How Buttons Work

Buttons follow the target-action design pattern. A button is a user interface object that sends an action message to a target when clicked. For more information on this design pattern, see Target-Action in Concepts in Objective-C Programming.

Most of the button’s work is handled by the NSButtonCell class. An NSButtonCell instance sends its action message to its target once if its view is clicked and it gets the mouse-down event, but can also send the action message continuously as long as the mouse is held down with the cursor inside the button cell. The button cell can show that it’s being pressed by highlighting in several ways—for example, a bordered button cell can appear pushed into the screen, or the image or title can change to an alternate form while the button cell is pressed.

An NSButtonCell object must work with an instance of a subclass of NSControl. If you need one button, such as a push button, use an NSButton object that contains a single NSButtonCell instance. If you need a group of related buttons, such as a group of switches or radio buttons, use an NSMatrix object that contains several NSButtonCell instances.

NSButton and NSMatrix both provide a control view. However, while NSMatrix requires you to access the NSButtonCell objects directly, most of NSButton’s methods are “covers” for identically declared methods in NSButtonCell. (In other words, the implementation of the NSButton method invokes the corresponding NSButtonCell method for you, allowing you to be unconcerned with the NSButtonCell object’s existence.) The only NSButtonCell methods that don’t have covers relate to the font used to display the key equivalent, and to specific methods for highlighting or showing the NSButton’s state (these last are usually set together with NSButton’s setButtonType: method).