As a software developer I'm usually forced to answer this question posed by the management, end-users or clients. They all want to know the future, and they want me to respond with a reliable estimate, because their businesses depend on my ability to meet deadlines. As we all know, development time estimation is hard, and even though there are many methodologies that help team leads cope with them, expectations are usually not fulfilled.
But wait a minute; is this problematic question valid? Does it even make any sense? Not really. Estimations are known to be useless if they don't include probabilities. In this talk we will learn how the question can be reformulated in more appropriate ways. We will see a simple stochastic model that simulates a Dev Team and provides insightful estimations and forecasts―the talk may also be seen as an introduction to Probability Distributions and Monte Carlo simulations in a domain familiar and interesting to the working smalltalker.
Chief Technologist at Caesar Systems and President of FAST, the Argentine Smalltalk Foundation. Recent talks: Software Development (Smalltalks 2007), Extreme Validation (ESUG 2007), Making the Most of User Changes (ESUG 2008), Instance Specific Behavior for the Working Smalltalker (ESUG 2009), Homological Algebra in Smalltalk (Smalltalks 2009).