# Public learner checks · fresh fictional cases

These short checks are new relative to the lesson cards and swaps. The teacher-targeted key is public, so a teacher should collect first responses before opening it. These are **not secure exams** or QCAA assessment instruments. Write, speak with captured notes, or build a labelled diagram with dictated reasoning; each route must show the same evidence-based analysis.

## Check A · Day 5 fresh transfer

**Fictional case: free programme pickup at a lantern festival.** Visitors ask at the entrance whether printed programmes are available. The current paper sign is changed by a volunteer at 10 am and 2 pm. Someone proposes a web page saying “live programme stock”, but no one counts programmes between sign changes. The team can publish a dated message twice daily. No visitor names are needed.

1. State the human need in one sentence and identify the evidence for it. Keep a proposed feature out of the need statement.
2. Choose a defensible first web-page scope and name two limits or missing facts. Explain why a live-stock claim would be unsafe from the given information.
3. Write one **testable** criterion for a paper or local prototype trial. Name the reader, task and observable success condition; do not invent a trial result. Name one possible social impact to watch.

**Routes:** write a three-part brief; explain it aloud while the teacher captures the evidence, two limits and criterion; arrange “need / first version / unknown / test / impact” strips and dictate the same reasoning.

## Check B · Day 10 fresh transfer

**Fictional case: community sound archive descriptions.** An offline kiosk runs a local web application with 24 approved, text-only descriptions of sound clips. Each description already carries one of four topic tags, and each tag has at least one approved description. Visitors want to find a description by topic. The team can edit the tags weekly but has no permission to upload audio, collect search histories or store visitor names. A designer suggests a four-topic filter; another suggests a machine-learning recommendation tool. No test results, training data, maintenance budget or evidence about visitors' search words are supplied.

1. Give the human need and decompose a first-version flow into input, process, output and one human maintenance step. State a scope boundary that follows from the permissions.
2. Compare the filter and recommendation proposals. Recommend a provisional first option with two source-based reasons, and name one fact to gather before revisiting the decision.
3. Give a testable criterion with reader, task and success condition. Explain one possible personal or social impact of the option and one limitation of the classroom evidence.

**Routes:** write a brief and four-stage flow; speak the flow and recommendation with notes captured; arrange four flow cards and two option cards, then dictate the same evidence, criterion and limit. This is a public formative task, not evidence of an implemented kiosk or trained model.