The First Test: Does the Loop Make Learning Legible?

yczz/oc-english now supports a useful checkpoint: a learner can answer a question, leave a durable trace in a unit, reach an exam, receive a score and points, and carry those points into a character or room. This chapter connects those already-established parts into one judgment about legibility. It asks what the application can directly show, what a learner may reasonably infer, and why a visible progression loop still falls short of evidence for durable language ability.

“Legible” here means that the important state changes can be seen and explained. It does not mean that every internal detail is exposed, or that the design is automatically effective. A learner should be able to tell what happened, why the next action became available, and what the resulting reward represents. The repository gives us enough material to test that standard.

Three layers, one decision

The loop is easiest to understand as three layers that must remain connected without being collapsed into one another.

The first is the answer result. On a practice card, the learner selects one of four options. The application can classify that selection as correct or incorrect for the card in front of them. This is the narrowest and fastest signal. It belongs to one interaction and answers one question: did the chosen option match the answer stored for this card?

The second is persisted progression. A correct word or listening card can cause the word to be recorded for the selected book and unit. That record is not the same as a click count. The word is added only when it belongs to the unit and is not already in that unit’s practiced-word list. The profile retains this state, together with exam best scores, pass status, and attempt count. Progress therefore answers a longer question: has this unit’s required word coverage reached the application’s completion condition?

The third is the visible consequence. Correct practice answers add points. A first passing exam adds a larger reward, and an improved best score can add a smaller retake reward. Points can then leave the study screen and become a purchase, an equipped item, or a room arrangement. This layer answers a product question: did the recorded effort change something the learner owns and can revisit?

These layers form a causal chain, but they do not have the same evidentiary strength. An answer result causes a progress update only when the command accepts it. Progress causes exam eligibility only when every unit word is present. A passing exam causes a reward only under the defined threshold and best-score rules. The visible item is therefore downstream of several conditions. It is not a second measurement of the learner’s English.

The state transitions make that relationship clearer:

stateDiagram-v2 [*] --> Unstarted Unstarted --> Practicing: Open practice Practicing --> Practicing: Wrong answer Practicing --> Practicing: Correct repeated word Practicing --> WordCovered: Correct new word WordCovered --> Practicing: More words remain WordCovered --> ExamReady: All unit words recorded ExamReady --> Passed: Score 12 or more ExamReady --> EligibleAgain: Score below 12 EligibleAgain --> Retake: Start another attempt Passed --> Retake: Start another attempt Retake --> Passed: Score 12 or more Retake --> EligibleAgain: Score below 12