High-integrity software systems are often so large that conventional development processes cannot get anywhere near achieving tolerable defect rates. This article presents an approach that has delivered software with very low defect rates cost-effectively. We describe the technical details of the approach and the results achieved, and discuss how to overcome barriers to adopting such best practice approaches. We conclude by observing that where such approaches are compatible and can be deployed in combination, we have the opportunity to realize the extremely low defect rates needed for high integrity software composed of many million lines of code.
I haven't had time to read over the article much yet, but find the premise interesting.