Different UIs for different cases
Let's start with the good news first. The desktop metaphor does not always need a replacement. It will still serve us well in a very specific user-base. These are the people who are indeed having a computer on their real desk in the office. They are actually using the computer for their office needs and they are quite happy with what they have. This is natural, since the desktop metaphor is targeted at them. They can easily do what they need with Windows/Office or GNOME/OpenOffice. Any User Interface which will not resemble at least on a basic level the desktop metaphor will make them unhappy if not angry.
Another area where the desktop metaphor will not be easily replaced because of inertia, is when the users don't spend enough time interacting with it in order to observe its flaws. This includes gamers and users of vertical market applications. A hard-core gamer only comes in contact with the desktop metaphor for 2 minutes before his/her favourite game "takes over" the computer (keyboard, screen, mouse). The Operating System is simply a portal to a big collection of games. Each icon represents a game and double-clicking on it makes the desktop metaphor irrelevant since the full screen is now managed by the game. These users would care less with the User Interface of their OS so they won't be directly affected by any radical changes.
Computers which are used for special tasks can also hide the inefficiencies of the desktop metaphor. Users who spend most of their time with special tasks (like Web development, Computer Generated Graphics, Desktop Publishing, Computer Aided Design, Hardware and Software Development etc.) do not really see the OS of the machine. For them the UI metaphor is Visual Studio/Maya/Photoshop/Dreamweaver... (insert vertical market application here). Their problems with User Interface will certainly have to do with the application they are constantly using, rather than the UI their OS is presenting on the screen. It is up to the companies who produce the respective applications to take into account user complaints and present to them the perfect user interface with each new version of their products. The OS has nothing to do with the metaphor used by the application which can by anything that suits the target group of the its users.
So this leaves us with the remaining user base. These are the people who use their computers for multiple purposes at their homes (and not for office use). They will listen to music, surf the web, write a letter, play a game, send an email, etc. They will organize their photos, use the Internet for information gathering, and in general spend most of their time interacting with the OS instead of specialized applications. We have already described these users in the previous article. Some of them are so young (12+) so that the desktop metaphor means nothing to them. They have never worked in an office in the first place!
We note the fact that this user base is ever increasing. Computers are reaching a broader audience each year. Wireless networks, digital cameras, and personal digital stereos make computers appear in many homes where the users need a User Interface NOT resembling the desktop metaphor. Still these users are using the same OS and UI which has stayed mostly the same for the last 10 years and is clearly aimed at office workers. Is this logical? Can we do something about it?
Unfortunately these users interact each day with the desktop metaphor considering it to be THE User Interface. They do not realize that this is just one possible method to use their computer. They think it is the only one. Companies on the other hand can sell their products as it is, so they see no need for a different metaphor. Why should they after all? Every day is business as usual...
The driving force behind this article is the fact that the latest buzz in user interfaces is to use 3D and transparency everywhere. Many users are excited and feel that the way they use their computer is going to change in some magic way. They cannot understand that eye candy is completely different from usability improvement.