Function allocation is a central component of systems engineering and its main aim is to provide a rational means of determining which system-level functions should be carried out by humans and which by machines. Such allocation, it is assumed, can take place early in design life cycle. Such a rational approach to work design sits uneasily with studies of work practice reported in the ACI and CSCW literature. In this paper we present two case studies of work in practice. The "rst highlights the di!erence between functional abstractions used for function allocation decision making and what is required to make those functions work in practice. The second highlights how practice and technology can co-evolve in ways that change the meanings of functions allocated early in design. The case studies raise a number of implications for function allocation. One implication is that there is a need for richer representations of the work context in function allocation methods. Although some progress has been made in function allocation methodologies, it is suggested that the method of Contextual Design might o!er useful insights. A second implication is that there is a need for better theories of work to inform function allocation decision making. Activity Theory is considered as a possible candidate since it incorporates a cultural-historical view of work evolution. Both Contextual Design and Activity Theory challenge assumptions that are deeply embedded in the human factors and systems engineering communities. In particular, that functions and tasks are an appropriate unit of analysis for function allocation.
2000 Academic Press -Men are best at}machines are best at. ? We do not wish to provide a full review of function allocation methods here (but see Dearden, Harrison and Wright, this volume for a full review and de"nition), instead we wish to point out the common characteristics of many methods of function allocation.