As with any other user interface, the only way to see whether a mobile interface is really any good is to use it. You can apply Fitts' Law and numerous other techniques to see whether the interface should be usable, but only putting it in the hands of an actual human will tell you if you've succeeded. Rules like these are approximations of human behaviorsometimes very rough approximations. Sometimes they're just plain wrong, or they make assumptions that are no longer applicable to your system.
Most mobile platform developer kits come with simulators for the target. These simulators are great for most programming tasks, but terrible for user interface testing. Running the iPhone or Android simulator on a laptop will let you test whether your interface is usable on a (relatively) low DPI screen, with a mouse and keyboard and no accelerometers. These settings have absolutely no relevance to determining whether the interface is usable on a device with no keyboard or mouse, a multitouch screen, and several accelerometers.
One final thing to remember: The gestures that Apple uses were first published several years before the first iPhone, as part of a research project going back another decade. They weren't used on any mainstream device, because multitouch screens were very expensive. It's always worthwhile to check papers from older human-computer interaction conferences and journals. They're full of papers that were ignored by the mainstream because they worked well on a high-end workstation, but weren't feasible on cheap commodity hardware. A few years later, a mobile phone has more processing power to spare than the high-end workstation can offer, and those ideas suddenly have become relevantas long as someone still remembers them.