HASLab/INESC TEC prevents errors in spreadsheets
A research team at the High-Assurance Software Laboratory (HASLab), INESC TEC’s Privileged Partner, has been applying strict methods – one of the group’s research areas – to spreadsheets in order to reduce errors.
Spreadsheet systems play a fundamental role in modern society. Because spreadsheets are flexible, they can be used for several purposes and by different types of users. A spreadsheet can either be created by a single individual in order to solve simple computational problems, or by a large company to integrate data from multiple systems, and as a foundation for business decision-making processes.
Although the importance of spreadsheets is undeniable, effective mechanisms to prevent errors have been somewhat disregarded. In fact, it is estimated that up to 94 per cent of the spreadsheets used in real situations contain errors which cause every year losses that amount to 10 billion dollars.
This is where HASLab comes in. This concrete scenario was the starting point for a coordinated research effort which started in 2007 with a doctoral project developed by Jácome Cunha, a researcher at HASLab. This was continued as part of the project “SSaaPP - SpreadSheets as a Programming Paradigm”, funded by the Portuguese Foundation for Science and Technology (FCT) and led by João Saraiva. This project’s research team studies strict approaches and how they can be used to develop spreadsheets. Their main purpose is to reduce the amount of errors that these spreadsheets contain.
The most important scientific contribution of this research team until now is a model oriented approach that is now widely accepted in other areas of engineering besides software engineering. Taking Civil Engineering as an example, bridges or houses are not built without a detailed abstract model of what is going to be built. Otherwise, correcting possible errors and inconsistencies during the construction phase would be extremely costly. The aim is similar in the context of Spreadsheet Engineering: a concrete spreadsheet where data are introduced will only be implemented when there is already an abstract model that fits the business model perfectly.
The results achieved so far have been widely accepted by the international scientific community and have been published in worldwide renowned conferences, such as the International Conference on Software Engineering (ICSE), the Visual Languages and Human-Centric Computing (VL/HCC), the International Conference on Fundamental Approaches to Software Engineering (FASE), the International Conference on Model Transformation (ICMT) and the Partial Evaluation and Program Manipulation (PEPM).
