Posted by: Jan Stafford
software development, Software Quality, Software testing
Software testers and quality assurance managers know that perfect software doesn’t exist.
Unfortunately their project managers, company executives and legal teams often don’t, and these people have “unreasonable expectations,” experience “constant disappointments” and make “disastrous decisions” regarding software testing, said Gerald M. Weinberg, author of many IT and software quality books, in our recent phone conversation.
Weinberg wants those people to read his new book, Perfect Software And Other Illusions About Testing.
“This book is designed for people who don’t fully know what testing encompasses, but make decisions that affect how and how much software testing is done,” said Weinberg.
The current economic downturn and resulting budget cutbacks in software development make informing decision makers and influencers about when, how and why software must be tested even more urgent. While it seems obvious to software testers that cutting software testing is a sure way to produce faulty software, it’s not as obvious to managers.
“Management thinks of testing as what is done after the software is developed,” Weinberg said. “They schedule testing at the end of the process, and that’s the easiest place to cut if there’s a cost overrun.”
Naturally, that’s when the QA manager should step in and explain why testing shouldn’t be cut; but often the lack of funds defeats such arguments at the end of a project.
Establishing processes wherein software is tested early and continually is a better practice than testing at the end of development. “There should be more people wearing testing hats at each stage of development, because mistakes made early are very hard to find later,” Weinberg said.
Reducing development costs by reducing testing is a penny-wise, pound-foolish approach. More and more, Weinberg said, software-related lawsuits and malpractice cases verdicts come down to what software wasn’t tested adequately.
Another common management mistake is not putting software support on the same ledger as development.
“Accurate cost accounting takes into account post-release costs,” Weinberg said. “Usually, the cost of fixing errors doesn’t get attributed to development managers and project managers. It should be.”
Weinberg hopes this book will help managers get “tuned into reality” about software testing. The key is getting it into the right people’s hands, he said, noting that testers and QA pros may want to keep it handy to do just that.
If you’re having trouble getting support for testing in any area or face testing cutbacks, our resident site experts can offer helpful advice. Just send your questions or describe your problem in an email to email@example.com. This software security pro wrote in and got advice on how to get management support for security quality.