Certitude and rectitude

There is a fundamental difference between certification (which is intended to give you the feeling that someone or something is doing the right thing) and correctness (for which you hopefully have some well-founded reason to believe that someone or something is doing the right thing - with respect to appropriate definitions of what is right). Certification is typically nowhere near enough; correctness is somewhat closer to what is needed, although often unattainable in the large - that is, with respect to the entire system. However, formal demonstrations that something is correct are potentially much more valuable than loosely based certification. So, a challenge confronting us here is to endow certification - of people and of systems - with a greater sense of rigor and credibility. Overall, we urgently need to explore alternatives within the context of the entire process of development, maintenance, and continued evolution. Although lowest-common-denominator certification of conventional programmers and simplistic metrics for judging organizational competence are likely to be palliatives at best, sensible procedures for certifying requirements engineers, system engineers, software engineers, debuggers, etc., could be just one of many potentially useful steps toward instilling greater discipline into the development process - particularly for critical systems.