Untitled Document

Giving up hope on hopeless problems seems rational – until the enormity of the problem forces us to re-think our assumptions. Software engineering measurement has suffered long under such a predicament. But when advancement occurs, the beneficiaries include several disciplines that primarily depend on intellectual labor. Today, with the proliferation of software as a development tool, such challenges are indistinguishable be they in hardware design, networking, dependability, or services.

Ten years ago the invention of Orthogonal Defect Classification [ODC] made us re-think how information captured in defects is leveraged to gain process signatures and insight. As the technology advanced, cost and efficiency of process improvement methods crossed an order of magnitude in savings - something unheard of in this business. Newer methods, algorithms and applications also grew. Test efficiency is one of the newer conceptual extensions, while rapid prototyping and robotics are some of the newer application domains. Today there is considerable penetration across the industry and experienced development organizations such as IBM, Nortel, and Telcordia regard it as one of their revered best practices.

Beyond the technology, my experience with ODC over the past decade has taught me two lessons. One, that new engineering methods need to be built not merely on physical abstractions but also on the spirit of current intellectual labor. Two, that problems in software engineering, a domain representing a cocktail of nerdy algorithms, human dynamics, and financial valuations, gives us researchers a fertile space to explore the future of management and engineering.

Audience: Technical, Academic
Length: 1 hour - discussion can follow.