This paper examines the current relation between pure and applied research on the comprehension of written information. It finds few points of contact. The suggestion is put forward that greater interaction could be mutually profitable. An information flow among researchers is proposed that starts with applied solutions to practical problems, continues through pure explanations of why these solutions are successful, and so enables the refinement of the original applied solutions. Because such an information flow begins with applied solutions, some of the problems of systematizing the findings of applied research are discussed. There can never be an applied theory of communication that specifies precisely how to design written information. Therefore a proposal is put forward for applying "quality-control" procedures to the preparation of documents. Such procedures indicate that several different kinds of research are necessary for determining the content, optimising the format and evaluating the effectiveness of written communications. The desirability of interactions among those who carry out these different kinds of research is discussed. Finally, the criteria for evaluating research in general are considered. The categorisation of particular studies as either useful or useless is found to be inappropriate. mmm * Thanks are due to Drs John Morton and John Marshall both for providing the incentive to put these thoughts in writing and for their advice and encouragement during the preparation of the paper. Many helpful comments on an earlier draft were received from