Starting with the focal question, "what should students know about technology?" we describe and illustrate a way of designing educational technology that is strongly informed by empirical studies of how students actually understand and use a technology. We also have theoretical aspirations in developing what we hope to be general principles that can, along with empirical data, orient design.

The type of technology used to illustrate this design methodology is scientific visualization software, in which spatially distributed data is given form as adjustable and often highly suggestive visual displays. Our primary contention is that what students need to know about this software is precisely those aspects of it that define it as a system of representations. More generally, we advocate representation as an important instructional target, and we examine what students know that can be enhanced by appropriate technology and learning activities.