With databases and information systems playing an increasing role in large scientific research projects, there is a growing stake in understanding how to design a useful information system and in broadening our understanding of what constitutes the scientific work involved in building these systems. Both experience and theory indicate that non-technical considerations, such as management and communication structures, are as important as technical decisions in system development. We examine four case examples of environmental information system development: the Ocean Biogeographic Information System, the Long Term Ecological Research Network, the California Cooperative Oceanic Fisheries Investigation, and SeamountsOnline. We then draw from a wide interdisciplinary literature, including science and technology studies and social informatics, to identify common myths and misconceptions about system development and consider alternatives. Our goal is both to provide a set of concrete models and a theoretical foundation useful to other projects.