Experimental Evaluation of Time Allocation Procedure for Technical Reviews

This paper proposes a new procedure for time allocation to technical review of software products. The procedure designed is based on an experienced knowledge that the lower the quality of a product, the higher the effort needed for reviewing it. In order to identify the low quality products, we use the idea of sampling inspection which is popular for quality inspection of manufactured goods. The proposed procedure consists of the following three steps: (1) each member of a team chooses any one of the products developed by himself or herself, (2) all members of the team jointly review the selected products and identify faults, and (3) review time for each product is determined based on both the number of faults detected in (2) and the size of the product. In the proposed procedure, step (3) is implemented by constructing a check list to be used in technical reviews such that the number of items in the check list is dynamically specified according to both the number of faults identified and the size of the product. Moreover, we apply the proposed procedure to software development projects in a certain industrial environment. The analysis results of experimental data collected from software developments show the usefulness of the proposed procedure.