The mnemonic medium’s design relies on light-weight prompts

The Mnemonic medium is designed to make answering prompts extremely lightweight.

  • Tiny tasks: Good recall prompts in a Spaced repetition memory system test just one thing, so they can be answered quite quickly. Quantum Country’s median prompt review time is 6 seconds (as of 2019/12).
  • One-tap marking: Because the prompts are self-graded, the interaction itself is quite lightweight: just one tap. No need to precisely target: the buttons each cover half the screen. See also Self-graded spaced repetition memory systems are more efficient than machine-graded systems.
  • Interaction-less reading: The questions and answers are short enough that they can be read without scrolling, which ensures that the only interaction necessary is for revealing and marking answers. This also means that there’s no state involved in the experience of reading a question.
  • Stability: The layout of the prompts and the review area are fixed, and neighboring prompts are stacked in the z dimension rather than vertically. So as each prompt is answered, the next appears in the same location on the screen. The user doesn’t have to “find” the starting point for the next prompt’s question or answer, because they’re always in the same absolute position.
    • Likewise, readers don’t need to “retarget” their mouse or finger from prompt to prompt because the buttons are always in the same absolute position.
    • Prompts don’t change their layout on different devices.
  • Tolerance: Most readers don’t care about every prompt. If prompts were weighty like a textbook exercise, many readers would pick and choose, making a conscious decision about whether answering each prompt. That extra decision makes each prompt even heavier! But if prompts are quite cheap to answer, a few uninteresting prompts aren’t so burdensome. Some readers might wish to “opt out” of certain prompts, but it’s possible to assume that readers will answer each prompt as a strong default.

Lightweight prompts are important to comprehensive coverage

If readers are to remember all the key material from a mnemonic text, authors must include many prompts (e.g. QCVC has 112 prompts and 20k words). As prompts become lighter-weight, readers will tolerate more of them at a given level of interest or time commitment. If each prompt takes 6s, we can add 112 prompts to QCVC at the cost of about 10 minutes—not so bad, since it takes most readers 3-4 hours to finish that essay.

If we only consider in-essay reading time, it wouldn’t be so bad if prompts took twice as long to answer. 20 minutes instead of 10 minutes in the context of a 3-4 hour reading experience doesn’t seem terribly material. But the efficiencies (or inefficiencies) compound. Most readers require 5-7 repetitions to reach multi-month retention, so at 6 seconds per prompt, we can expect that retaining those 112 prompts “costs” about 90 minutes in the ensuing months. Doubling that number would be a significant burden. (See Mnemonic essays may offer detailed retention of their contents in exchange for 35-50% reading time overhead)

Lightweight prompts can be interleaved more regularly

Textbooks traditionally punctuate sections with exercises, but these problems are often fairly involved. They might require extended thought or pen and paper—in any case, a significant context switch away from the reading experience. You wouldn’t want to consider that context switch every few paragraphs.

But because the mnemonic medium’s questions are lightweight, we can interleave them every few hundred words in Quantum Country. We were concerned about the disruption this might cause to the reading experience, so we interviewed readers extensively about prompt timing. No readers reported finding the review sets disruptive or annoying. Instead, we’ve found a number of positive effects; see Mnemonic medium prompts are interleaved into the reading experience for more discussion.