Rapid serial visual presentation: An approach to design

摘要

A familiar action in the physical world is the riffling of a book's pages in order to gain, in a very short time, some idea about its content. That action is appropriately termed rapid serial visual presentation (RSVP). A computational embodiment of RSVP can provide such a variety of possible image presentation modes and flexibility of control as to offer a powerful tool to support a wide range of tasks involving the presentation of a collection of images. Those images can be conventional pictures [e.g. the Mona Lisa), the current image of a television channel, a frame of a video recording, the layout of a letterhead or page, a line drawing, a logo or the layout and shape of some text as in a motorway sign. Although our understanding of RSVP is not complete, sufficient experimental evidence and application experience is now available to lead to a useful set of guidelines for the interaction designer interested in exploiting its potential. These guidelines relate to six specific areas of design: the design style, the rate at which images are presented, image size on screen, the value of user control, image density on screen and image overlap. In this paper those guidelines are presented and justified by experimental evidence, much of it, unsurprisingly, emerging from the nature of human visual processing. In formulating the guidelines we fully acknowledge the importance, in a practical application, of aesthetic considerations; likewise, we take account of the need to minimise constraints on the creativity of the interaction designer. While RSVP addresses a specific class of tasks, we suggest that the principles discussed in this paper will have a broader relevance to image presentation techniques.