Convergence, viewed as the union of disparate technical solutions, is frequently proposed as a way of maximising value for end users: reducing the number of distinct technologies users have to purchase, learn and use. Yet few empirical studies of use and convergent technology have been reported. Though convergence as a catchphrase has had currency for over a decade now, a tension remains between those who argue for strong-specific solutions, i.e. carefully targeted ‘information appliances’, and those who prefer weak-general approaches, the ICT equivalent of the Swiss army knife. We describe the dynamic nature of the trade-off between usability and functional complexity that is inherent in convergence. We contrast current products as examples of ‘Convergence-by-Design’ with empirical data that illustrates users’ practices in both combining previously non-converged solutions and separating previously converged solutions. We conclude that effective resolution of the usability/complexity trade-off involves both designers and users, and that sub-optimal resolution can trigger the workarounds that we call ‘Convergence-in-Use’ and ‘Divergence-in-Use’.