I would just like to point out that this is a highly subjective argument. The iPhone 5 might be too large for Dustin's hand, but I'm quite comfortable using a 4.3 or even 4.5 inch phone. I naturally hold phones with the bottom right corner in pushing the bottom of my hand like the second image on the page, and my hand quite naturally wraps a phone up to ~4.5". For me, emphasis me, any phone over 4.5 becomes a problem because my index finger will have a hard time hitting a power switch on the top right of the phone, but overall I'm able to reach any point of the screen with my thumb and small muscle movements in my hand. But again, this is all subjective to the size of my hand.
Exactly: an iPhone 4 is to my girlfriend's hand nearly exactly as a Galaxy Note is to mine. The idea that people make these reachability arguments as if hand size is universal down to silly half-inch increments is downright insane.
"Have you ever heard of 'variance'?" The "common case" makes sense when you are dealing with a discrete variable that has a mode (a single highly common value): it does not make as much sense for a continuous one like hand span.
Put simply, the issue is that no one actually has the "common case": the probability that someone's hand span exactly matches the "average" or "mean" of the distribution is approximately 0.
Instead, we have to first discretize the variable into a histogram and then assign a common case afterwards. The way you normally might do that is with standard deviations. However, now you have to ask "what is the variance".
With hand size, it is quite obviously very large: again, my girlfriend's hand is to an iPhone 4 as my hand is to a Galaxy Note, and people generally do not go around calling either of us freaks of nature with regards to our hands (I actually have fairly effeminate hands ;P).
I will therefore repeat with emphasis: the idea that people make these reachability arguments as if hand size is universal down to silly half-inch increments is downright insane. However, I will now show some data.
This chart comes from a piano company measuring hand span for purposes of optimal piano sizing (concentrating on people whose hands have already finished growing). (Their conclusion, btw, was that they needed to sell a few more piano sizes.)
Now, this is full hand span; however, as your hand is fixed in the middle by the thing you are holding, you won't get this full spread, but I would easily argue that you will get at least half of this spread in that case.
The result, as you can see, is that there is a very very large range of reasonable values: hands are really not that common; you might think you have a common hand, but you really don't: hand span varies by numerous inches.
(There are multiple factors for this, of course: it isn't just hand size; you also have to take into account flexibility of joints, muscle strength, etc.; even if everyone's hands were the same size, the ability to reach the upper-corner of a device would vary a lot.)
I will then argue that even if you want to just attack a lot of people, and still fail for numerous others, you need to at least be making a device for women and a device for men: the difference between genders is really striking in that chart.
However, most people don't argue things like this: they accept that there are massive variations in the size of peoples' feet, hands, torsos, and pretty much every other body part, and this is why we don't just sell one-size-fits-all clothing.
Apple, FWIW, would not only be selling a single size of jeans, they would claim that they had optimized the size of the jeans to fit some optimal size, and their user community would then defend them with diagrams that show variances of less than half an inch for how perfect the fit is.
