Designing innovative computer systems is no longer simply a technical issue. There is now a growing awareness that a grounded understanding of the social and organisational context into which these systems are to be deployed, until recently overlooked, may be critical, particularly in the early phases of design. It is in these stages, known as requirements elicitation, capture or analysis, or more generally Requirements Engineering, that researchers have begun to acknowledge both social and technical concerns in the requirements for systems. This thesis aims to mediate between a detailed appreciation of the social organisation of the workplace and the technical structure of information technology. In order to do this, we explore the relevance of recent developments in the social sciences, principally ethnomethodology and interaction analysis, for providing an alternative analytic orientation for requirements capture. In particular, we outline the principal characteristics of an approach that takes into account the details of the moment-to-moment production of work activities and communicative practices. It is argued that attendance to such interactional features will improve requirements practice, and hence will ultimately lead to more sensitive designs for supporting collaborative work. To illustrate this approach, we take as an example the development of particular technologies for a complex work setting - financial trading rooms. Results of the analysis are then used to discuss requirements for systems to support trading. In particular, we consider how issues emerging from this analysis, developed from an ethnomethodological orientation, could inform requirements analysis. In this regard, we investigate approaches to modelling interactional resources, including the use of formal notations developed for sequential and communicating processes, and provide requirements analysts with sensitivities by which to consider naturalistic settings. This forms the basis for mutually dependent investigations: on the system design side, drawing on the model to allow consideration of conflicts introduced by technological choices; and on the social science side, providing an agenda for renewed investigation into the domain. We conclude by discussing the pre-requisites necessary so that approaches in this thesis could be integrated within the software development process.