Changes

The central interactive model for this project is the idea of the user administering tests to the computer--an exchange suggesting the need to diagnose the computer (and through it, the test) to discover its underlying pathology, and prompting the viewer to form some sense of the parameters and limitations of testing tools. Multiple-choice tests are closed systems with finite possible outcomes and descriptive states--and as such they are already essentailly "machines" for producing diagnoses, and thus are ripe for interface and exploitation by statistical or learning machines. Computer vision (CV) and statistical machine learning technologies are the means to orchestrate this encounter between the viewer and the test and facilitate the ongoing interaction between the two.

*technical possibilities:

**optical character recognition / computer vision interface. (natural interface)