Most analyses of how humans use artifacts, and interactive software in particular, have a strong tendency to assign 1:1 correspondences between goals and methods: to see software as supporting one task, users as having one way of executing a task, one thing to learn when learning a command, and one source for discovering the information. In fact this is a rare case, and multiplicity of goals, methods, information needs, and information resources is the rule even in simple software. How this causes problems for the design and testing of user interfaces can be illustrated by examples from a wide range of domains and levels of design, including studies on learning by exploration, the effect of machine delays on user strategies, the Icamability of icon sets, evaluation studies of Computer Assisted Learning, and an analysis of the concept of affordance. Such plurality can he a source of robustness for the performance of interfaces: it is a problem main for analysis and HCI research, which struggle to account for the frequent case of high average performance levels mixed with a few residual problems. To address this plurality, we must extend our analyses to cover sets of alternative methods for tasks rather than single user procedures, and perhaps draw on concepts such as Activity Theory to address users' mental organisation of such plurality.