It's good to remember Jeff Raskin's theorem "intuitive = familiar". If you come from the C++ world, Objective-C is going to look weird. If you come from the Smalltalk world, not all that weird.
Also, syntax is the worst place to start learning Objective-C imho 'cos there can be a lot of it to learn if you come from a non-smalltalk world. The best place I've found is to dive straight into the core runtime function objc_msgSend. Once you grok that, and see that you could write down the core runtime in a handful of C functions in an hour or so, everything else -- classes, categories, protocols, delayed invocation, remote invocation, proxies, posing, key-value coding -- finds a "natural" slot in your brain. As a bonus, as you get to the more "advanced" features unique to the system (relative to, say, C++) such as key-value coding, you see how the dynamism of the language plays to support all of that. (Disclaimer: Yes, this is how I got it, but I don't know whether it is generally good way to approach it, though I'd recommend it. Maybe I should write a tutorial on it.)
If you start by looking at the syntax and going "ugh", you'll be missing all the neat ideas in the system ... including, imo, the older memory management system that many complain about. I've, for example, used the "auto release pool" idea in C++ to relieve colleagues of the need to think about ownership and lifetime in relatively isolated corners of a system while considerably simplify api design and staying performant. If you're looking for "predictable performant garbage collection", this is a reasonable design candidate.
Also, syntax is the worst place to start learning Objective-C imho 'cos there can be a lot of it to learn if you come from a non-smalltalk world. The best place I've found is to dive straight into the core runtime function objc_msgSend. Once you grok that, and see that you could write down the core runtime in a handful of C functions in an hour or so, everything else -- classes, categories, protocols, delayed invocation, remote invocation, proxies, posing, key-value coding -- finds a "natural" slot in your brain. As a bonus, as you get to the more "advanced" features unique to the system (relative to, say, C++) such as key-value coding, you see how the dynamism of the language plays to support all of that. (Disclaimer: Yes, this is how I got it, but I don't know whether it is generally good way to approach it, though I'd recommend it. Maybe I should write a tutorial on it.)
If you start by looking at the syntax and going "ugh", you'll be missing all the neat ideas in the system ... including, imo, the older memory management system that many complain about. I've, for example, used the "auto release pool" idea in C++ to relieve colleagues of the need to think about ownership and lifetime in relatively isolated corners of a system while considerably simplify api design and staying performant. If you're looking for "predictable performant garbage collection", this is a reasonable design candidate.