The contributions of this paper are in two main areas. The first is an "integrated" approach to the development of a practical adaptive control algorithm. In particular, we bring many existing ideas together and explore the effect of various design parameters available to a user. We also extend the theory in the following areas: we show how the problem of pole-zero cancellation in the estimated model can be overcome by running parallel estimations and we show how asymptotic tracking can be achieved in the presence of unmodelled dynamics. A feature of our presentation is that we treat continuous and discrete systems in a unified framework.