PROCESS is a reactor systems code – it assesses the engineering and economic viability of a hypothetical fusion power station using simple models of all parts of a reactor system. PROCESS allows the user to choose which constraints to impose and which to ignore, so when evaluating the results it is vital to study the list of constraints used. New algorithms submitted by collaborators can be incorporated – for example safety, first wall erosion, and fatigue life will be crucial and are not yet taken into account. This paper describes algorithms relating to the engineering aspects of the plant. The toroidal field (TF) coils and the central solenoid are assumed by default to be wound from niobium-tin superconductor with the same properties as the ITER conductors. The winding temperature and induced voltage during a quench provide a limit on the current density in the TF coils. Upper limits are placed on the stresses in the structural materials of the TF coil, using a simple two-layer model of the inboard leg of the coil. The thermal efficiency of the plant can be estimated using the maximum coolant temperature, and the capacity factor is derived from estimates of the planned and unplanned downtime, and the duty cycle if the reactor is pulsed. An example of a pulsed power plant is given. The need for a large central solenoid to induce most of the plasma current, and physics assumptions that are conservative compared to some other studies, result in a large machine, with a cryostat 36 m in diameter. Multiple constraints, working together, restrict the parameter space of the optimised model. For example, even when the ratio of operating current to critical current in the TF coils is increased by a factor of five, the total coil cross-section decreases only a little, because of the need for copper stabiliser, insulation, and structural support. The result is that the plasma major radius hardly changes. It is these surprising results that justify the development of systems codes.