Business bears burden of buggy software costs

Business is bearing the spiralling cost of buggy software, which adds up to billions of dollars a year, as IT shops continue to expend resources fixing software errors.

While Australian IT managers blamed the rush by vendors to get products to market, local software developers said features, rather than quality, drive the market.

The National Institute of Standards and Technology (NIST) attributed a hefty price tag to the problem, claiming software bugs are costing the US economy $US59.5 billion yearly, with users incurring 64 per cent of the cost, and developers 36 per cent.

An IT manager from a local manufacturing firm, who wished to remain anonymous, said software is one of the few markets where buyers actually accept faulty products. There's plenty of room for improvement when it came to testing by vendors, he said.

However, Australian Agile Methods consultant and author Steve Hayes believes the market is unwilling to pay for higher quality software.

Instead, he said, it is driven by features.

"Users know that early versions of software are notoriously buggy, but still plan developments around features that aren't going to be available until the next untested release," Hayes said.

"Until the market demands higher quality rather than new features, the situation is unlikely to improve substantially."

Hayes said there is hope in agile software development methods, which emphasise continuous, integrated, automated testing.

Dr Jeff Gosper, CSIRO group leader of software architecture and components group of the mathematics and information science division, said business is having to pay for this costly problem because so many IT staff are forced to diagnose and correct software errors.

Gosper advocates more rigorous testing by suppliers and a more components-based approach, reusing components that are reliable.

"I speak to companies that want to upgrade [software], but hold back because the first release is too buggy which makes it costly for them. There are two levels to this problem: software you buy in and software you develop in-house, but you are still dealing with software that isn't necessarily robust," he said.

The study did not propose specific action for improved testing but called for the development of standards as well as standardised testing tools.

Join the newsletter!

Or

Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.

More about Agile SoftwareCSIROHayes

Show Comments