Abstrakt
Restarting automata were introduced as a tool for modelling a method of checking correctness of sentences called analysis by reduction. The first ten years of research on restarting automata brought out their interesting properties and introduced several extensions to the original model.
Here we concentrate on recent developments in the opposite direction - from complex to simpler models. Such development aims to find automata models which are somewhat limited in their power but have other desired properties, like e.g. simpler definition, more effective recognition or learnability.
All these properties are crucial for possible applications of restarting automata.