gistfile1.txt

- (void)method1:(id)somethingElse {
    NSString *identifier = [NSString stringWithFormat:@"%@.%@", [SomeClass someMethod], somethingElse];
    MyObject *obj = [[MyObject alloc] initWithIdentifier:identifier];

    …
}

- (void)method2:(id)somethingElse {
    NSString *identifier = [NSString stringWithFormat:@"%@.%@", [SomeOtherClass someOtherMethod], somethingElse];
    MyObject *obj = [[MyObject alloc] initWithIdentifier:identifier];

    …
}

- (void)method3:(id)somethingElse {
    NSString *identifier = [NSString stringWithFormat:@"%@.%@", [YetAnotherClass anotherMethodWithSomethingElse:somethingElse], somethingElse];
    MyObject *obj = [[MyObject alloc] initWithIdentifier:identifier];

    …
}

If you apply DRY to the design decisions rather than the code detail, an answer everyone agrees on may be more forthcoming. If the format of those strings is essential to the correct operation of your code, I'd wrap that up in a utility method with a name documenting that design decision. If it's just an opaque identifier the above is fine.

That probably makes the most sense. I mean I could see both sides of the argument with it, I just struggled to justify why you'd do either other than "just because".