Procedural User Testing
The first two tests were often wildly inaccurate, however subjects quickly began to understand the concept of looking at a flattened version of the horizon and reached a consistent error rate (pages 1, 2). In general more points of data improved accuracy however in the second test (page 2) where many points of data were supplied; the accuracy was not substantially improved. The single most important tool for improving accuracy seems to be anything that shows the direction one is looking on the map, either through “book ending” (page 14) or showing the cone of vision as un-shaded on an otherwise shaded map (pages 6, 8, 4, 10).One possible conflict might prove to be a tendency toward landmarks. Even if the site to be found was nondescript, our subjects tended to assume that the point belonged to a feature on the horizon.