"It talks here about storms and hurricanes. Tell me, is every hurricane a storm?" "It says the name of this storm was Ivanka. Does every storm have a name?" "It says this hurricane evolved into a tropical storm. Should we treat the hurricane and the tropical storm as the same object, with changing properties during its life cycle, or should we treat them as different objects?" The questions will be difficult and challenging, and I don't know how long it will take, but asking such questions is the only way I know to do it. Many of the questions don't have a right answer, by the way: we're building a model, and the test of a model is not whether it is correct, but whether it is useful. I did this with a group of people from the BBC once who were modelling sports results. It was a fascinating exercise; it was even more interesting to hear ten years later how the BBC had built a data model for the 2012 Olympics. Michael Kay Saxonica |