Scope is an object that refers to the application model. It is an execution context for expressions. Scopes are arranged in hierarchical structure which mimic the DOM structure of the application. Scopes can watch expressions and propagate events.
After evaluating the expression, the $apply method performs a $digest. In the $digest phase the scope examines all of the $watch expressions and compares them with the previous value. This dirty checking is done asynchronously. This means that assignment such as $scope.username=”angular” will not immediately cause a $watch to be notified, instead the $watch notification is delayed until the $digest phase. This delay is desirable, since it coalesces multiple model updates into one $watch notification as well as guarantees that during the $watch notification no other $watches are running. If a $watch changes the value of the model, it will force additional $digest cycle.
The root scope is created during the application bootstrap by the $injector. During template linking, some directives create new child scopes.
During template linking directives register watches on the scope. These watches will be used to propagate model values to the DOM.
For mutations to be properly observed, you should make them only within the scope.$apply(). Angular APIs do this implicitly, so no extra $apply call is needed when doing synchronous work in controllers, or asynchronous work with $http, $timeout or $interval services.
At the end of $apply, Angular performs a $digest cycle on the root scope, which then propagates throughout all child scopes. During the $digest cycle, all $watched expressions or functions are checked for model mutation and if a mutation is detected, the $watch listener is called.
When child scopes are no longer needed, it is the responsibility of the child scope creator to destroy them via scope.$destroy() API. This will stop propagation of $digest calls into the child scope and allow for memory used by the child scope models to be reclaimed by the garbage collector.