HCI is a lazy subject. Too often it stumbles upon new topics well after industry has moved on. For example, much of HCI was still focussing on text editors and window systems when commercial developers were producing multimedia authoring tools. HCI also has a tendancy to develop orthodoxies. For example, I believe that the notions of `task' and `context' have done much to obscure innovation within the field.
This talk will provide two concrete examples of the malaise that is affecting HCI. Firstly, I will argue that few interface design principles can be constructively applied to guide the development of pseudo-3D systems. I will illustrate this argument by showing a range of VRML (Virtual Reality Mark-up Language) and QuickTimeVR applications that I have helped to develop over the last two years.
Secondly, I will argue that designers cannot rely upon concepts such as `task' and `context' to assess the value or utility of these new media. I will illustrate this argument by describing a number of studies that have failed to distinguish 2D and pseudo-3D techniques in terms of task performance in several different domains or contexts. These results are problematic because they fail to explain the profound differences that users report when interacting with this new generation of pseudo-3D interfaces.
I think that some of our findings can be explained in terms of Gibson's early work on the perception of pictures. By extending his theory to pseudo-3D systems, I believe that we will be better equipped to exploit the techniques provided by VRML and QuickTimeVR.
I also hope to persuade the audience to look again at some of the underlying user characteristics that transcend `task' and `context'.
For more information about this talk please contact johnson@dcs.gla.ac.uk