Present advancements in the design of computer systems work towards making them “user friendly”; this paper tries to model Human-Computer Communication (HCC) using hardware and cognitive ergonomics. Earlier models, such as the state transition model, sites/modes/traits model, human processor model, and abstract model, are proposed in the literature.

The paper initially defines terms like models, modeling procedures, and modeling tools. The modeling procedure proposed by the author is based on the Channel and Agencies network (CA-net) and the Means/Activity network (MA-net). This approach is a graphic representation of the HCC. Based on these representations, a meta-model of HCC is proposed, consisting of user, designer, and dialogue system. According to the author, users of interactive systems cannot easily model them because of various reasons.

Based on the meta-model, various known models in the literature are discussed: finite state mechanics model; IFIP model of interactive systems; the human processor model; the TOTE model of users based on cognitive processes inside the user; sites/modes/traits model; and the desk-top model.

Until now, generally, the user was left in a passive role when HCCs have been designed. In this paper, a co-operative model between user and designer using the proposed meta-model is suggested. Self-explaining dialogue systems are not produced by the system, but instead by the designer. Such explanations may be inadequate for the user because there is no feedback to the system about the user’s understanding of such messages.

As a first step, the user must learn the semiformal model using CA-net and MA-net. Then, as a next step, the users and designers can develop an acceptable net showing static amd dynamic characteristics. A prototype should be the following step. Efficient implementation is the final step, along with the training of users.

The present model is still an approach. It has to be tested before it can be accepted as a standard approach in designing HCCs. Hence, this paper may prove to be interesting to the designers of human-computer communication systems. The paper concludes with some future directions of research and a good list of references.