A property-access expression is equivalent to a message expression:
[object setTexture:tex];
A property declaration is equivalent to one (readonly
) or two (readwrite
/default) instance-method declarations. Keywords like retain
tell the compiler how to implement the method if you tell it to do so (@synthesize
).
However, you can skip the property declaration and declare the methods directly, as shown in your question. You can't synthesize their implementations, since you need a property declaration for that (otherwise, it wouldn't know what memory-management policy to use: assign
, retain
, or copy
), but you can always implement the methods yourself.
Then, even though you declared and implemented the methods yourself, since property-access syntax and message syntax are equivalent to each other, you can use the methods whichever way you want: With a message expression, or with a property-access expression.
Some would consider it bad form, though, to use property access expressions on anything but a formal @property
(e.g., myString.length
or myArray.count
or myView.frame
). It definitely is bad form to use a property-access expression to send a message that doesn't access any kind of property; foo.retain.autorelease
, for example, is bad and wrong: It reeks of trying to pretend you're programming some other language than Objective-C.
Incidentally, a property and a variable are unrelated. A @property
will ordinarily be backed by an instance variable, but this is not required: You could store the property's value inside another object, or convert it to and from some other format, or both. Likewise, accessing a property (which is an accessor message) and accessing an instance variable (which is just accessing a variable, nothing more) are very different.