About Face 2.0(c) The Essentials of Interaction Design

Overlapping windows fail to make it easy to navigate between multiple, running programs; so other vendors continue to search for new ways to achieve this. The virtual desktop of session managers on some Unix-based platforms extends the desktop to six times the size of the visible window. In a corner of the screen are small superimposed, thumbnail images of all six desktop-spaces, all of which can be running different things simultaneously and each of which can have many open windows. You switch between these virtual desktops by clicking on the one you want to make active. In some versions, you can even drag tiny window thumbnails from one desktop to another.

Microsoft braved a double-barreled breach-of-contract and patent infringement lawsuit from Apple in order to add overlapping to Windows 2.0. In all this controversy, the basic problem seemed to have been forgotten: How can the user easily navigate from one program to another? Multiple windows sharing a small screen—whether overlapping or tiled—is not a good general solution (although it certainly may have its occasional uses). We are moving rapidly to a world of full-screen programs. Each application occupies the entire screen when it is "at bat." A tool like the taskbar borrows the minimum quantity of pixels from the running application to provide a visual method of changing the line-up. (Amusingly, this concept is similar to the early days of the Mac with its Switcher, which would toggle the Mac display between one full-screen application and another). This solution is much more pixel-friendly, less confusing to users, and highly appropriate when an application is being used for an extended period of time. In Windows, users have the choice of making their applications full-screen or overlapping at will. Apple is sticking to the overlapping window model alone, at its peril.

Much contemporary software design begins with the assumption that the user interface will consist of a series of overlapping windows, without modes, informed by a global metaphor. The PARC legacy is a strong one. Most of what we know about modern graphical user interface design came from these origins, whether right or wrong. But the well-tempered designer will push the myths aside and approach software design from a fresh viewpoint, using history as a guide, not as a mandate.