Objective: Choose a small set of inputs that carries most of the information: one representative per equivalence partition, the six boundary values of any range, and a decision table for combinations of conditions.
⏱ ≈75 min full lesson · Practice spec: bedrooms field accepts whole numbers 1–6 · English edition — Español · العربية · বাংলা in translation
🧽 The practice business: FreshNest Home Cleaning — a fictional local cleaning company (owner: Maya Torres) invented for this course. Every example and exercise uses FreshNest, so you learn on a complete, realistic business without ever touching a real one.
Equivalence partitioning divides inputs into groups the program should handle identically. For the bedrooms field, 3 and 4 are the same kind of thing — valid counts inside the range — while 0, 9, 2.5 and 'two' each fail for a different reason. One representative per partition usually exposes the group's behaviour; testing every value from 1 to 6 teaches little extra.
Boundary value analysis observes that programmers make mistakes at edges: a > typed where >= was meant, a range checked at one end only. So you test six values — the minimum, one below it, one above it, the maximum, one below it, one above it. They bracket the range and catch the classic off-by-one defects that mid-range values never touch. The common misconception: 'boundary testing means testing the biggest value.' The maximum is one of six; most boundary defects live at the lower edge and just outside the range.
Decision tables handle combinations. Single-field techniques can't say what should happen when a booking is a Move-Out AND the phone is invalid AND a field is blank. A table lists every combination and its expected action — and exposes combinations nobody specified, which become questions for the owner before they become defects. The trade-off is combinations versus cost: three conditions give 8 rules, eight conditions give 256. Prune with risk — full tables only where wrong actions hurt: money, contact details, legal claims.