Insights into Decision Making

Making decisions is part of our everyday life, and software projects are no exception. What architectural style would suit our application best? Should we go for the simpler design, or the more complex one that in return offers more efficiency? Should we use generic containers in our API or will typed objects be the better option? Who of the available team members should be working on the project? And who should be the team lead? Should we agree to the deadline the customer wants, or would it be safer (and more honest) to announce we’ll need a month longer? Questions like these force us to make decisions, and each such decision must aim at the best possible solution, whatever ’best’ means in a specific context. Yet as we learn more and as conditions change, we may find that our decision wasn’t perfect and that we have to adapt our strategy. As cognitive psychologists don’t fail to remind us, questioning past decisions and looking for better strategies is an integral part of the human mind (Pinker 1997). However, changing decisions too frequently will cause ambiguous and inconsistent behaviour, and will result in a lack of decisiveness. The goal must be to find strategies that are both adaptive and straightforward. This paper takes a look at other disciplines in search of suitable analogies: physics, biology, physiology, control engineering and economics. The paper presents five patterns which all address a context that is best described as the decision-making processes in software engineering, and it takes its examples mainly from this area. The underlying principles, however, can be applied to decision making in general. The following diagram gives an overview.