Authors:
Dietmar Jannach
;
Arash Baharloo
and
David Williamson
Affiliation:
TU Dortmund, Germany
Keyword(s):
Spreadsheets, Debugging, Artificial Intelligence, Software Risk.
Related
Ontology
Subjects/Areas/Topics:
Service-Oriented Software Engineering and Management
;
Software and Systems Development Methodologies
;
Software Engineering
;
Software Quality Management
Abstract:
Spreadsheet applications can nowadays be found nearly everywhere in companies and are used for a variety of purposes. Because of the high risk that arises when business decisions are based on faulty spreadsheets, in recent years new approaches for spreadsheet quality assurance have been proposed. Among them are techniques that allow for more intelligent tool support during the spreadsheet test and debugging process.
The design and evaluation of such new methods and tools, which are for example based on model-based techniques is however challenging. On the one hand, real-world spreadsheets can be large and complex, thus requiring highly efficient and scalable error-location algorithms. On the other hand, as spreadsheets are usually developed by non-programmers, special care has to be taken when designing the debugging user interface.
In this paper, we discuss these challenges and present the design and architecture of an integrated framework for spreadsheet debugging called Exquisit
e. Furthermore, we report details and first experimental results of a constraint-based debugging approach implemented in the framework, which supports the automated identification of possible spreadsheet errors based on user-provided test cases and model-based diagnosis techniques.
(More)