The Software Process Assignment Two – Simulating a Process This assignment is worth 10% of the final mark for this unit. The report is due on saturday the 3th September at 3pm. The report should include the graphs and a discussion covering the results and conclusions. Extra marks are available for early submission. The attached sheet contains data adapted from an actual test process. The process had two stages: finding defects and fixing defects. The table represents the rate of finding defects by one competent tester working 25 hours per week over a twenty-week period and 300 detects were discovered in that time. They were classified as having a major impact on the usability of the system or only a minor effect – and is hard or easy to fix. The hard ones will take five hours to fix and the easy one can be done in two hours (on average). You are the project manager with three experienced software engineers (each working 25 hours per week) available for testing the system and fixing the defects. The users have stated that the major defects are "seven times as damaging" on average as the minor ones. Clearly state your objectives and any assumptions you have to make in the simulations. Discuss the measurement process and the data. What other data might have been collected? Fit defect detection curves to the data. Do they fit well? How many defects do you estimate are still in the system after the 300 defects have been found? Choose at least three appropriate metrics for measuring the test and repair process. Some possible metrics for illustrating the progress in each week (in each case the smaller the value, the better) are: [1] Estimate of number of unfixed defects (found & unfound) still in the software [2] Total importance of found defects still to be fixed (= 7 x #major + 1 x #minor) [3] Average time in queue of major defects still to be fixed [4] Ratio of defects fixed to defects found One allocation of your staff might be to have one person testing and the other two fixing the defects. However, you cannot fix all the defects at once. Repair strategies might be: • Solve them in a random order • Solve the easy ones first • Solve the hard ones first • Solve the major ones first • Solve the minor ones first • Solve them in the order in which they are found (FIFO) Produce a table (or graphs – Excel is acceptable) of your metrics each week for each of the six strategies. Relate your results to system reliability and also customer satisfaction. Consider the use of your resources. How could the performance of your team be improved if the tester was shifted to repair activities for a few weeks? - or if one of the repair engineers was given the task of testing for a few weeks? If you wished to minimize the time major defects went unfound and unfixed, what would your optimum allocation of staff be? Simulating a Process APPENDIX - ORIGINAL DEFECT DETECTION DATA TABLE Week Hard & Major Hard & Minor Easy & Major Easy & Minor Total 1 10 6 7 25 48 2 7 9 8 20 44 3 6 5 4 15 30 4 7 4 8 8 27 5 2 5 1 14 22 6 3 3 2 16 24 7 1 2 1 15 19 8 1 1 3 9 14 9 1 4 1 7 13 10 2 2 2 5 11 11 1 1 1 5 8 12 1 0 4 3 8 13 2 1 3 2 8 14 2 1 3 2 8 15 1 1 0 1 3 16 0 0 1 1 2 17 2 2 0 0 4 18 0 1 0 1 2 19 1 2 0 1 4 20 0 0 1 0 1 Total 50 50 50 150 300 Hours to Fix 5 5 2 2 Total Work 250 250 100 300