Original practice scenario: A hypothetical memory-product configurator combines capacity, interface, timing and board constraints. How would you detect incompatible combinations and explain failures?
Read model answer +
Represent compatibility as explicit, versioned constraints with understandable reasons. A list of accepted product names does not explain why a new combination should be rejected.
Work through it
Separate component identity from its attributes, validating required fields and units before evaluating dependencies among interface, timing, voltage and board support.
Evaluate constraints in dependency order, distinguishing failed comparisons from checks that cannot run because an input is missing or invalid.
Return stable rule identifiers, observed values and supported alternatives where justified, with fixtures covering boundary values and conflicting constraints.
Tradeoffs to discuss
Declarative rules improve auditability but need a clear precedence model; procedural validation can express complex behavior while becoming harder to explain.
Common pitfalls
Do not assume a newer component revision is compatible with every older board.
A skipped rule must not be counted as a passed compatibility check.
Original practice scenario: A hypothetical memory manufacturing dashboard averages daily pass percentages across test stations. Stations process different quantities and some units are retested. Design a corrected metric.
Read model answer +
Define the tested-unit population and retest policy before aggregating. An unweighted average of percentages gives a low-volume station the same influence as a high-volume station.
Work through it
Retain unit, test attempt, station and outcome identities, selecting the intended first-pass or final-pass measure with an explicit rule for incomplete attempts.
Aggregate passed and eligible counts at compatible product and time grains, deriving the combined percentage from summed counts rather than averaging station percentages.
Reconcile units entering and leaving the test population, showing retest frequency and coverage alongside yield to preserve operational meaning.
Tradeoffs to discuss
First-pass yield exposes rework while final-pass yield describes eventual acceptance; reporting both avoids conflating those purposes.
Common pitfalls
Do not count a unit once for every successful retest in the numerator.
A station with no eligible tests has an undefined rate, not automatically zero or one hundred percent.
