Software Testing Fundamentals

In software testing terminology, a failure is when the system deviates from expected behavior during execution - it's the visible symptom to the end-user. A fault is the defect in code that causes the failure. An error is a human mistake that leads to a fault. A defect is often used interchangeably with fault.

Regression testing should be performed after software changes (W) and when the environment changes (Y), as these are the situations where existing functionality might break. Testing every week (V), as often as possible (X), or when the project manager says (Z) are not standard practices - regression testing is event-driven, not schedule-driven.

The IEEE 829 standard outlines the structure of a Test Plan, which includes sections for test items, deliverables, and environmental needs. While a "test specification" is a separate IEEE 829 document type altogether, it is not a structural component or section within the Test Plan document template itself.

Testing should be stopped based on risk assessment for the system being tested, not simply because all tests are done, time is up, or faults are fixed. Risk-based testing considers the cost of remaining defects versus the cost of additional testing. Even with all planned tests run, high-risk areas may need more testing.

The valid range is 10000 to 99999 inclusive. Valid boundaries are the exact edges of the valid range (10000 and 99999), and a valid equivalence class member lies inside the range (like 50000). Other options either include invalid boundaries (like 9999 or 100000) or omit the equivalence class member.

Early test design (i) prevents fault multiplication by catching issues before they propagate, (iii) finds faults in requirements and design, and (iv) can cause requirement changes when testability issues emerge. Statement (ii) is false because early faults are CHEAPER to fix, not more expensive. Statement (v) is false as early design reduces overall effort.

Non-functional testing evaluates quality attributes of a system rather than specific functional behaviors. This includes properties like performance, usability, reliability, security, and scalability. The distractors incorrectly define non-functional testing as testing for functional bugs or isolating specific software features for functional execution.

Configuration management includes status accounting of configuration items (A), identification of test versions (C), and recording changes to documentation over time (D). ISO9001 conformance auditing (B) is part of quality assurance or audit processes, not configuration management itself. Configuration management focuses on version control and change tracking.

Integration testing in the small focuses on combining software units and testing their interfaces. The main purpose of the integration strategy is to determine the order and combinations in which modules should be integrated - whether incrementally (bottom-up, top-down, sandwich) or all at once (big bang). It specifies which modules to combine when and how many at once. Testing individual modules adequately is unit testing, and testing external system interfaces is integration testing in the large.

Test completion criteria, also known as exit criteria, define the specific requirements or metrics (like test coverage or defect density thresholds) that must be met to decide when it is safe and appropriate to stop testing. The distractors describe test execution, test case design, or input generation instead.

Incidents can be closed without being fixed if they are duplicates or invalid (i), and incidents must be raised during independent testing (v). However, incidents can be raised against documentation (ii), the final stage is closure (iii), and incident records must contain environment details (iv), making the other options incorrect.

Statement coverage requires every statement to be executed at least once. With test values P=60, Q=50, both IF conditions are true and all 4 statements execute, achieving statement coverage in 1 test. Branch coverage requires both true and false outcomes for each decision. There are 2 IF statements (branches): P+Q>100 and P>50. The second branch (P>50) is only reached when the first is true. With P=60, Q=50 both are true; with P=30, Q=40 both are false. Thus 2 tests achieve full branch coverage.

One test case where the IF condition evaluates to True is sufficient to execute all statements. However, for full branch coverage, a second test case where the condition evaluates to False is required to cover the implicit else branch, making one test for statement coverage and two for branch coverage correct.

Two test cases are sufficient for both statement and branch coverage. The first test (e.g., A=10, B=5, D=5) covers the true branch of the first IF and the true branch of the second IF. The second test (e.g., A=5, B=10, D=10) covers the false branches, achieving full coverage.

To achieve full statement coverage, we need two test cases: one where you watch television (covering the IF block) and one where you continue reading and complete the crossword (covering the ELSE block and second IF). Achieving full decision coverage requires three test cases to evaluate all True/False outcomes of both decisions.