Semantics of Objective-C Methods

Question

Coming from a Java background, I'm still don't quite understand the semantics of Objective-C methods as opposed to their syntax. Take as an example the following method:

- (NSInteger)numberOfSectionsInTableView:(UITableView *)tableView;

Easy enough. There is a method named numberOfSectionsInTableView which takes a UITableView as a parameter and returns a NSInteger. Now, how about these methods:

- (NSInteger)tableView:(UITableView *)tableView numberOfRowsInSection:(NSInteger)section;
- (CGFloat)tableView:(UITableView *)tableView heightForRowAtIndexPath:(NSIndexPath *)indexPath;
- (UITableViewCell *)tableView:(UITableView *)tableView cellForRowAtIndexPath:(NSIndexPath *)indexPath;

How should I interpret these methods? I've stumbled upon three possible explanations:

1) All three the methods are named tableView and they are overloaded with different parameters. This seems unlikely as people talk about 'calling the numberOfRowsInSection method'.

2) The methods are named numberOfRowsInSection, heightForRowAtIndexPath and cellForRowAtIndexPath, and the tableView is an artifact due to delegation. In that case, what exactly does the tableView part mean syntactically? If the format of a method is '(return)name:parameters, where does tableView fit in?

3) As Objective-C uses message passing, it's wrong to think about methods. Rather think about passing messages to the object directly. In other words, if the object receives the messages named tableView and numberOfRowsInSection, it knows to execute a certain part of the code. If this is the case, does order matter? Is passing numberOfRowsInSection and tableView the same as passing tableView and numberOfRowsInSection?

Answer 1

The name (or selector, in Objective-C's parlance) of a method includes all of the parts of the name. For example, look at these three methods, all of which can coexist in the same class:

- (void)foo;                       // named 'foo'
- (void)foo:(int)bar;              // named 'foo:'
- (void)foo:(int)bar baz:(int)qux; // named 'foo:baz:'

Note the colon in the name of the version of "foo" that takes a parameter: that's crucial. The number of parameters that a method takes is part of its name. This is actually strictly defined, because at times in Objective-C you need to refer to the full selector of a method (using the @selector directive). See the Selectors section of the Objective-C Programming Language document.

Additionally, the order is significant. These two methods are different:

- (void)foo:(int)bar baz:(int)qux; // named 'foo:baz:'
- (void)baz:(int)quz foo:(int)bar; // named 'baz:foo:'

There is no overloading in Objective-C. The compiler won't dispatch to different methods with the same name but different parameter types.

The leading tableView: in your examples is simply a common naming convention used for delegates; all delegate selectors will start with a parameter that specifies the object that is delegating functionality.

Answer 2

In Objective-C, the arguments are interspersed with the method name (called the "selector") instead of passed as a list in parentheses after the name. The full selector is that whole long thing (tableView:numberOfRowaInSection:). The selector is split at the colons and that's where the arguments are placed. There is one argument for each colon. A selector without any colons signifies a method that takes no parameters.

Answer 3

In Objective-C the method names are referred to as signatures and the methods are often called selectors.

The signature of the four selectors in your example are.

numberOfSectionsInTableView:
tableView:numberOfRowsInSection:
tableView:heightForRowAtIndexPath:
tableView:cellForRowAtIndexPath:

cellForRowAtIndexPath:tableView: is not the same as tableView:cellForRowAtIndexPath:

There cannot be more than one selector with the same signature per class.