The problem with any revolutionary change on the desktop is that it inevitably leads to totally inconsistent user experience and most likely many MANY new breakages. I remember when we had discussion about application menu titles. Now the “File” is a crapy name for a menu that includes “Quit.” But if you change it to some other menu (I remember a “foot” menu was suggested) meant that you would close each app on the system differently. There are many apps which say still use Motif even though Motif should have been dead 10 years ago. There are many useful apps (xfig) that are heavily used that used xaw, even though xaw should have been dead before it was created. Hence, your new model must integrate nicely with such apps. Otherwise you’re making the users life harder, not easier.

The argument that “But in 10 years all apps will use this new model” is bogus, because in 10 years, this model will be considered old and there will be a new innovation that breaks old habits.

I would disagree that linux desktop works at all satisfactorily. There are millions of little issues that should be solved before we tackle big changes that will bring yet more breakage. Now that I’m mostly a user, rather than a developer on GNOME, I see a lot more breakage. And I don’t have time to describe and work on it and try to find how to fix/improve it. That’s because I need to use GNOME, not work on it.

So I would say incremental improvement is FAR superior. I would be much happier if GNOME/Linux just worked, were consistent, even if they would look “so 90′s” and would not look so flashy like vista or macos. I tried using vista on my wife’s new computer and it does the “web desktop innovation” by making sure that every window looks like a differently designed webpage. It is no longer easy to see what you can click, and what does what. Everything has different interface, except for old legacy apps which are the only consistent thing.

Another “innovation” I would disagree with is the desktop trying to be “too smart.” Windows does this to the max. They rename things with some supposedly easy to understand term, but never explain what is going on. Hence, you mostly don’t know what the system is doing because it thinks you are too stupid to tell it what to do. Most users are not going to use the system in the way it was designed to be used. If you design an interface with a certain workflow in mind, chances are, no a single user is actually following that workflow.

Anyway too much ranting … gotta go write a final for my class … (I’m procrastinating now that no students are in my office hours) …