Another One-button Dynamic Mode for Dasher: ‘Two-click mode’

The arithmetic-coding-based communication system, Dasher, can be driven by a single switch. In MacKay et al. (2004), we proposed two versions of one-button Dasher, ‘static’ and ‘dynamic’, aimed at a theoretical model of a user who can click with timing precision g, and who requires a recovery time D between clicks. While developing and testing those two versions we invented a third way of using one button, presented in MacKay and Ball (2006) and Mead et al. (2007) and included in Dasher 4 (released 2006) under the name ‘one-button dynamic mode’. This note describes another design for a one-button version of Dasher, which is expected to be more user-friendly. I’ll call this ‘two-click mode’ because the principal gesture made for each navigation action is a pair of clicks. 1 How to use one button in Dasher? As in MacKay et al. (2004), MacKay and Ball (2006), and Mead et al. (2007), we assume that the user controls only the times of presses, not the times of releases. The ‘dynamic one-button mode’ in Dasher version 4 can be likened to a bang-bang controller for a car. The car lurches either drunkenly to the top or to the bottom of the Dasher world, with the button presses being used to switch from the one sort of lurch to the other. Dynamic one-button mode has just one advantage over dynamic two-button mode – it uses only one button; and it has several disadvantages. First, whereas in two-button mode we recommend to users that they click as precisely as possible, identifying the exact moment when the desired target is alongside the fiducial, in one-button dynamic mode we have to give vaguer advice, along the lines of ‘click when the target is close to the edge of the screen – but not too close, because clicking too late would be bad news, requiring an unzooming activity to get the target back on screen’. One-button mode is thus unable to exploit super-accurate presses. The typical flow velocity of the Dasher canvas is bigger in one-button mode than in two-button mode (assuming the two modes are zooming at the same rate), because the point being zoomed in on is at the edge rather than the centre of the display. Another weakness of the one-button mode is that it requires a special action – such as a long press, or a double-click or triple-click, or a press of a second button, to indicate that the user wants to get out of writing mode, for example to correct errors or to stop using Dasher. The motivation for this note is to discuss other ways of using one button that fix these defects. Most important of all, we want to be able to instruct the user to click precisely, and to exploit the large information content of precisely timed clicks.