Although the article mentions the topic, it has such a broad sweep that it obscures the real takeaway, which is the way that 3d is employed. All the stuff about the incredibly detailed CG, backstory etc. is valid and interesting, but epic levels of such are not really new - Star Wars has a very detailed alternative universe, and any Tolkien nerd above a certain age can remember the vast level of supplementary data on Middle Earth that existed before the movies (Tolkien was a philologist by profession so he had fully developed multiple languages etc. to invest his imaginary world with a sense of realism).

The real kicker (from studio and filmmaker's point of view) is how Cameron is employing 3d. On most films until now it's been a novelty, used to make a weapon or explosion seem to imperil the audience, or to abuse perspective for comic effect. Cameron's achievement is to integrate it tightly with the existing cinematic language of the lens.

Basically lenses can be 'wide' or (confusingly) 'long'. A wide-angle lens presents a wide field of view while exaggerating perspective, while a long lens isolates small detail and can reduce the sense depth. Obviously this is important for framing, but it can be used in other ways; if you wish to increase tension, shoot someone running towards the camera with a long lens. S/he will hardly seem to be moving forward at all, and as viewers we fear the character will be unable to reach their objective in time. Conversely, the same motion when shot with a wide lens will suggest rapid movement from background to foreground, giving us a sense of power an inevitability (or conversely, of weakness and impotence if a character is falling away from the camera). Obviously I'm just scratching the surface here, but the various distortions of perspective possible with different focal lengths are a big part of what gives cinema its dramatic power.

What Cameron has been doing with Avatar is to shoot in deep focus (no using the aperture and focus controls to blur out the background, a favorite technique for isolating the subject from the environment) but instead create depth by altering the angle between the two lenses dynamically, creating the illusion of a large space in which attention to depth is focused stereoscopically. Until now most 3d projects have kept the stereoscopic distance fixed, which yields the feeling of watching the story take place on a stage in front of one and occasionally having one of the props or actors protrude outwards toward the audience. By varying the angle between the lenses in the same fashion as our eyes, Cameron presents a far more immersive way of experiencing the third dimension.

Much of the skepticism towards Avatar trailers and so forth stems from the fact that the background is often fully in focus - a technique which has been used to great cinematic effect (eg in Citizen Kane) but which has fallen out of fashion over the years. In videogames, by contrast, such deep focus is common (since you don't know where the player will want to focus attention in advance, and also because simulating narrow depth of field dynamically is computationally expensive). This deep focus is necessary to provide a credible sense of depth with variable-angle stereoscopy (if we threw the background out of focus too much, the foreground characters would just seem to be floating in space), but when you see it projected in regular 2d it looks old-fashioned and videogame-y.

It's not the CG that looks weird; if you shot an ordinary scene with the same 3d technique it would look boring and flat in 2d (indeed, the difficulty of emphasizing depth by adjusting focus on a consumer video camera is a big part of what makes it 'look like video'). So the primary reason Avatar is a Big Deal for Hollywood is that Cameron seems to have succeeded in developing a 3d photographic technique that is much more compelling and realistic than the standard fixed-angle 3d which has been used until now, which actually emphasizes the separation of the audience from the action. And in order to fully appreciate this...yes, you'll need to go to the theater.

Cameron has also done other amazing things - being able to integrate the CG with the live action at production time (instead of imagining it and then integrating the two image sources over months or years of post-production work) is also a very big deal, but that's much more 'inside baseball' as the main beneficiaries are the filmmakers rather than the audience, although it will lead to better results insofar as it will make it easier for less skilled or funded production teams to produce convincing blends of real and artificial images. And there are a host of other technical innovations - there will be multiple books written for the professional market on how to improve a film by employing Cameron's techniques; indeed, I dearly wish Cameron himself would write a book (or several) on his approach to directing, since he is the master of staging and editing complex visual action - if you need proof of this, watch a long action sequence from Terminator 2 (or any of his other films actually) and then a similar action scene from Michael Bay or someone similar; he too has access to huge budgets and armies of technical professionals, but in films like Transformers audiences can literally lose track of what's going on, who's winning a battle etc.

But the bottom line is that Avatar will live (or die, but probably live) by virtue of the way Cameron leverages 3d to put the audience 'in' the visual story, which is a fundamentally different photographic technique from that employed up to now.