- End users' programs are fraught with errors, costing companies millions of dollars. One reason may be that researchers and tool designers have not yet focused on end-user debugging strategies. To investigate this possibility, this dissertation presents eight empirical studies and a new strategy-based end-user debugging tool for Excel, called StratCel.
These studies revealed insights about males' and females' end-user debugging strategies at four levels of abstraction (moves, tactics, stratagems, and strategies), leading to specific implications for the design of tools. Particular results include a set of ten debugging stratagems, which generalized across three environments: the research spreadsheet environment Forms/3, commercial Windows PowerShell, and commercial Microsoft Excel. There was also evidence of the stratagems' generalization to a fourth set of environments: interactive application design environments, such as Dreamweaver, Flash, and Blend. Males and females statistically preferred and were effective with different stratagems, and females' stratagems were less well-supported by environments than males' were. In addition to what stratagems our participants used, we also investigated how these stratagems were used to find bugs, fix bugs, and evaluate fixes. Furthermore, a Sensemaking approach revealed end-user debugging strategies and the advantages and disadvantages of each.
We then built StratCel, which demonstrates a strategy-centric approach to tool design by explicitly supporting several of the implications from our studies. StratCel's evaluation revealed significant benefits of a strategy-centric approach to tool design: participants using the tool found twice as many bugs as participants using standard Excel and fixed four times as many bugs (including two bugs which had not been purposefully inserted by researchers and had gone unnoticed in previous studies). Further, StratCel participants did so much faster than participants using standard Excel. For example, participants using StratCel found the first bug 90% faster and fixed it 80% faster than participants without the tool. Finally, this strategy-based approach helped the participants who needed it the most: boosting novices' debugging performance near experienced participants' improved levels. The fact that this approach helped everyone -- males and females, novices and experts —- demonstrates the significant advantage of strategy-centric approaches over feature-centric approaches to designing tools that aim to support end-user debugging.