Abstract

In this paper we describe how we have adopted the laboratory notebook as a metaphor for interacting with computer simulation models. This 'virtual' notebook stores the simulation output and meta-data (which is used to record the scientist's interactions with the simulation). The meta-data stored consists of annotations (equivalent to marginal notes in a laboratory notebook), a history tree and a log of user interactions. The history tree structure records when in 'simulation' time, and from what starting point in the tree changes are made to the parameters by the user. Typically these changes define a new run of the simulation model (which is represented as a new branch of the history tree). The tree shows the structure of the changes made to the simulation and the log is required to keep the order in which the changes occurred. Together they form a record which you would normally find in a laboratory notebook. The history tree is plotted in simulation parameter space. This shows the scientist's interactions with the simulation visually and allows direct manipulation of the parameter information presented, which in turn is used to control directly the state of the simulation. The interactions with the system are graphical and usually involve directly selecting or dragging data markers and other graphical control devices around in parameter space. If the graphical manipulators do not provide precise enough control then textual manipulation is still available which allows numerical values to be entered by hand. The Virtual Laboratory Notebook, by providing interesting interactions with the visual view of the history tree, provides a mechanism for giving the user complex and novel ways of interacting with biological computer simulation models.