I was at the AIC 2010 conference yesterday, which I enjoyed more than last year. The most interesting session was Ivar Jacobson. He discussed how immature our industry is, with its disjoint between software engineers and academic computer scientists. Something I have commented on before when discussing if our industry should be licensed.
He discussed how we need to build our industry upon commonly agreed kernel of simple best practice activities, not upon some currently fashionable process whether it be Waterfall, CMMI, Agile, Lean etc. We must identify what are the best practices from all process (and all processes have something to offer) and teach them to students at University as this kernel of activities, as well as teaching how to compose them into more complex practices and hence whole development processes. Thus providing new entries to our industry with the base of reusable techniques that the can use for their whole career, irrespective of fashion. The key idea being that any current process model can be build by composing these kernel of basic activity.
So if this sounds interesting, and it does to me, have a look at www.semat.org. The signatories aim to have results within the year, if they achieve this aims this could well be the first step to making Software Engineering on a par with other chartered engineering disciplines such as Structural or Mechanical Engineering, with there is a long term set of industry accepted best practices that people can be judged against.