It has become a commonplace design slogan, though seemingly attended to more in the breach than in practice, that distributed system design needs to attend to the sociality of work. Part of this involves understanding how technologies become embedded and are understood within day-to-day working practices. This paper presents some results from a long term empirical investigation of computer systems in use in financial services. It reports on an ethnographic study of the ‘sociality of work’ of a major bank undergoing massive and simultaneous change - organisational, technological, and cultural. The focus for this paper is on ‘legacy’ issues and their impact on everyday working. It suggests that ‘legacy’ is not merely a technological problem facing organisations with ageing mainframes and dated software, pointing instead to a far wider organisational purchase and relevance.

This case study employs ethnomethodologically informed ethnographic methods to highlight how computers, procedures and rules are enmeshed in a system of everyday working - thereby illustrating those ‘human factors’ which most closely pertain to system design, whether 'system' in this context is restricted to technology, or is expanded to include organizational processes. The paper has a particular concern with notions of ‘process’ and standardisation’ central to so many attempts to reconfigure the workplace and illustrates how an understanding of ‘process’ and the implementation of ‘standardisation’ are locally achieved accomplishments of parties to the work such that even when standard procedures are implemented, local knowledge and skills are both persistent and necessary to the very flow of work that is presumed to result from process design or redesign.

The central point of the paper is that an appreciation of legacy needs to move away from a purely technological stance to admit the importance and impact of organisational issues. Empirical examples are used to illustrate how the rapidly changing nature of commercial and organisational life means that legacy issues can arise relatively soon even after the introduction of comparatively new technologies. Furthermore, understanding ‘legacy’ and its impact on business 'processes' and everyday working requires a nuanced view of various factors, including working practice, communication and control problems, and indeed any number of complex articulations of structure, process, technology, and 'situated' knowledge. Straightforward process approaches, despite their attraction to system modellers, are unlikely to discover or appreciate the various interactional subtleties involved in everyday work. Consequently, any attempt to resolve legacy issues will depend for its success not only on finding the right answers but upon deciding the right questions to be asked in the first place. We suggest that the descriptive and analytic techniques to be found in ethnographic studies provide alternative ways of asking questions, or ways of 'respecifying the problem'.