Objective-c – What’s are the differences between @dynamic and @synthesize

These are good questions, and its great to see that you're doing this research and seem concerned with learning how to "do it right" instead of just hacking it together.

First, I agree with the previous answers which focus on the importance of putting data in model objects when appropriate (per the MVC design pattern). Usually you want to avoid putting state information inside a controller, unless it's strictly "presentation" data.

Second, see page 10 of the Stanford presentation for an example of how to programmatically push a controller onto the navigation controller. For an example of how to do this "visually" using Interface Builder, take a look at this tutorial.

Third, and perhaps most importantly, note that the "best practices" mentioned in the Stanford presentation are much easier to understand if you think about them in the context of the "dependency injection" design pattern. In a nutshell, this means that your controller shouldn't "look up" the objects it needs to do its job (e.g., reference a global variable). Instead, you should always "inject" those dependencies into the controller (i.e., pass in the objects it needs via methods).

If you follow the dependency injection pattern, your controller will be modular and reusable. And if you think about where the Stanford presenters are coming from (i.e., as Apple employees their job is to build classes that can easily be reused), reusability and modularity are high priorities. All of the best practices they mention for sharing data are part of dependency injection.

That's the gist of my response. I'll include an example of using the dependency injection pattern with a controller below in case it's helpful.

Example of Using Dependency Injection with a View Controller

Let's say you're building a screen in which several books are listed. The user can pick books he/she wants to buy, and then tap a "checkout" button to go to the checkout screen.

To build this, you might create a BookPickerViewController class that controlls and displays the GUI/view objects. Where will it get all the book data? Let's say it depends on a BookWarehouse object for that. So now your controller is basically brokering data between a model object (BookWarehouse) and the GUI/view objects. In other words, BookPickerViewController DEPENDS on the BookWarehouse object.

Don't do this:

@implementation BookPickerViewController

-(void) doSomething {
   // I need to do something with the BookWarehouse so I'm going to look it up
   // using the BookWarehouse class method (comparable to a global variable)
   BookWarehouse *warehouse = [BookWarehouse getSingleton];
   ...
}

Instead, the dependencies should be injected like this:

@implementation BookPickerViewController

-(void) initWithWarehouse: (BookWarehouse*)warehouse {
   // myBookWarehouse is an instance variable
   myBookWarehouse = warehouse;
   [myBookWarehouse retain];
}

-(void) doSomething {
   // I need to do something with the BookWarehouse object which was 
   // injected for me
   [myBookWarehouse listBooks];
   ...
}

When the Apple guys are talking about using the delegation pattern to "communicate back up the hierarchy," they're still talking about dependency injection. In this example, what should the BookPickerViewController do once the user has picked his/her books and is ready to check out? Well, that's not really its job. It should DELEGATE that work to some other object, which means that it DEPENDS on another object. So we might modify our BookPickerViewController init method as follows:

@implementation BookPickerViewController

-(void) initWithWarehouse:    (BookWarehouse*)warehouse 
        andCheckoutController:(CheckoutController*)checkoutController 
{
   myBookWarehouse = warehouse;
   myCheckoutController = checkoutController;
}

-(void) handleCheckout {
   // We've collected the user's book picks in a "bookPicks" variable
   [myCheckoutController handleCheckout: bookPicks];
   ...
}

The net result of all this is that you can give me your BookPickerViewController class (and related GUI/view objects) and I can easily use it in my own application, assuming BookWarehouse and CheckoutController are generic interfaces (i.e., protocols) that I can implement:

@interface MyBookWarehouse : NSObject <BookWarehouse> { ... } @end
@implementation MyBookWarehouse { ... } @end

@interface MyCheckoutController : NSObject <CheckoutController> { ... } @end
@implementation MyCheckoutController { ... } @end

...

-(void) applicationDidFinishLoading {
   MyBookWarehouse *myWarehouse = [[MyBookWarehouse alloc]init];
   MyCheckoutController *myCheckout = [[MyCheckoutController alloc]init];

   BookPickerViewController *bookPicker = [[BookPickerViewController alloc] 
                                         initWithWarehouse:myWarehouse 
                                         andCheckoutController:myCheckout];
   ...
   [window addSubview:[bookPicker view]];
   [window makeKeyAndVisible];
}

Finally, not only is your BookPickerController reusable but also easier to test.

-(void) testBookPickerController {
   MockBookWarehouse *myWarehouse = [[MockBookWarehouse alloc]init];
   MockCheckoutController *myCheckout = [[MockCheckoutController alloc]init];

   BookPickerViewController *bookPicker = [[BookPickerViewController alloc] initWithWarehouse:myWarehouse andCheckoutController:myCheckout];
   ...
   [bookPicker handleCheckout];

   // Do stuff to verify that BookPickerViewController correctly called
   // MockCheckoutController's handleCheckout: method and passed it a valid
   // list of books
   ...
}

The last two are identical; "atomic" is the default behavior (note that it is not actually a keyword; it is specified only by the absence of nonatomic -- atomic was added as a keyword in recent versions of llvm/clang).

Assuming that you are @synthesizing the method implementations, atomic vs. non-atomic changes the generated code. If you are writing your own setter/getters, atomic/nonatomic/retain/assign/copy are merely advisory. (Note: @synthesize is now the default behavior in recent versions of LLVM. There is also no need to declare instance variables; they will be synthesized automatically, too, and will have an _ prepended to their name to prevent accidental direct access).

With "atomic", the synthesized setter/getter will ensure that a whole value is always returned from the getter or set by the setter, regardless of setter activity on any other thread. That is, if thread A is in the middle of the getter while thread B calls the setter, an actual viable value -- an autoreleased object, most likely -- will be returned to the caller in A.

In nonatomic, no such guarantees are made. Thus, nonatomic is considerably faster than "atomic".

What "atomic" does not do is make any guarantees about thread safety. If thread A is calling the getter simultaneously with thread B and C calling the setter with different values, thread A may get any one of the three values returned -- the one prior to any setters being called or either of the values passed into the setters in B and C. Likewise, the object may end up with the value from B or C, no way to tell.

Ensuring data integrity -- one of the primary challenges of multi-threaded programming -- is achieved by other means.

Adding to this:

atomicity of a single property also cannot guarantee thread safety when multiple dependent properties are in play.

Consider:

 @property(atomic, copy) NSString *firstName;
 @property(atomic, copy) NSString *lastName;
 @property(readonly, atomic, copy) NSString *fullName;

In this case, thread A could be renaming the object by calling setFirstName: and then calling setLastName:. In the meantime, thread B may call fullName in between thread A's two calls and will receive the new first name coupled with the old last name.

To address this, you need a transactional model. I.e. some other kind of synchronization and/or exclusion that allows one to exclude access to fullName while the dependent properties are being updated.