On present-day devices much effort is devoted to develop state-of-the-art diagnostics with a continuous drive towards higher accuracy, better spatial and temporal resolution and more diagnostic channels. Diagnostic innovations often lead to better physics insight and they are often a driver for improving theoretical models. In future fusion devices the operation of diagnostics is strongly limited by the hostile environment. In ITER many of the presently used diagnostics are still marginally applicable, but in DEMO the amount of diagnostics that can be used is severely constrained - at the one hand because of the tough environmental effects and at the other hand because access to the machine will be limited. Theoretical modeling will be very important for DEMO diagnostics. Firstly, simulations based on synthetic diagnostics should lead to the optimum choice of diagnostics and secondly, theoretical modeling should complement the rather sparse diagnostic data set that can be obtained in DEMO.