Previous work has developed the What You See Is What You Test (WYSIWYT) methodology for testing spreadsheets. This methodology has been shown to help end users test, debug, and modify spreadsheets. To date, however, this system has provided no support for creating, reusing, and managing test cases, a process that can be tedious and time-consuming. To alleviate this, we have developed automated test case generation and test reuse methodologies for the WYSIWYT methodology. For test case generation, we have prototyped two techniques, and performed a study to assess the effectiveness and efficiency of these techniques. The results of this study show that we can efficiently exercise a large percentage of a spreadsheet under test. We also implemented a test reuse methodology and performed a study that shows that we are able to find a relatively small subset of test cases to reuse after a modification to a spreadsheet, and that these test cases re-establish much of the coverage lost by modifications made to the spreadsheet. / Graduation date: 2003

Is it possible to achieve some of the benefits of formal testing within the informal programming conventions of the spreadsheet paradigm? We investigate an approach that attempts to do so via the development of a testing methodology for this paradigm. The "What You See Is What You Test" (WYSIWYT) methodology for testing spreadsheets supplements the automatic immediate visual feedback about values with automatic immediate visual feedback about "testedness". In this thesis, we present empirical data about the methodology's effectiveness resulting from two controlled experiments. The first experiment provided interesting but inconclusive results which spurred us to consider ways to improve the design of our experiment. We used the Cognitive Walkthrough method to evaluate and improve our design and readministered the experiment. Our results from the redesigned experiment show that the use of the methodology was associated with significant improvement in testing effectiveness and efficiency, even with no training on the theory of testing or test adequacy that the model implements. / Graduation date: 2000

End-user programmers are writing an unprecedented number of programs, due in large part to the significant effort put forth to bring programming power to end users. Unfortunately, this effort has not been supplemented by a comparable effort to increase the correctness of these often faulty programs. To address this need, we have been working towards bringing fault localization techniques to end users. In order to understand how end users are affected by and interact with such techniques, we conducted a think-aloud study, examining the interactive, human-centric ties between end-user debugging and a fault localization technique for the spreadsheet paradigm. Our results provide insights into the contributions such techniques can make to an interactive end-user debugging process. / Graduation date: 2004

Test case generation in software testing is a process of developing a set of test data that satisfies a particular test adequacy criterion. It is desirable to automate this process since doing it manually is not only technically difficult but also tedious and time-consuming. Although there has been considerable research in automatic test case generation directed at imperative languages, we find no research exists addressing the problem for spreadsheet languages. This problem is particularly important for spreadsheet languages, since spreadsheet languages are widely used by end users and most of them lack testing backgrounds. To address this need, in this thesis, we present an automatic test case generation methodology for spreadsheet languages. Based on an analysis of the differences between imperative languages and spreadsheet languages, we developed our methodology by properly adapting existing test case generation techniques for imperative languages. Our methodology is integrated with a previously developed methodology for testing spreadsheets, and supports incremental automatic test case generation and visual feedback. We have conducted a family of empirical studies to assess the effectiveness and the efficiency of the essential techniques underlying our methodology. The results of our studies show that the test cases generated by our methodology can exercise a large percentage of a spreadsheet under test. The results also provide insights into the tradeoffs between two test case generation techniques for spreadsheet languages. / Graduation date: 2001