i-nth logo


Elisabeth Getzner


Spreadsheets are widely used by end-users in the corporate environment, as they allow the display and calculation of large amounts of data. Real-life application spreadsheets are often the basis for critical computations, yet research has found that many of them contain errors, which could cause financial loss if left uncorrected.

In this paper, we examine fault localization techniques which support the end-user in locating such errors. We provide an overview and comparison of the research in this area, including trace-based, model-based and combined approaches.

Among others, we discuss the techniques WYSIWYT, SFL, Sendys as well as ConBug and Exquisite and explain their process to locate faults. These approaches are compared with regards to their user input, runtime complexity, their ability to handle multiple faults and their outputs.

Additionally, we provide an overview of the evaluations for the presented tools, comparing their setup and the metrics used to measure the success of these approaches.


Partial program dependency graph
Partial program dependency graph

This example shows the spreadsheet references and their data dependency arrows. The faulty cell D2 has a red border and the faulty reference to C3 is indicated by a red arrow.


2014, Graz University of Technology, October

Full article

Survey of fault localization techniques for spreadsheets