This paper deals with estimation of the state of semi-Markov processes. It attempts to give a short survey of relationships existing between the Viterbi and the forward-backward algorithm (used in the context of hidden Markov models, widely used for speech recognition) on one hand, and Kalman filtering and Rauch-Tung-Striebel smoothing on the other hand. It is shown how those algorithms are related. For illustration, an example of a nonlinear hybrid system that can be filtered with a mixed algorithm is introduced.