit of the information ubiquity.
Your computer will begin to predict your behavior. Preparing to open applications based on usage. Responding to your temperment. Altering their interface base upon the activity your engaging in. The desktop will provide more services that used to be handled by applications. You'll have new ways of interacting with your computer (voice, eye movement, touch).
The difference between what is on the web, and what is on your computer will begin to fade. Rarely will people ever interact with thier filesystem, they will have the files they need by the context of their actions. The separation between the computer and the web will begin to dissappear, and the files one needs will be available on demand.
There will be programming languages that allow almost anyone to write simple applications, and application will begin to become customizable based upon users describing what they want. Designers will be able to quickly prototype the applications they want, and have functional applications ready to give to developers.
Application design will become more user friendly by incorporating a narative process into their user interface desgin. The metaphores of computers will continue to resemble real things more and more. What is now considered common anoyances, will begin to be things for "real techies" People will wonder how windows XP was ever considered "easy".
Users will begin to have so much control over their PC's that vendor lock in for applications will be effectively impossible. Computers will be very personal items, and hard to share with others directly because they will reflect a users personal narative. Services will be the way software companies make money.

Next 30 years: (Desktop 1,400,000 time more powerful, internet 200,000 time more powerful)
I wont even venture to guess here. Here we find ourselves in the realm of science fiction. I will say that computers will be everywhere. The national parks will be lit by wireless. Businesses we walk past will know our name. We'll be in a world where everything is in a constant dance of communication. Perhaps we'll begin to have robot's that act as customer service. The modern internet will be as archaic as the telephone. Who knows? I just try and imagine my laptop as a supercomputer, and I am stunned. When I write a aritificial neural network right now, it trains sometimes in a day. In thrity years the same problem will be solved by my computer in a fraction of a second. My laptop in 30 years will be able to solve that problem with more ease than my current laptop displays a single character of this essay on this screen.

I must point out that I am pontificating here. I don't know, this is just a guess. But I will point out that it is practically impossible given what we now about computing trends that the desktop operating system will dissapear anymore than computing itelf.