Originally posted by Maeglamor
[color=66ff00]I've seen one or two of them, none are implemented to the standard that was exibited by the SUN system and are usually fairly akward to get around in. Can you cite any examples offhand that you think are especially relevant? I found most of them to be poorly thought out and often quite buggy.
[/color]
That would be my point precisely. There's the two that SUN has worked on, Microsofts Gallery, no less then six or seven open source projects, the CAVE system, etc. All of them fail at their task because they try to map a 2d interface into 3d, but either keep the same controls (keyboard and mouse) or require truly awkward actions (standing in the middle of a series of screens with body tracking). To top it off, the interface remains--at the interaction level--2d.
[color=66ff00]That's not really a good argument, claiming that a GUI and all of the complexities therein can be compared to a game is a bit of a stretch. No game has ever had to use 3D space to handle anywhere near the amount of onscreen information conveyed by a desktop GUI (except perhaps one written by DSmart and we all know how great those are. 
(and are in actual fact 2D) ).[/color]
No, no game ever has, but looking at the average user's desktop, there's really not a hell of a lot of data going on. Its mostly buttons, empty space, and a small area of interaction. 3d is actually going to be faster than 2d here, just from the accelleration angle. I reference you back to Descentrace. His description is almost exactly the right way to use 3d concepts in a 2d desktop environment (and very likely what Vista will be doing).
[color=66ff00]
Using a 3D card to render a 2D interface does not amount to the same thing as using a 3D interface either. That's like saying that 5.1 is exactly the same as stereo simply because they both can utilise a sound processor. The important feature here is to make the desktop more than a layered 2D world and thus allow the kind of skewing, rotating and stacking of windows evident in SUN's system. [/color]
No, and I never said it was. I, in fact, have kept the discussion of 3d interfaces and 3d-accellerated 2d interfaces seperate. To respond to the second half of that, I have to ask one thing: what does a rotated or skewed window gain you in the interface? You'll still be using 2d desktop concepts for everything. You'll have windows over windows (the UI as its stands does that), windows side by side (yep), etc. Tossing a window off into 3d space to let it hang about where you can't see it is just fancy eye-candy for minimising it. You haven't gained anything.
Exercise: list all the things you would be able to do with a 3d interface. For my part, I'll show you how each and every one of them maps 1:1 to a 2d desktop interface concept.
[color=66ff00]As for the mouse and keyboard argument; we haven't seen a device that can really replace one or either, at some point speech may replace the keyboard but only to a certain extent. A lot of research has gone into 'look and click' interfaces i.e. a camera tracks where you're looking on the screen and uses this to plot where the pointer goes. Perhaps one or both of these systems may supercede the keyboard or mouse which have lasted so long due to their simplicity?
[/color] [/B]
Look-and-click devices are just fancy mice. They're still, essentially 2d: you're interacting with an XY cartesian plane, and you are NOT doing anything in the Z direction. To do so would require being able to sense a person's depth of focus at the exact moment (possible, true), but again I ask, what does it gain you? I can already have windows behind windows and I don't need either fancy hardware or fancy interfaces to handle that.
