BD00 has been following the development of Ivar Jacobson et al’s SEMAT (Software Engineering Method And Theory) work for a while now. He hasn’t decided whether it’s a revolutionary way of thinking about software development or a bunch of pseudo-academic malarkey designed to add funds to the pecuniary coffers of its creators (like the late Watts Humphrey’s, SEI-backed, PSP/TSP?).
To give you (and BD00!) an introductory understanding of SEMAT basics, he’s decided to write about it in this post. The description that follows is an extracted interpretation of SEMAT from Scott Ambler‘s interview of Ivar: “The Essence of Software Engineering: An Interview with Ivar Jacobson”.
As the figure below shows, the “kernel” is the key concept upon which SEMAT is founded (note that all the boasts, uh, BD00 means, sentences, in the graphic are from Ivar himself). In its current carnation, the SEMAT kernel is comprised of seven, fundamental, widely agreed-on “alphas“. Each alpha has a measurable “state” (determined by checklist) at any time during a development endeavor.
At the next lower level of detail, SEMAT alphas are decomposed into stateful sub-alphas as necessary:
As the diagram below attempts to illustrate, the SEMAT kernel and its seven alphas were derived from the common methods available within many existing methodologies (only a few representative methods are shown).
In the eyes of a SEMATian, the vision for the future of software development is that customized methods will be derived from the standardized (via the OMG!) kernel’s alphas, sub-alphas, and a library of modular “practices“. Everyone will evolve to speak the SEMAT lingo and everything will be peachy keen: we’ll get more high quality software developed on time and under budget.
OK, now that he’s done writing about it, BD00 has made an initial assessment of the SEMAT: it is a bunch of well-intended malarkey that smacks of Utopian PSP/TSP bravado. SEMAT has some good ideas and it may enjoy a temporary rise in popularity, but it will fall out of favor when the next big silver bullet surfaces – because it won’t deliver what it promises on a grand scale. Of course, like other methodology proponents, SEMAT’s advocates will tout its successes and remain silent about its failures. “If you’re not succeeding, then you’re doing it wrong and you need to hire me/us to help you out.“
But wait! BD00 has been wrong so many times before that he can’t remember the last time he was right. So, do your own research, form an opinion, and please report it back here. What do you think the future holds for SEMAT?