Software Engineering: 40 Years

During the summer of 1998, I visited Germany for the first time, backpacking solo cross-country throughout former West Germany while not neglecting to visit some key sites in former East Germany. Some of the most pleasurable destinations in former East Germany included Berlin (most notably the incredible Pergamonmuseum), the Harz Mountains where my mother was born (including Brocken), and the Reformation museum in Lutherstadt-Wittenberg. Garmisch-Parmenkirchen in the German Alps was my favorite destination in former West Germany due to the hiking at the Partnachklamm, and I find it amusing that the first NATO "Conference on Software Engineering" took place in that locale 40 years ago.

Tom DeMarco wrote an article in the recent IEEE Software issue on the evolution of the software engineering discipline. As a member of IEEE, I subscribed to IEEE Software earlier in my career, and have since turned to other sources of information, but came across the August 20, 2009 InfoQ post that sited this article, which discusses software engineering in light of project control and metrics. The intial replier to the post seemed confused about the point of the article, prompting me to post a reply, which reminded me of DeMarco's collaborative work with Timothy Lister called Waltzing with Bears, the best material I have read on risk management, in which the authors deem risk management "project management for adults".

My reply to the initial poster, who did not seem to understand DeMarco's positioning on what types of projects should be pursued:

The ability to obtain what DeMarco refers to here as "precise control" of a project would seem to indicate that the project is of a low risk nature. Outsourced routine maintenance is an example of low risk project work. The decision to choose "projects where precise control won't matter so much" points to projects of a high risk nature. High risk projects generally have the ability to produce high value, whereas low risk projects generally do not. It is typically easier to control projects where risk is low because there exists more certainty. Pursuit of high risk projects should not be abandoned due to lack of predictability. At the same time, risk should always be managed.