A few years ago, while I was working at Neco, I was involved in the development of the OPEN/Metis software development methodological framework. In addition to including a metamodel and some extra stuff (of not much direct use to software practitioners), OPEN/Metis is basically a software development methodology. I am a professional software engineer, and therefore the idea that no single methodology can solve all the problems is embedded in each molecule of my body. That is, the old saying that one size fits all is plainly wrong, at least in academe.
However, my experience in industry shows that, for any given company, most of its projects are very similar in scope, size and even area of expertise. Perhaps we need to exclude here some hugely diversified companies that work in a multitude of areas. But, for most of the mid-size companies, this is, in my experience, quite valid. Therefore, the methodology that they use can be specified to a high degree of concretion leaving little room for variation. As long as the methodology is a good one, of course.
My point is that, in software methodologies for mid-size companies that develop custom business applications (my area of expertise), one size fits most. A single methodology with a few customisable parts can work and, in fact, is cheaper and easier to deploy.
Massively customisable solutions, such as method fragment repositories, are the top end products in the customisation scale. They can be used to generate an optimal methodology for each project or endeavour. But they are also expensive and need a considerable time and effort to convert “raw matter” (i.e. method fragments) into a usable methodology. This process usually involves expensive consultancy fees.
This is why the major component of OPEN/Metis is a methodology, targetted to mid-size companies that develop custom business applications. Its customisability is not excellent but, in exchange, its usability and affordability are extremely good.
What are your views out there?