A great deal of traffic has flowed through the internets recently concerning the possibility of Apple's introducing a television, its impact on the TV manufacturing business, and the place such a product would occupy into the iOS and OS X product line: Would it incorporate and replace AppleTV? Will it run apps? Will it have a Retina display? This is all fascinating speculation, but for a typical user it's really rather unimportant. Except that, from the way I see it, one aspect of such a product would have important consequences that would reverberate throughout all of its products, and it is of such fundamental importance that I suspect Apple is trying to get it right before they release a television-type product.

That aspect is the interface required to interact with a television.

Apple's current product lines divide into two OS camps, OS X and iOS, and of course each has its own interface paradigm. OS X relies on a screen pointer manipulated by a mouse or trackpad (or other more exotic devices for the handicapped). iOS uses direct manipulation via natural finger gestures. In fact, Apple has a third interface – the one used by AppleTV, wherein a remote control with a small set of buttons enables a user to select icons or menus and activate them. These three interfaces overlap in some ways, and have been designed to suit the capabilities and limitation of the devices they control. The elemental difference between them all is the way in which the manipulating device (mouse, remote, finger, etc.) is associated with a location on the display: in the OS X and AppleTV worlds, the device is visually disconnected from the display (and is often out of the line of sight entirely), so a pointer or highlight is required to indicate visually to the user where an action will take place. in the iOS world, the manipulation is directly to the display, no pointer is needed, and the action occurs where the user touches.

If we assume that Apple's television is also a computer – in other words, can run apps – using iOS, then none of the interface paradigms currently in use by Apple is sufficient. Further, moving Apple's computers to iOS has the same problem. Here's why.

Let's take a simple example: selecting an item and activating it. (This could be clicking an icon or a menu item.)

• Using OS X, a user moves an on-screen pointer with a mouse or the trackpad to the location of the icon or menu, then clicks or taps to activate it.
• Using iOS, a user touches the icon or menu on the display, either once to select it or twice to activate it.
• Using AppleTV, a user uses the control "ring" to move amongst icons (which then highlight via a border) or menu items (which are highlighted), then presses the selection button in the center of the ring to activate the selection.