Simplified Domain Coverage: A Complete Guide for the ISTQB CTAL-TA Exam
Introduction
Simplified domain coverage belongs to the data-based test techniques in the ISTQB Certified Tester Advanced Level Test Analyst (CTAL-TA) syllabus. It is the coverage criterion used with domain testing. Domain testing extends equivalence partitioning and boundary value analysis so they also work when inputs interact, for example when a condition involves several variables at once, such as x + y ≤ 10.
This guide explains why the technique matters, what it is, how to derive test points step by step, and how to answer exam questions on it.
1. Why Simplified Domain Coverage Is Important
Classic boundary value analysis (BVA) treats each input variable on its own. Many real requirements combine variables in a single condition, for example:
- loan amount ≤ 5 × annual income
- weight + luggage ≤ 120 kg
- age ≥ 18 AND score > 60
In these cases, testing each variable's minimum and maximum separately can miss defects in how the boundary itself is implemented. Simplified domain testing matters for four reasons:
- It targets real boundary defects. Developers often write < instead of ≤ (a closure defect), use a wrong constant (a shifted boundary), or use a wrong coefficient (a tilted boundary). ON and OFF points are designed to expose exactly these faults.
- It is efficient. The full domain-testing strategy needs many points per boundary. The simplified strategy, based on the work of Jeng and Weyuker, needs only about two points per inequality boundary while keeping strong fault-detection ability.
- It is measurable. It defines clear coverage items, so a test analyst can report coverage objectively.
- It handles multiple dimensions. It works for one-dimensional and multi-dimensional input domains, where simple BVA becomes unclear.
2. What It Is: Key Concepts
Domain: a subset of the input space in which the system should behave the same way, like an equivalence partition, possibly over several variables. Each domain is enclosed by boundaries.
Boundary (border): the edge of a domain, defined by a condition such as x ≤ 100 or x + y ≤ 10.
Closed vs. open boundary:
- A closed boundary includes its points in the domain. It uses ≤, ≥ or =.
- An open boundary excludes its points from the domain. It uses < or >.
ON point: a point that lies exactly on the boundary.
- For a closed boundary, the ON point is inside the domain.
- For an open boundary, the ON point is outside the domain.
OFF point: a point as close as possible to the ON point, but on the opposite side of the boundary from the ON point. Its domain membership is the opposite of the ON point's.
- For a closed boundary, the OFF point is outside the domain.
- For an open boundary, the OFF point is inside the domain.
- The distance between ON and OFF depends on the data's precision: 1 for integers, 0.01 for money in cents, and so on.
IN and OUT points: points clearly inside or outside the domain, away from the boundary. They are sometimes used as additional, typical values.
The simplified domain testing strategy:
- For each inequality boundary (<, ≤, >, ≥), select one ON point and one OFF point.
- For an equality boundary (=), select one ON point and two OFF points, one on each side, because values on both sides are outside the domain.
Coverage items: the ON and OFF points identified for all boundaries.
Coverage measure:
Simplified domain coverage = (number of ON and OFF points exercised by tests) / (total number of ON and OFF points identified) × 100%
3. How It Works: Step by Step
1. Identify the domains from the specification, with their defining conditions.
2. List every boundary of each domain as a separate condition.
3. Classify each boundary as open (<, >), closed (≤, ≥) or equality (=).
4. Determine the precision of the data, which sets the step between ON and OFF points.
5. Select the ON point(s) exactly on each boundary.
6. Select the OFF point(s) as close as possible to the ON point, on the other side.
7. Remove duplicates. Adjacent domains often share a boundary, so the same ON/OFF pair can serve both. Count each point once when calculating coverage.
8. Define expected results using domain membership: a point inside the domain gets that domain's behaviour, and a point outside gets the neighbouring domain's behaviour or an error.
9. Measure coverage using the formula above.
Example 1: One-dimensional domain
Requirement: a valid quantity satisfies 0 < x ≤ 100, with integer values.
- Boundary x > 0 is open:
- ON point = 0, which is outside the domain.
- OFF point = 1, which is inside the domain.
- Boundary x ≤ 100 is closed:
- ON point = 100, which is inside the domain.
- OFF point = 101, which is outside the domain.
Result: 4 coverage items {0, 1, 100, 101}. Testing 0, 1 and 100 only gives 3/4 = 75% coverage.
Note: in one dimension, this matches two-value BVA.
Example 2: Two-dimensional domain
Requirement: x ≥ 0, y ≥ 0, x + y ≤ 10, with integer values.
- Boundary x + y ≤ 10 is closed:
- ON point = (4, 6), since 10 is on the line and inside the domain.
- OFF point = (4, 7), since 11 is outside the domain.
- Boundary x ≥ 0 is closed:
- ON point = (0, 5), which is inside the domain.
- OFF point = (-1, 5), which is outside the domain.
- Boundary y ≥ 0 is closed:
- ON point = (5, 0), which is inside the domain.
- OFF point = (5, -1), which is outside the domain.
Result: 6 coverage items. Choose ON points away from corners where possible, so that each point checks one boundary and does not hide a defect in another.
Example 3: Equality boundary
Requirement: a discount applies only when items = 5.
- ON point = 5.
- OFF points = 4 and 6, both outside the domain.
Result: 3 coverage items.
Defects detected
- Closure defect (≤ coded as <): the closed ON point is wrongly rejected.
- Shifted boundary (≤ 100 coded as ≤ 101 or ≤ 99): the OFF or ON point gives the wrong result.
- Tilted boundary (wrong coefficient in x + y): ON or OFF points in the multi-dimensional case reveal it.
- Missing or extra boundary: often revealed when the expected domain change does not happen.
4. Relationship to Other Techniques
- Equivalence partitioning gives the domains.
- BVA is the one-dimensional special case.
- Simplified domain testing generalises BVA to boundaries defined over several variables.
- Compared with the full domain testing strategy (several ON points plus OFF points per boundary), the simplified strategy is cheaper but slightly less thorough for non-linear or tilted boundaries.
Exam Tips: Answering Questions on Simplified Domain Coverage
Tip 1: Classify each boundary first. Before choosing any value, mark each boundary as open, closed or equality. Most wrong answer options come from mixing up whether the ON point is inside or outside the domain.
Tip 2: Memorise the rule.
- Closed boundary: ON inside, OFF outside.
- Open boundary: ON outside, OFF inside.
- In every case, the ON point is on the boundary value itself.
Tip 3: Count coverage items correctly.
- Inequality boundary: 2 items.
- Equality boundary: 3 items.
- Then remove points shared between adjacent domains, if the question concerns the whole input space.
Tip 4: Respect precision. If values are in cents, the OFF point for ≤ 50.00 is 50.01, not 51. Questions often hide the precision in the scenario text.
Tip 5: Calculate coverage carefully. Coverage = exercised ON/OFF points ÷ total ON/OFF points. Check whether each test value in the options actually matches a required point. A value that is merely inside the domain (an IN point) does not count.
Tip 6: For multi-variable conditions, check the boundary equation. Substitute the candidate point into the condition. An ON point must make the expression equal to the boundary constant. An OFF point must be the nearest value that flips the result.
Tip 7: Find the minimal test set. When asked for the minimum number of tests, consider whether one test can cover an ON point of one boundary and an OFF point of another. Keep in mind that the syllabus recommends avoiding corner points that mask defects.
Tip 8: Link points to defect types. Scenario questions may ask which point detects a defect such as ≤ coded as <. The answer is usually the ON point of the closed boundary.
Tip 9: Eliminate distractors. Wrong options typically:
- use IN/OUT points instead of ON/OFF points,
- swap ON/OFF membership for open boundaries,
- omit the second OFF point for equality, or
- use the wrong step size.
Tip 10: Practise under time pressure. K3 (apply) questions on this topic take time. Sketch a quick number line, or an x-y grid for two variables, and label ON and OFF points before reading the options.
Summary
Simplified domain coverage requires one ON and one OFF point per inequality boundary, and one ON and two OFF points per equality boundary. Coverage is measured as the percentage of these points exercised. Master the open/closed rule, data precision and the counting rules, and you can answer CTAL-TA exam questions on this topic confidently.
