Qualitative Deviation Models for Spreadsheet Debugging
Birgit Hofer, Iulia Nica, Franz Wotawa · 2017
Methods and tools for supporting programmers when localizing the fault after revealing a failure help to reduce effort and costs. This applies in particular to enduser programming where programmers usually do not have experiences in software engineering methodologies. Tools for end-user programming have additional requirements like ease of use and interactivity constraints. Debugging approaches like model-based debugging have high computational requirements and can hardly be directly used especially for larger programs. In this paper, we focus on improving model-based debugging of spreadsheets. In particular, we formally introduce a new improved model that makes use of the deviations obtained between the expected value of cells and the actual ones. The qualitative deviation model allows to specify whether a value is higher, lower, or equivalent to its expectations. In addition, the diagnosis time is small even for large spreadsheets. This allows interactive debugging for spreadsheets.