Btukfyl

Let's say I have to store customer information, and to manage two-way binding I'll use $scope here.
So my doubt here is, which approach is better?

$scope.firstname = "foo";

$scope.lastname = "bar";

$scope.cellno = "1234567890";

$scope.email = "foobar@example.com";

OR

$scope.customerDetailsObj = {};

$scope.customerDetailsObj.firstname = "foo";

$scope.customerDetailsObj.lastname = "bar";

$scope.customerDetailsObj.cellno = "1234567890";

$scope.customerDetailsObj.email = "foobar@example.com";`

I've been wondering about this because I have a large angular application and sometimes the scope watchers count goes beyond 1500. I'm using a chrome extension to see watchers count.

Please share your views. Thank You.

In my opinion and as-per angularjs guide first approach (binding to properties directly) will be more efficient. Excerpt from the guide:

Dirty checking can be done with three strategies: By reference, by
collection contents, and by value. The strategies differ in the kinds
of changes they detect, and in their performance characteristics.

• Watching by reference (scope.$watch (watchExpression, listener)) detects a change when the whole value returned by the watch expression
  switches to a new value. If the value is an array or an object,
  changes inside it are not detected. This is the most efficient
  strategy.


• Watching collection contents (scope.$watchCollection (watchExpression, listener)) detects changes that occur inside an
  array or an object: When items are added, removed, or reordered. The
  detection is shallow - it does not reach into nested collections.
  Watching collection contents is more expensive than watching by
  reference, because copies of the collection contents need to be
  maintained. However, the strategy attempts to minimize the amount of
  copying required.


• Watching by value (scope.$watch (watchExpression, listener, true)) detects any change in an arbitrarily nested data structure. It is the
  most powerful change detection strategy, but also the most expensive.
  A full traversal of the nested data structure is needed on each
  digest, and a full copy of it needs to be held in memory.

Talking only about performance, I think option 1 (having plain variables) it's more efficient.

Watching plain properties

the $watch() function, by default only checks object reference equality. So within each $digest it will check to see if the new and old values pointing to the same object.

Watching objects and arrays

If you want to $watch() an object, you have to switch to the deep watch and that means that within each $digest, it will check the whole tree to see if the structure or values are changed. So, bigger the object, more expensive the check.

$watchGroup(), an alternative to objects/array watchers

If you have a group of related properties like ie a person (name, surname, age), you could define it like plain properties, and use $watchGroup(), that works on the same way like a common $watch() but taking a list of properties to watch rather than one.