This article interrogates assumptions about the on-line/off-line distinction in internet research. Much research has presumed "virtuality" and a strong division between the on-line and the off-line as a methodological and analytical given and starting point for research. We argue that when actual internet use is approached from an ethnographic perspective, this distinction turns out to be a highly contingent social accomplishment: some actors deploy great skill to establish a sharp on-line/off-line distinction in their internet practices; for others, it is meaningless or undesirable and the distinctions is not established in reality. A comparison of cybercafes in Trinidad is used to demonstrate this variability and the wide range of different on-line/off-line relationships that are in fact observed. We relate this material to the overall issue of establishing "contexts" in ethnographic research.
cybercafes; ethnography; internet; virtuality