Groupware use can be described as a process of social (human-computer-human) interaction. For example, small groups can use a group-based geographic information system (GIs) to share maps and decision tables during a discussion about selection of sites for salmon habitat improvement in Seattle, Washington. Empirical research about groupware use is intended to improve our understanding of the dynamics of the process, as well as improve our understanding of the development requirements for information technology. Gaining a detailed understanding of the human-computer-human interaction process requires reasonably unobtrusive observation-for example, using video cameras to capture and replay the ebb and flow of interaction. From each replay of videotape we can abstract a different research view, hence characterize the ebb and flow of interaction from a different perspective, giving us deeper insight into the interaction. Interpreting and synthesizing the raw observations to make sense of "what went on during interaction" can be accomplished through the use of interaction coding systems. In this article, we report on the development of three interaction coding systems that were created for studying the use of a group-based, research prototype GIs software, called Spatial Group Choice. We wrote this article to help researchers compare approaches to the development of coding systems and compare the value of their use. Despite previous use of coding systems by others, there are no detailed reports in the literature of how researchers devised their coding systems. We discuss in detail the process of creating and using such coding systems, describing the advantages and disadvantages of performing interaction coding to foster an understanding of group dynamics in different settings and for designing new groupware.