Public computer systems (PCS) are systems designed for use at the interface between organizations and their clients. PCS intervene in client-organization relations; the questions discussed in this thesis are what role they play in the client-organization encounter, what role they could and should play, and what theories might be available to guide the development of such systems. While the fields of Human-Computer Interaction (HCl) and Computer Supported Cooperative Work (CSCW) focus on (one) man - (one) machine interaction and computer-mediated interaction in small groups respectively, this study focuses on computer-mediated or computer supported interaction between organizations and individual clients. This focus is chosen because the emerging electronic markets make social dimensions not covered by HCl and CSCW relevant to information systems design. While PCS and electronic markets have so far been studied mainly from technical and economic perspectives, this study takes a communications perspective. The nature of actual PCS implementations is studied with respect to changes in the communicational style of the client-organization encounter. The relations between organizations and clients concern not only the actors directly involved. They also affect the general social climate, the societal dialogue, particularly so when public agencies are concerned. What does it mean to change the arenas for the societal dialogue? One candidate theory pertinent to PCS impact on the societal dialogue, participatory theory as of Rousseau, J. S. Mill and Cole, is investigated. Based on this theory, a model for participatory information systems (PARTIS) is developed. This model is proposed as a base for PCS design. The Feedback Learning Strategy (FLS) is then outlined as a method for the design and redesign of the computerized parts of a PARTIS. The strategy is based on built-in functions aimed at encouraging and facilitating participation. An example of a FLS prototype system, the LiveBetter, is introduced and discussed. Conclusions are in short: • PCS are important ingredients in societal information systems, and must therefore be designed to support communication according to democratic principles. • To be effective, PCS must be well integrated with organizational structures. • Participatory information systems must include a redesign forum that supports discovery, fair interest articulation, multiple descriptions, equality, and conflict resolution. • The specific design of those facilities must be done in each case. This is a challenge for systems design which I call conversation management; it is more an organizational challenge than a technical one. • Computerized tools may be used to enhance the participatoriness of the systems. / <p>Diss. Umeå : Univ.</p> / digitalisering@umu