Buh, this forum needs the 'edit' button enabled, I always think of more stuff to add after posting, I tend to think of forum posts as miniwikies now days. Anyway:
Depth would also allow for browsers inside of 3D environments to have content popup in the 3D world, such as videos games or SecondLife kind of stuff rather than just rendering a flat 2D surface. That's something that would be much harder to do with the full Web3D (if at all possible). You might not be able to get something 3D spinning cubes (unless you did manage to tap the Web3D canvas into the full 3D environment, or added depth to the embedded SVG), but text and buttons that hover above the wall would be doable.
3D user interfaces (once again starting to popup in video games like Dead Space).
Actual implementation would need a little though, you would probably have to give each depth level a texture and if you have the layers changing in 3D at runtime they could split and merge but provided people aren't too crazy with the number of different levels it shouldn't be a problem.
What coordinate system to use might need some thinking too, there's no pixels for depth, if you use fixed units, say millimetres, different systems are capable of different levels of depth and would have much different impact even if they both have the same capabilites, so a mobile phone will be different to a home cinema. An embedded 3d browser will be able to handle basically infinite depth within its virtual world, that also rules out percentage levels. Maybe an attempt to extrapolate based on the screen 2D parameters (ie 1.0 is the same depth back as the width of the screen).
You can see the Dead Space UI for an example of the kind of thing I'm thinking of, I believe it was done in an embedded flash, it also has the rotation too making it a bit more than a flat UI:
http://www.youtube.com/watch?v=xLF0DjFF ... re=related
Understand I'm talking about adding depth to normal 2D content, not trying to project full 3D models and such.