Editor's Note: Software Quality Is Everyone's Problem

One topic that we've been discussing with readers of InformationWeek and Optimize lately is intangible assets--things that are difficult (sometimes impossible) to measure but often contribute significantly to the health of the business and that all-important line at the bottom of financial statements. Intangibles, which may include knowledge sharing, incentives, company culture, and brand awareness, can account for as much as half of the market value of large industrial and services companies, according to some experts. But they certainly play a role in companies of all sizes and in all industries.

Here's another intangible to consider: software dependability. How do you place value on dependability? How do you measure it? You hope for it in your car, your child care, your banking system, and many other areas in your life (I should know: My car battery, washing machine, and the air conditioning in my house all conked out in the same week!). In the software market, one major vendor claims its products are unbreakable and another is trying to make its software more trustworthy. Even with that, it's difficult for buyers to walk away from a purchase feeling as if they have a guarantee of dependability.

Dealing with software problems is an experiment in conditioning and behavior modification. Business-technology professionals, it seems, have been conditioned to accept a certain amount of software failure. It's not unusual to be bombarded by a series of bug fixes or security patches after a new application is installed. But for a growing number of software users, this is no longer acceptable--not in their home PCs, their businesses, and certainly not in the nation's critical infrastructure.

A group of influential consumers of software and vendors have joined Carnegie Mellon University in the newly launched Sustainable Computing Consortium (see story, "Consortium Looks For A Long-Term Solution"). The goal is to create standards and methodologies to reduce defects and to provide ways to measure and quantify quality, dependability, and security. It's an admirable and lofty goal. Finding ways to measure software quality will surely lead to better ways of managing it.

Getting back to psychology, negative reinforcement is the easy way--customers can put added pressure on vendors to create more dependable, less buggy software by voting with their wallets. But as any psychologist will tell you, positive reinforcement is more successful. Vendors that provide better quality and assurances also may be able to attach a premium to their software or boost their market share. Ideally, it would have a domino effect. Ty Sagalow, chief operating officer of AIG e-Business Risk Solutions and a consortium member, reminded me of the first carmakers that installed airbags. Now, of course, you can't buy a car without one.

Fortunately, vendors are getting serious about software quality. Microsoft and Oracle, for example, have joined the new consortium and will be able to collaborate with other members. After all, is it really fair to put all the blame on the software vendors? (See p. 104 for more perspective.) Says Sagalow, "We have to share the risk, and we don't want to curb innovation or creativity."

Stephanie Stahl
Editor
[email protected]