Page 35 - The Art of Designing Embedded Systems
P. 35

22  THE ART OF DESIGNING EMBEDDED SYSTEMS



                               What does it cost to inspect code? We do inspections because
                          they have a significant net negative cost. Yet sometimes manage-
                          ment is not so sanguine; it helps to show the total cost of an inspec-
                          tion assuming there’s no savings from downstream debugging.
                               The inspection includes four people: the Moderator, Reader,
                          Recorder, and Author. Assume (for the sake of discussion) that these
                          folks average a $60,000 salary, and overhead at your company is
                          100%. Then:
                               One person costs:   $120,000  = $60,000 x
                                                  2 (overhead)
                               One person costs:   $58/hr  = $120,000/2080 work
                                                  hours /year
                               Four people cost:   $232/hr  = $58/hr x 4
                               Inspection cost/line:  $1.54  = $232 per hour/l50 lines
                                                  inspected per hour
                               Since we know  code costs $20-50  per line to produce, this
                          $1.54 cost is obviously in the noise.


                            For more information on inspections, check out Soware Inspection,
                       Tom Gilb and Dorothy Graham,  1993, TJ Press (London), ISBN 0-201-
                       63 18 1-4, and  Software  Inspection-An  Industry Best  Practice,  David
                       Wheeler, Bill Brykczynski, and Reginald Meeson,  1996 by  IEEE Com-
                       puter Society Press (CA), ISBN 0-8 186-7340-0.


                            Step 4: Create a Quiet Work Znvironment
                            For my money the most important work on software productivity in
                       the last 20 years is DeMarco and Lister’s Peopleware (1987, Dorset House
                       Publishing, New York). Read this slender volume, then read it again, and
                       then get your boss to read it.
                            For a decade the authors conducted coding wars at a number of dif-
                       ferent companies, pitting  teams against each other on a standard set of
                       software problems. The results showed that, using any measure of  per-
                       formance (speed, defects, etc.), the average of  those in the first quartile
                       outperformed the average in the fourth quartile by a factor of 2.6. Surpris-
                       ingly, none of the factors you’d expect to matter correlated to the best and
                       worst performers. Even experience mattered little, as long as the program-
                       mers had been working for at least 6 months.
   30   31   32   33   34   35   36   37   38   39   40