AR is a rarified technology, guarded in the crypts of academic research. While several groups have sought to remove the mantle and put AR on the street, much of the medium is still reliant on those silly goggles and half a CRAY computer in the backpack. The people at headmap seek to change that by making consumer grade glasses, with dependent processing in a jacket or backpack. Still, i'm not convinced this is going far enough. AR guru Ignacio Mondine and I were talking about this a few months ago; throwing around the possibilities for using a mobile phone in camera capture mode as a 'looking glass' for a composited layer of 3D content. Ideally the content would be positioned with GPS, or better, using an existing urban mesh-network within a 'live play' or ubiquitous gaming context. Finally someone's cracked it and the possibilites for Pervasive Gaming this opens up are enormous..

The particular branch of Augmented Reality that's interested in compositing a layer of 3D content over a real world captured scene usually works with image-recognition. So far this has only been possible using expensive Head Mounted hardware and a processor intensive application layer.
The software 'eye' is trained to recognise a unique (usually) black and white image on a card in the real world onto which the 3D model is fixed. From there the carrier of the capture device can 'move around' the opaque model in the realworld and it thus appears to be a part of the scene.

What would the Situationists have done with this technology? Is the derive already enough? One thing is sure, public domain AR will have a great influence over the way we navigate, deploy and engage with information. In this way it can't help being anything other than a psychogeographical medium.