Just read this blog post, containing an interview with a guy named Robert Rice. While much of it is a bit over the top self-promotion, pushing his "new vision" (read: things he's going to be trying to sell this year or next), it's refreshing to at least read commentary by someone who gets that AR isn't really about "putting graphics in the world", but about the what, who, why and how. Or, as he puts it:
it is inherently about who YOU are, WHERE you are, WHAT you are doing, WHAT is around you, etc.
Of course, he then uses this to irrationally broaden the definition of AR to dismiss the idea that it includes putting graphics in the world. As those who know me would expect, this irritates me for a few reasons. First, it makes "discussion" impossible if everyone just goes off and redefines terms that have perfectly good existing definitions, just to suit their purposes (read: lots of people want AR to be more broad because then it includes whatever it is they are working on). Second, AR drives peoples imagination precisely because of the "putting graphics in the world" aspect of it; there are lots of other terms to describe location-based, context-aware, mobile applications (e.g., ubiquitous computing, pervasive computing, mixed reality, etc) without having to try and latch your non-AR concept onto the excitement AR generates!
All that said, it sounds like these folks are going to try and actually do mobile AR, which will be very cool. Of course, it's easier said then done, but I'll be pretty excited if they manage to release something!
Until then, those of us doing AR research will continue plugging along. I'm excited by this years projects, especially the mirror worlds work we are doing, where we are going to start combining terrascale virtual worlds with mobile AR worlds! Just the kind of thing this guy is hinting at, although not on the scale he wants (we'll stick with the campus for now, thanks!)