An observational study was conducted on a professional designer working on a design project in aerospace industry. The protocol data were analyzed in order to gain insight into the actions the designer used for the development of a solution to the corresponding problem. Different processes are described: from the "simple" evocation of a solution existing in memory, to the elaboration of a "new" solution out of mnesic entities without any clear link to the current problem. Control is addressed in so far as it concerns the priority among the different types of development processes: the progression from evocation of a "standard" solution to elaboration of a "new" solution is supposed to correspond to the resulting order, that is, the one in which the designer's activity proceeds. Short discussions of * the double status of "problem" and "solution," * the problem/solution knowledge units in memory and their access, and * the different abstraction levels on which problem and solution representations are developed, are illustrated by the results.