JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD MCA 2023 SOFTWARE TESTING METHODOLOGIES Previous Question Paper (March/April - 2023 - R20)

Testing is a critical activity in the software development lifecycle, but it is inherently constrained by several fundamental limitations that prevent it from guaranteeing a completely error-free product:

• Exhaustive testing is impossible: For any non-trivial software system, the number of possible input combinations, execution paths, and environmental states is astronomically large. Testing every possible scenario is computationally and economically infeasible.
• Testing shows the presence of defects, not their absence: Testing can identify the existence of bugs, but it cannot prove that a system is entirely free of defects. Even after extensive testing, latent errors may remain in untested code paths or edge cases.
• Dependency on test data and environment: The effectiveness of testing is limited by the quality and representativeness of the test data. If the test environment does not accurately mirror the production environment, critical bugs related to hardware, network latency, or integration may go undetected.
• Human error and bias: Test design and execution are performed by humans. Testers may overlook specific scenarios, misinterpret requirements, or suffer from confirmation bias, leading to gaps in test coverage.
• Cost and time constraints: Testing is often performed under strict project deadlines and budget limitations. This forces teams to prioritize certain features over others, meaning that lower-priority modules may receive less rigorous testing, increasing the risk of residual defects.

In domain testing, domains are categorized based on their geometric and algebraic properties. A nice domain is characterized by properties that make testing straightforward and predictable, while an ugly domain possesses features that complicate test case design and increase the likelihood of errors.

Properties of Nice Domains:

• Linear Boundaries: The boundaries are defined by linear equations, making it easy to identify test points.
• Complete and Consistent: The domain is defined by a complete set of inequalities with no missing or overlapping regions.
• Small Number of Boundaries: The number of boundaries is small relative to the number of input variables, simplifying the selection of on-points and off-points.
• Orthogonal Boundaries: Boundaries are often orthogonal (perpendicular) to the coordinate axes, which simplifies the calculation of boundary values.
• Convexity: The domain is convex, meaning any line segment connecting two points within the domain lies entirely within the domain.

Properties of Ugly Domains:

• Non-linear Boundaries: Boundaries involve complex curves or higher-order polynomials, making it difficult to determine exact boundary points.
• Discontinuity: The domain may be fragmented or have holes, leading to gaps in the input space that are difficult to test.
• High Dimensionality: A large number of variables creates a complex boundary structure, increasing the number of required test cases exponentially.
• Non-convexity: The domain has concave regions or irregular shapes, which can hide errors within the "dents" of the boundary.
• Boundary Overlap or Ambiguity: Boundaries may be defined by complex logical conditions that lead to ambiguous regions where it is unclear whether a point belongs to the domain or not.