Independent Practice and Transfer
Complete the prerequisite check, concept sequence and worked examples first. Keep the solution page closed until you have written an attempt. These cases deepen the existing practice and lab evidence; they do not replace the original topics, project or practical gates.
Case A — Predict and justify
A proposal uses a benchmark from a different workload to justify a database choice. Write a blocking finding or question with a decisive next experiment.
Record the expected result or state transitions, the assumption that makes them valid, and one plausible but incorrect answer with a counterexample. A result without its reasoning is incomplete.
Case B — Change the assumption
New evidence invalidates an assumption after an ADR is accepted. What should change in the record?
Identify exactly which part of Case A survives and which must be recomputed or qualified. Keep both attempts; do not overwrite the first with the calibrated answer.
Experiment — Connect the claim to an actual observation
Use the proposal-review kit and one actual or explicitly labeled solo architecture review. Preserve alternatives, prioritized findings, decision owner and a revisit trigger; generated summaries do not replace reviewed artifacts.
Use the existing guided labs and their setup/contract instructions. Save the exact version or source revision, input, command/procedure, predicted outcome, actual output and interpretation. Add a negative control that would expose the particular mistaken claim, then restore the correct setup and rerun. Use a disposable copy for deliberate failures.
If this task requires a runtime or reviewer you do not have, mark that observation pending. A written prediction, simulator, reference implementation or generated review cannot close an actual-system or human-evidence requirement.
Review, repair and transfer
Compare your attempt with separate practice calibration. Repair the smallest misunderstanding and answer a fresh variant unaided. At the module's existing delayed recheck, change a value, failure boundary or assumption and defend what transfers. Reuse this evidence in the module assessment; its per-gate thresholds and critical errors remain authoritative.
Count this work inside the original practice, lab and recheck allocation in semester workload and progression. Reuse a matching original exercise artifact and extend it with the changed case instead of repeating identical work. Record actual time and overruns for pilot calibration; the published estimate is not a deadline.