Measurement In The handPAM Environment

First it seemed that structure merely offered humans a way of getting their hands on the image. But now it turns out that in thus getting their hands on the image, they leave 'tracks'. Given sufficient structural agility, users will create groupings wherever they need them. The structure left behind reflects their grappling with the image to manipulate it for their cognitive purposes. Indeed, the structure offers a shared framework of orientation (for user and handPAM) precisely because it has been formed through purposeful manipulation of the image.

To take full advantage of these structures, we must measure them. Measuring involves: first, recording the structural dynamics of a user session; and second, attempting to determine which of those dynamics reflect the cognitive groupings of the user.

The handPAM environment facilitates this task in a semi-Procrustrean fashion (figure 8). Since the human needs an artifact to manipulate text and graphics anyway, why not supply one which makes for easier measuring of the use of the artifact? The paramount task in measurement is to define the units of text-graphic activty      the temporal and spatio-visual chunks which are meaningful to the user. A recording of the user session in these units is a history. A history of a user session can be helpful to both the user and researchers. For the user, it provides an InterLISP style history of events; for researchers, a corpus of visual linguistic activity for analysis.

This 'chunkwise' history list provides cognitive psychologists with a tool for exploring the role of graphics and external memory in problem solving. As small insights are garnered toward this end, then they can be applied to improving the system's representation of what the user is up to, and supplying assistance based on this knowledge.