Software development: Quality code saves time, says Compuware

While software quality is often an afterthought for programmers eager to complete assignments, Compuware says development can actually be completed faster by taking time to check quality.

According to Marianne Prins, application reliability specialist at Compuware South Africa, the issue of software quality is not a new problem. "Software development has always been prone to bugs, badly compiled code and poor performance in the live environment. These issues are a factor in the programming difficulties, as well as the tendency for developers to approach quality as an afterthought," she explains.

Prins says software problems can be resolved through the utilisation of quality control measures in the four factors involved in software development: people, process, methods and tools. "By focusing on these, developers will dramatically reduce the time taken to iron out problems, while producing code that is more effective and efficient," she says.

However, introducing quality to the process has not been widely accepted as yet by developers - a study by software development specialist and author Capers Jones found that fewer than 6% of organisations have clearly defined software development management processes in place; with the result that development is not the tightly controlled activity that it should be.

"Additionally, research indicates that in most companies, the cost of poor quality software runs at 20% to 40% of the total project cost. In other words, up to two-fifths of development effort is spent re-doing things that went wrong because of poor quality," she states.

Prins notes that accounting for the cost of quality requires a calculation of two metrics - the costs incurred due to a lack of quality, and those incurred in the achievement of quality. "Adding quality control across the software development process will incur a higher cost in terms of more stringent attention to detail as well as additional tools and methodologies. However, this has to be measured against the savings that will be made by removing the need for revisiting and recoding when poor programming is discovered later in the project and cost to the business when an application falls over," she explains.