The way we handwrite characters is a deeply individual matter, as bank tellers who ask for your signature and graphologists who claim to be able to study your personality from your handwriting know well. The handwriting samples that they work with are static, in the sense that they consider the trace left behind well after the signature is formed, and thus are at one remove from the person who actually did the original writing. In this sense, any attempts to identify an individual, let alone to claim to reconstruct aspects of their personalities, have the flavor of archaeological digs. What if we could use the online time course of the formation of a signa-ture? Would we not see things as the signature unfolds in time that could not be observed in the static image? Could we see, for example, when a person was nervous, in a hurry, suffering from the onset of Parkinsonism, or rejoicing in a state of profound tranquillity and peace? Surely we could discover new ways by which a handwriting sample characterizes a specific individual, and perhaps use this to make forgery harder than it is now. In this chapter we use what we call a dynamic model for handwriting. We demonstrate how the model can be fitted to the writing of a particular individual using repeated samples of their printing. We also investigate how well the model separates one person from another. Our first task, however, is a brief and nontechnical account of some simple dynamic models. Those familiar with differential equations may well be happy to skip ahead, but many readers will find this next section important.