FURTHER READING For a 5 discussion of egoless in Visual C#.NET

Painting EAN13 in Visual C#.NET FURTHER READING For a 5 discussion of egoless

programming, see The Psychology of Computer 7 Programming, 2d Ed. 8 (Weinberg 1998).

I had a personal experience using inspections on the second edition of Code Complete. For the first edition of this book I initially wrote a rough draft. After letting the rough draft of each chapter sit in a drawer for a week or two, I reread the chapter cold and corrected the errors I found. I then circulated the revised chapter to about a dozen peers for review, several of whom reviewed it quite thoroughly. I corrected the errors they found. After a few more weeks, I reviewed it again myself and corrected more errors. Finally, I submitted the manuscript to the publisher, where it was reviewed by a copy editor, technical editor, and proofreader. The book was in print for more than 10 years, and readers sent in about 200 corrections during that time.

You might think there wouldn t be many errors left in the book that had gone through all that review activity. But that wasn t the case. To create the second edition, I used formal inspections of the first edition to identify issues that needed to be addressed in the second edition. Teams of 3-4 reviewers prepared according to the guidelines described in this chapter. Somewhat to my surprise, our formal inspections found several hundred errors in the first edition text that had not previously been detected through any of the numerous review activities. If I had had any doubts about the value of formal inspections, my experience in creating the second edition of Code Complete eliminated them.

Inspection checklists encourage focused concentration. The inspection process is systematic because of its standard checklists and standard roles. It is also selfoptimizing because it uses a formal feedback loop to improve the checklists and to monitor preparation and inspection rates. With this control over the process and continuing optimization, inspection quickly becomes a powerful technique almost no matter how it begins. The Software Engineering Institute (SEI) has defined an Capability Maturity Model (CMM) that measures the effectiveness of an organization s softwaredevelopment process (SEI 1995). The inspection process demonstrates what the highest level is like. The process is systematic and repeatable and uses measured feedback to improve itself. You can apply the same ideas to many of the techniques described in this book. When generalized to an entire development organization, these ideas are, in a nutshell, what it takes to move the organization to the highest possible level of quality and productivity.

more details on the SEI s concept of developmental maturity, see Managing the Software Process (Humphrey 1989).

0 CC2E.COM/ 2199

Do you have checklists that focus reviewer attention on areas that have been problems in the past Is the emphasis on defect detection rather than correction Are inspectors given enough time to prepare before the inspection meeting, and is each one prepared Does each participant have a distinct role to play Does the meeting move at a productive rate Is the meeting limited to two hours Has the moderator received specific training in conducting inspections Is data about error types collected at each inspection so that you can tailor future checklists to your organization

Is data about preparation and inspection rates collected so that you can optimize future preparation and inspections Are the action items assigned at each inspection followed up, either personally by the moderator or with a re-inspection Does management understand that it should not attend inspection meetings

Other kinds of collaboration haven t accumulated the body of empirical support that inspections or pair programming have, so they re covered in less depth here. The kinds covered in this section include walkthroughs, code reading, and dogand-pony shows.