Per this thread, Apple/WebKit's SVG Transforms are NOT competing or in the same area of interest as O3D. Instead, SVG Transforms (which incorporates WebKit's CSS 2D and 3D Transforms, CSS Transitions and CSS Animations) is meant to display and interact with flat pieces of web content (in this case, the traditionally-2D SVG) in 2.5D/3D space, while O3D and WebGL are meant to display fully 2.5D/3D *scene graphs* and *models* separately from any web content (even though both efforts are aiming for an in-browser user experience).

Interestingly, Maciej mentions that it would be possible to codify a means to join arbitrary (2D/3D) web content and arbitrary 3D models together, but that it may be "a bigger challenge than anything that anyone has done so far".

So if this same argument could also be applied to other related initiatives in both fields, from canvas(3D) to X3D, then I'm assuming that the next frontier for the 3D Web initiatives to cross is how to bridge the divide of perception between 3D and 2D in the same network-centric application. Certainly, the fact that most of us do not have 3D-ready navigation hardware (like the SpaceNavigator) is a core part of that dialogue over why the 3D Web initiatives are not scaling to better expectations of ease of use and accessibility (not to mention hardware acceleration of graphics).

So O3D or WebGL may have a future in the web browser (or at least a better one than VRML and X3D), but its not like we're that much closer to bridging the gap between 3D web content and 3D scene content.

Stereopsis is an effect that allows for the eyes + brain to perceive depth, and is utilized in an optical illusion that, when equipped with the right method of visual perception, allows for the viewer to perceive a sense of depth in an otherwise flat surface area.

Many methods have been devised since the 19th century to give off an illusion of depth in a media-dedicated flat surface area (a technique known as stereoscopy), ranging from "wiggle stereoscopy" to the manual side-by-side-so-cross-your-eyes stereoscopy to the iconic anaglyph image + polarized "red and blue" glasses.

Most recently, a few folks on YouTube took to demonstrating the combination of anaglyph imagery and polarized 3D glasses with MIT graduate Johnny Chung Lee's WiiMote headtracking method (resulting in hundreds of demo videos on YouTube from inspired users) to demostrate how such a combination of anaglyph imagery with headtracking might work.

Sadly, I'm not convinced by the look of the videos. Plus, after thinking it through, I 've started to think that it seems redundant to combine digital CGI anaglyphs with headtracking polarized glasses.

I mean, if you demonstrate Compiz Fusion with its Wiimote-recognition plugin for headtracking separately from Compiz Fusion with its anaglyph filter plugin for polarized glasses, you get a fairly stereoscopic view from either perception; in fact, the Wiimote plugin for headtracking seems to possess much further potential range depending upon the range of motion that the viewer has around the screen, while the Wiimote anaglyph plugin for polarized glasses gives the classic feel of being able to almost reach inside the imagery on the screen.

The drawbacks, therefore, may dedicate either method of perception to particular media and particular audiences:

• the Wiimote headtracking method may be best for visual environments - preferably 2.5D - which concurrently allow for manipulation of the environment's contents through interactive controls (like games)
• the anaglyph method may be best for visual environments - namely 2D - which don't allow for manipulation of the environment's contents through interactive controls (like movies).

I'll expand upon this a bit later, but here are the two demos posted to YouTube: