Difference between self.var and simply var

Question

What is the difference between using self.var vs. just var in an Objective-C class? Are there benefits or dangers to one or the other?

Answer 1

self.var calls the property for var. Behind the scenes, Objective-C automatically generates a getter for properties (or you can make one yourself, if so inclined), so self.var uses that getter. Plain var accesses the instance variable directly (i.e., it doesn't go through the getter to get the value).

Answer 2

foo = self.var;
self.var = foo;

is conceptually identical to

foo = [self var];
[self setVar: foo];

So using dot notation, you are really sending messages to self.

foo = var;
var = foo;

is conceptually the same as

foo = self->var;
self->var = foo;

So not using dot notation to access an instance variable is the same as treating self as a pointer to a C struct and accessing the struct fields directly.

In almost all cases, it is preferable to use the property (either dot notation or message sending notation). This is because the property can be made to automatically do the necessary retain/copy/release to stop memory leaks. Also, you can use key value observing with a property. Also subclasses can override properties to provide their own implementation.

The two exceptions to using properties are when setting an ivar in init and when releasing it in dealloc. This is because you almost certainly want to avoid accidentally using a sub class override in those methods and you don't want to trigger any KVO notifications.