ITPEC FE Morning April 2023 Question 48

Error Embedding Method — estimate latent errors by seeding known bugs.

Seed S errors into code, then test. Testers find bugs equally regardless of origin.

Variables:
- S — number of embedded (seeded) errors
- m — discovered seeded errors
- T — latent (real) errors before testing
- n — total discovered errors (seeded + real)
- n − m — discovered real errors

Core proportion: detection rate is equal for both types:

m / S = (n − m) / T

Cross-multiply: m × T = S × (n − m), so T = S × (n − m) / m

Why not others:
- (b) m/S = T/(n−m) — inverted right side, gives wrong equation
- (c) m/S = n/T — uses total n instead of real discovered n−m
- (d) m/S = T/n — both numerator and denominator wrong on right side

Key rule: "Catch rate for seeded = catch rate for real" → m/S = (n−m)/T

AI-generated — may contain errors