Learning is important in software projects, so how do we maximize learning? In Tom and Mary Poppendieck’s Lean Development: An Agile Toolkit … they say something about the level of learning in scientific experiments – that it peaks when the success rate is about 50 percent. I don’t remember whether they quoted some study about this, but it sure feels right to me that a balance between success and failure would increase learning. You need some friction.

If we accept that learning is very important, it becomes apparent that whatever you do to learn about the problem to be solved and its context, it must be something that will give you good-quality feedback as quickly as possible whenever you pose a hypothesis. From where would that feedback come? Obviously from people with the best knowledge of the problem and its context – people that are experts within the domain. Also, the future users of the system might not be domain experts, but they should be included because they are experts on themselves; they might not be able to say what the system should be like, but they will definitely be able to say whether the actual system feels right to them – if they understand it and feel that it helps them do whatever it is that they do.

If we want to maximize learning, we will have to do whatever gives us the best feedback from the experts until we have learned enough to build the system. I don’t know about you, but I can’t think of anything but actually building the system and having the domain experts try it out. There will always be room for different interpretations (between developers and experts) in documents, specifications, diagrams, mock-ups, interface dummies, and so forth – and doing richer prototypes comes closer and closer to the actual thing.

This is what Extreme Programming (XP) advocates, but my point wasn’t to argue for doing incremental releases and involving experts continuously, but to ask how to maximize learning. Googling for that phrase turns up a few articles that might be interesting; for instance: A Guide to Maximizing Learning in Small Groups by Igor Kudyszyn.