Why do we need to name method arguments?

Question

What we have now:

@interface BGTest
-(void)someMethodWithName:(NSString *)methodName tag:(NSUInteger)tag path:(NSURL *)path;
@end

Why not to write:

@interface BGTest
-(void)someMethodWithName:(NSString *) tag:(NSUInteger) path:(NSURL *)
@end

It is stated that ObjC method names should be self documented, and they are. Why to write more if methods can be clearer with lesser code?

I mean, why do we need to name method arguments in interface part, but not just in implementation?

Answer 1

You can't declare a method without a parameter name because in this case there is no way for the compiler to tell if the next token after the type is a parameter name or it is a part of the method name. Let me demonstrate it on this example:

-(void)testMethodWithParam:(int)a secondParam:(int)b;

This is a correct method declaration. Now let's edit it as you propose:

-(void)testMethodWithParam:(int) secondParam:(int);

It's pretty obvious for us what we mean. We can tell, that secondParam: is clearly a part of a method name. But Objective-C allows us to write a column after a space like this:

-(void)testMethodWithParam:(int) secondParam :(int);

And this case (from my point of view) is not that clear anymore. Especially if we remove one more space:

-(void)testMethodWithParam:(int)secondParam :(int);

Now it looks like secondParam is the name of the first parameter, and part of a method name is missing. I guess, to make resolving such issues easier, Objective-C makes us write a complete method specification. I think, they could've changed method declaration format a little bit, so that it allowed such declarations (I like your idea), but it is like this for now.

Still, you can use different variable names in @interface and in @implementation. So this:

@interface TestClass : NSObject
-(void)testMethodWithParam:(int)a secondParam:(int)b;
@end

@implementation TestClass
-(void)testMethodWithParam:(int)firstParam secondParam:(int)secondParam
{...}
@end

is legal.

P.S. By the way, @CodaFi's idea with using underscores as variablesNames could be nice in this case:

@interface TestClass : NSObject
-(void)testMethodWithParam:(int)_ secondParam:(int)_;
@end

The interesting thing is that it compiles, despite the fact that I use a single underscore for both variable names.