A resolve is simply a value that is passed to the controller upon instantiation (which are used like an injected value). The neat thing about them is that if the value returned is a promise, the view/controller won't load until the promise has resolved.
The way you use them is by adding a resolve key to your route state, and returning the object you want injected into your controller (also naming it). For example:
.state('example', { url: '/page', templateUrl: 'sometemplate.html', controller: 'SomeCtrl', resolve: { injectionName: function(){ // return a value or promise here to be injected as injectionName into your controller } } }); Then inside your controller you simply add the resolve name to the controller injected values:
.controller('SomeCtrl', function($scope, injectionName){ // do stuff with injectionName }); Just note that if you do return a promise, the value that is injected is the result of the promise (not the promise itself). Also note that if the promise errors the view/controller will not load, and an error will be thrown. As @koox00 commented, this error will fail silently unless $stateChangeErrorError is handled (usually in your apps primary run() function).
So why would you use this? Well if not inferred from above, you do this usually when you want your view/controller to wait until some async process has completed before loading a particular state. This saves you from creating loaders or loading processes for every single view/controller, moving it to a simple definition of what needs to be loaded.