Interesting, I guess sometimes there are fun coincidences. In recent emails with old friends, I found many parallels in the conversation with my schooling's current lessons. My current class is Creativity Measurements. Since the definition of creativity is hard to pin down, much research has been done in a qualitative fashion (as well as quantitative, but I will get to that another time).
When friend mentioned that we were trained in the constrictive way as system analysts (as engineers are trained, which is what computing grew from), we did it in a quantitative fashion. Rigorous, to the numbers, and by the book. Everything is generic, but constrictive and defined. Works well in an engineering background. This is why I think CMMi doesn't work well for software development. It grew from engineering quality control which is restrictive.
What we need to look at (and as the group on the emails alluded, the practice is evolving to) are qualitative development measures. Things that deal with the human aspect. That your system is affected by the subjects as they are affected by the system. Emergent. The crux of qualitative measurements are validity. The aforementioned chum is alluded to it, in the "do no harm by allowing this field/code" (I know I am paraphrasing) as long as it is a valid use. Hmm.. a parallel. This is also why Agile works better for project management at least in software engineering. The Organic Software Development Approach OSDA(patent pending, trademarked, copyrighted, etc....).
I guess what I am looking at is the former had too much science and little art. The later has a balance of art and science. My thoughts are if qualitative development is used too much, we could get pseudo-science, which is the worse case scenario. Too much art and not enough science. You need validity in your qualitative research.
Interesting article: Validity in Qualitative Research, Whittemore, Chase, Mandle in the Qualitative Health Research Vol.11 No. 4 July 2001
Some food for thought.
~Jeffrey
No comments:
Post a Comment