[…] the kind of control scrolling gives to users seems completely meaningless in the context of the task the user is engaged in. She’s reading a book. It’s a mostly linear affair. Her main goal is to go through the text from beginning to end. The additional control isn’t helping with that goal, it’s just creating more work.
This is true in some use cases, probably the majority, but not all. The other day I was reading a technical book in iBooks and the scroll mode allowed me to accomplish two things that the paginated interface didn’t:
I could scroll down to align the heading of the current section to the top of the page. I was reading tutorial style content in a technical book, most of which was broken down into short sections. I would follow the instructions on my computer and then get back to the section to keep going. Aligning the current section to the top of the page helped me get back to it without searching the page.
More important: the scroll interface did not insert artificial breaks in the content. When you’re looking at code or a list of instructions, you really don’t want them broken up by an artificial page, forcing you to go back and forth to verify that everything you’ve typed is correct.
I agree with Mathis that for a lot of reading the scroll introduces additional work. If you’re reading prose, it’s easier to just hit the next page button or tap the edge of the screen to keep going. That said, there are some use cases, such as technical material, that can benefit from the increased level of control of the scroll and it’s lack of page breaks. Ultimately, we don’t need need to pick one mode or the other; having the choice of the two is good.