State and decomposition

A conveyancing matter file holds hundreds of fields: the parties, the property, the price, the searches, a dozen rounds of enquiries, every email, the billing ledger and the file notes. When a question about that file comes back wrong, the instinct is to blame the model. More often the request was badly shaped: too much state, a question that hid three judgments, or a date comparison the model was never good at. This lesson is about shaping requests so that Jev answers the questions it is good at, and your code does the rest.

In Your first Jev request you sent a state and a set of questions. In Choice, Score and Noul you learned to write each question type. Here the focus moves to the two things around the questions: what goes into state, and how a broad judgment becomes several narrow ones.