Software interfaces undervalue peripheral vision! (a thread)
My physical space is full of subtle cues. Books I read or bought most recently are lying out. Papers are lying in stacks on my desk, roughly arranged by their relationships.
Peripheral vision spontaneously prompts action.
If I need to fix a door, I’ll be reminded each time I see it. Digital task lists live in a dedicated app. I have no natural cause to look at that app regularly, so I need to establish a new habit to explicitly review my task list.
Peripheral vision emphasizes the concrete.
Unread digital books and papers live in some folder or app, invisible until I decide that “it’s reading time.” But that confuses cause and effect.
If I leave books lying on my coffee table, I’ll naturally notice them at receptive moments. I'll read a book if I feel an actual, concrete interest in it. By contrast, the motivation to read a digital book comes from abstract interest in the habit of reading.
Peripheral vision offers context.
If I mark up a physical book then later flip through to see my margin notes, I’ll always see them in the context of the surrounding text. By contrast, digital annotation listings usually display only the text I highlighted, removed from context.
The primary “unit” in such systems is a single highlight or note, but that’s not how I think. Marginalia have fuzzy boundaries, and I often think of a page’s markings as a single unit.
LiquidText is a lovely counterexample: it works hard to display annotations in context.
In digital note systems, the UI centers on the experience of writing one note. The core operations and representations fixate on “the note you have open,” not on larger structures. I often can’t simultaneously see another note I’ve just finished writing—let alone the last four.
Most systems barely support multiple windows, but even if I can open multiple windows, it’s awkward to arrange them into the spatial relationships I might naturally use for physical index cards. Rather than peripheral vision, it’s like I’m wearing horse blinders and mittens.
Backlinks are a weak peripheral vision, and they help, but they’re generally about switching the one note you have open, not an effective means of sense-making across many notes. Contextual backlinks help, but if you navigate, you lose object permanence.
If I read an old digital note, I get the unnerving sense that it’s part of some “whole” that I can’t see at all—no matter how much hypertext is involved. Working with physical notes, I’d shuffle notes around to make sense of the structure. There isn’t a digital equivalent.
I have most of my drafts in Scrivener, because it’s the only tool I’ve found that gives you any tools at all for large-scale structure. Unfortunately the only usecase it’s willing to address is writing paper books (and basically only pulp fiction), so it’s still terrible for me