Abstract:

Field testing of industrial software may be insufficient due to risk and other constraints. Aiming at the above problems in practice, we propose an innovative approach of field testing. The concept of testing subset and the related attributes are defined. Testing subset is designed according to a comprehensive variety of factors concerning customers' production, through which the initiative of the field testing is improved. Every testing subset can be driven by corresponding auxiliary program for field testing. Preparation databases for different testing subsets are established based on historical data. The input data for active tests are designed based on the search results of the preparation databases in combination with the factors on the field through multiple steps. The integrated implementation process in practice is illustrated. The practical case of field testing on the optimal control system of petrochemical production proves that the proposed method can improve the sufficiency and efficiency of field testing significantly on the premise of risk prevention.

Key words: field test;testability;test coverage;risk;software testing