Username or EmailPassword
So you're saying the system architecture of Gnome 2.x is the ultimate possible, and will never need to change? Sorry, but if you want to make something better in a fundamental way, you have to break it first. Either that or create a lot of compatibility layers to keep the old API, and only Microsoft has the resources to do that.
That's not exactly what I'm saying. GNOME will replace components of the platform, but without breaking the entire stack while they're at it. In fact, it's been doing that for a long while now (see Project Ridley, D-Bus, gvfs).
The D-Bus adoption is a perfect example. GNOME adopted D-Bus without breaking anything.
gnome-vfs is not perfect and will be replaced by gvfs, which is being developed. But GNOME will not depend on gvfs until it's ready; that's the main difference in strategy between the current development model of KDE and GNOME. KDE decided to fix its entire platform in one go, while GNOME is incrementally replacing components as the next generation ones are completed and ready to go.
Of course, supporting gnome-vfs and gvfs, ORBit/Bonobo and D-Bus, etc, is not a lifetime solution, which is why the old libraries will eventually become unsupported and the platform released as version 3.0.
Btw, GTK+ adopted Cairo without having to break its API, but that's because GTK+ was well designed to begin with.