This prompted a re-design of the GUI, for which the main objectives were to substantially reduce the total number of windows, and to decrease the number that needed to kept open simultaneously. To do this I needed to find more efficient data representations, combine multiple graphical charts into one, and to compress the contents of as many pop-ups as possible into the main window.
In the language of Tufte's theory, I needed to maximize the data density of the interface. Data density is the ratio of the size of the data matrix behind a graphic to the area of the graphic, and Tufte suggests the following editing principle: Maximize the data density and the size of the data matrix, within reason. He also advises us to maximize the data-ink ratio, where data-ink is the non- redundant information content of the graphic. And, we should aim to erase both non-data ink and redundant data-ink. ( Tufte 1983.)
The original design consisted of one primary window and eight secondary pop- up windows. They all contained dynamically-updated information about the health of the network. Figure 1 shows the main window and one of the pop-ups.
The main window displayed summary and overview data, highlighting any problems or other "interesting" conditions for which more detailed data could be obtained from one of the pop-up windows. Unfortunately, due to space restrictions, the three "thumbnail" charts had no scales and few actual data values, so 3 of the 8 pop-ups were simply enlarged versions of the same charts.