SYSTEMS

By grouping the existing 350-plus metrics for object-oriented systems, a Penn State researcher has determined that using the right metrics at the right time can help avoid the costly mistakes that can hobble small and large-scale projects. "No one had proposed a systematic approach to deciding when to apply specific metrics whether at an early stage, throughout the development or at the end," said Dr. Sandeep Purao, associate professor of information sciences and technology. "Knowing this can help managers on a regular basis keep tabs on software being developed, and that will mean fewer failures and fewer problems." From their review, Purao and co-researcher Vijay Vaishnavi, professor of MIS at Georgia State University, discovered gaps in and overlaps among the existing metrics that are applicable to any IT development project using the object-oriented approach. To do this, the researchers developed a mathematical framework that groups metrics around the stages of project requirements; design specifications; implementation; and operation in the software system's intended environment. The framework revealed that some metrics are applicable only at certain product stages while others are applicable at multiple times. The researchers also realized few metrics exist for measuring the processes employed and resources expended during software development. "There is a lot of research on object-oriented techniques, but researchers and managers need to know which metrics are appropriate at different stages of the system development process," Purao said. "We have done that." Information system snafus affect everyone from PC users to cities and corporations and can carry hefty price tags. The errant computerized baggage handling system that chewed up luggage at Denver International Airport contributed to that project's $4 billion cost, more than double original estimates. The Baltimore (MD) City school district, already pinched, is paying out close to $500,000 a month because of a software problem that cropped up in a new $16 million computer system. Rather than waiting to test projects until they are completed, evaluating them at increments can prevent such troubles. That can keep projects from having cost overruns and needing time extensions, Purao said. The researchers' future work will determine relationships among the metrics, so that interdependent metrics can be used simultaneously. Ultimately, Purao and Vaishnavi hope to identify which metric suites may be most effective in different situations.