In financial engineering today we see a growing collection of mission critical and enterprise critical applications. These demand the highest levels of software quality and reliability in order to maintain confidence among end users. At the same time, market forces in the banking and IT sectors call for ever shorter time to market and shorter product lifecycles.
Fortunately, the financial sector is unusually well suited to a mathematical approach to software quality and testing. Much trading software relies on precise mathematical models from financial calculus and statistics (see for example Baxter and Rennie [1997]). Furthermore, packages often integrate standard numerical routines such as Monte Carlo and Finite Elements.
We propose that a productive new approach to software testing for this sector could be based on combining ideas from formal methods (see for example Loeckx and Sieber [1987]) and traditional software testing (see for example Beizer [1995]). We will describe a general approach to automated software testing based on logical constraint solving. We discuss what scientific and industry barriers must be overcome to make this approach viable in practise. Finally, we suggest a model for industrial and academic collaboration that could overcome these obstacles and make progress.
References
M. Baxter and A. Rennie, Financial Calculus, Cambridge University Press, 1997.
B. Beizer, Black-Box Testing, John Wiley and Sons, Chichester, 1995.
J. Loeckx and K. Sieber, The Foundations of Program Verification, John Wiley and Sons, Chichester, 1987.
Back to Karl Meinke's Home Page