Continuous-scroll digital reading uncomfortably disrupts object permanence

It’s painful to read endlessly-long articles on the web, which scroll continuously. I always set my e-book readers to flip, not scroll continuously. One key cause seems to be that continuous scrolling interactions disrupt object permanence.

When I read a physical book, if I find myself lost, I can flip backwards a few pages, keeping my finger in my current reading spot. LiquidText’s solution probably suffices when flipping forward to understand the book’s structure, but it doesn’t help as much when flipping backwards because I’m also using the discrete physical layout of the page in this operation: for instance, I might vaguely recall that the sentence I’m looking for is on a right-facing page, towards the bottom, etc.

If web publishing is to be ubiquitous, we’ll need some alternative to continuously scrolling through an enormous, undifferentiated block of text.

Related: Applications don’t reliably maintain scroll positions

Vandendorpe (2009) makes a similar argument (p123-124):

As we have seen, the papyrus scroll provides no markers to facilitate the reader’s task. Designed to be unrolled from left to right, it can be read only by going through its columns of text from beginning to end. Readers are thus supposed to follow the thread of the text just as they would follow the thread of a speech or a conversation. Unlike the modern book, which since the emergence of the codex has seen the gradual creation of a large number of markers based on the tabularity of the text, the papyrus provides few clues that allow readers to manage their reading activity effectively. Contemporary readers suddenly confronted with a familiar book in the form of papyrus scrolls would not recognize it. They would have trouble finding a specific section of the text, since there would be no table of contents; locating a particular chapter, since there would be no running heads; finding a particular quotation, since there would be no page numbers; locating references to a specific author, since there would be no index; or determining the sources used in a scholarly work, since there would be no bibliography. And it would not be easy for readers to find where they had left off reading, since there would be no paragraphs or page numbers.


Vandendorpe, C. (2009). From Papyrus to Hypertext: Toward the Universal Digital Library. University of Illinois Press.