Monday, November 16, 2009

A Theoretical Basis for Maturity Models (Part 2)

This is another article discussing my proposed Genealogical Maturity Model (GMM).

Last time, I abruptly closed in the middle of discussing using levels in maturity models. I’ll try and pick up right where I left off.

Levels Enable Aggregation of Multiple Factors

“Multiple, independent input variables” is a fancy way of saying that many factors determine maturity. For example, doctors may categorize premature infants by gestational week. Infant survival is dependent on many factors: respiratory development, birth weight, gender, single vs. multiple birth, brain blood vessel health, ductus arteriousus closure, intestinal health, liver development, anemia, immune system development, mother’s health, medical team, and medical facilities. Some of these factors are interdependent; infants without sufficient respiratory development are most likely to suffer bleeding around the brain.1 Other factors are independent; the degree of respiratory development in the womb is largely independent of medical personnel available to the case.

Still, the medical community finds value in grouping preterm infants by weeks gestation. Likewise, maturity models of complex systems like software development and genealogy can also find value using groupings, despite the existence of multiple input variables—factors.

The result is a model that only approximates reality. The better the approximation, the more useful the model.

Totally Ordered Sets

Last time I also threw the term totally ordered set at you. That’s a fancy way of saying that to match the definition of “maturity,” the levels have to progress from one to the next. For example, if I gave you the numbers from 1 to 5, you could place them in ascending order. If I replaced one of the numbers with an apple, and another of the numbers with the color purple, there won’t be any natural order to the five items.

To match the definition of maturity, a maturity model must define levels such that they are a totally ordered set.

Narrow Objective

After reading all your comments, I believe one reason Humphrey was successful with his Capability Maturity Model (CMM) was by narrowly defining the objective of the model. The objective of CMM is not to rate a company’s ability to successfully develop software. The focus is much finer. CMM rates a company’s ability to successfully develop software on time and on budget. (Some of you that have used CMM might be able to narrow this even further. Might it predict success only for large, DoD sized projects?)2


     1. “Quick Reference Fact Sheets: Premature Birth,” March of Dimes (www.marchofdimes.com : dated January 2009, accessed 15 November 2009).

     2. Wikipedia contributors, "Capability Maturity Model," Wikipedia, The Free Encyclopedia (http://en.wikipedia.org : accessed 11 November 2009).

No comments:

Post a Comment