> I think that there is conceptually no superior way of scrolling.

...which means you should go with the mechanism people have already internalized. I thought that was a key principle of UI design: if you don't have a good reason to violate established expectations, don't do it.

One argument for the new behavior is that more people have internalized scrolling behavior on a modern touchscreen browser than have internalized the old MacBook way of scrolling. It makes even more sense looking forward. Steve Jobs has predicted that classic keyboard-equipped computers are going the way of "trucks" - if that's the case, it makes some sense to standardize toward touchscreen conventions.

> ...which means you should go with the mechanism people have already internalized. I thought that was a key principle of UI design: if you don't have a good reason to violate established expectations, don't do it.

Sure that's good advice for beginners, but Apple is good at design. They know the rules and the know when to break them. This is how design evolves; because great designers can imagine how something new could ultimately be better once users get used to it.

If you just follow all the rules you end up like Jakob Nielsen, with a data-backed justification for every detail, all of which adds up to something inelegant and ugly as sin.

This reminds me of some usability trials I did last year - relatively few of the users employed the mousewheel to scroll (only one, as I remember it) when viewing web pages. But almost all of them had iPhones so obviously they knew how to scroll there.

I'm a two-finger scroller on my Macbook, I'd be willing to accept the argument that there are actually more people doing it in the iOS way now.