FWF - iDEOS - Interactive Spreadsheet Debugging

Project: Research project

Description

Spreadsheets are omnipresent in organizations. They are used for a variety of purposes and in many cases, calculations in spreadsheets directly serve as a basis for reporting and for subsequent managerial decision making. One reason for the widespread use of spreadsheets is that they can be developed (“programmed”) by end users, e.g., an employee in the controlling department. Being able to develop one’s own data aggregation and decision support tools increases the flexibility for the end users. However, one potential problem with spreadsheets is that they are typically not subject to quality assurance (QA) processes that are common in traditional
software development, like code inspections or systematic testing. In many cases, spreadsheet developers also lack the awareness regarding the importance of software testing. Moreover, today’s spreadsheet environments provide only limited functionality for fault avoidance, detection, and removal.
As a result, faults are not uncommon in spreadsheets and numerous cases have been reported where such faults led to substantial financial losses for companies1 or to miscalculations in scientific investigations, for a prominent example in the field of economics.
Due to the risks that can arise from faulty spreadsheets, researchers have proposed a variety of approaches to provide better tool support for spreadsheet developers. These techniques range from advanced spreadsheet visualizations, over algorithmic spreadsheet testing and debugging methods, to solutions that aim at the automated repair of erroneous spreadsheets.
The proposed project continues these lines of research and specifically focuses on improved debugging support for spreadsheets. While existing work in that area often focused on algorithmic fault localization, recent research indicates that to be truly helpful for users, debuggers should offer more functionality than only providing ranked lists of fault candidates. In this project, we will therefore explore novel mechanisms for interactive spreadsheet debugging, which, for example, proactively guide the user to the true location of the fault. From a methodological perspective, one key aspect of our joint research project is that we follow a dual approach. Besides traditional computational experiments that, e.g., compare the fault ranking performance and the computation times of different techniques, we plan to run different controlled user studies, where study participants solve debugging tasks with the help of the tools that will be developed in the project.
StatusActive
Effective start/end date15/09/1914/03/23