Cambridge A-Level Computer Science 9618 – Program Testing and Maintenance
12.3 Program Testing and Maintenance
Why Ongoing Maintenance Is Essential
Software is never truly finished. After a system is released it continues to operate in a changing environment. Ongoing maintenance is required to:
Fix defects that were not discovered during testing.
Adapt to new hardware, operating systems, or external APIs.
Improve performance, usability or security in response to user feedback.
Prevent future problems by refactoring or updating documentation.
Neglecting maintenance can lead to system failures, security breaches, increased cost of later fixes, and loss of user confidence.
Types of Maintenance
The IEEE defines four primary categories of software maintenance. Each type addresses a different need and has distinct activities.
Maintenance Type
Purpose
Typical Activities
When It Is Triggered
Corrective
Fix faults discovered after deployment.
Bug identification and debugging.
Patch creation and distribution.
Regression testing to ensure no new faults.
When errors are reported by users or detected by monitoring tools.
Adaptive
Modify the system to work in a changed environment.
Porting to new operating systems or hardware.
Updating third‑party libraries or APIs.
Changing configuration files for new network settings.
When external dependencies, standards, or platforms evolve.
Perfective
Enhance functionality, performance, or usability.
Adding new features requested by users.
Optimising algorithms for speed or memory.
Improving user interface design.
When stakeholders request improvements or competitive pressures arise.
Preventive
Reduce the risk of future failures.
Refactoring code to improve structure.
Updating documentation and comments.
Implementing additional test cases.
Proactively, often as part of a scheduled maintenance cycle.
Relationship Between Testing and Maintenance
Effective testing supports all four maintenance types:
Regression testing ensures corrective changes do not introduce new faults.
Compatibility testing validates adaptive modifications against new environments.
Performance testing measures the impact of perfective enhancements.
Static analysis and code reviews help identify areas for preventive work.
Illustrative Workflow
Below is a simplified workflow that demonstrates how a maintenance request moves through testing and implementation.
Suggested diagram: Flowchart showing “Report → Classification (Corrective/Adaptive/Perfective/Preventive) → Impact Analysis → Development → Testing (unit, integration, regression) → Deployment → Review”.
Key Points to Remember
Maintenance is a continuous activity that extends the useful life of software.
Each maintenance type has a distinct focus but they often overlap in practice.
Testing is integral to every maintenance activity to maintain system reliability.
Preventive maintenance, though less visible, can significantly reduce future corrective costs.
Example Calculation
Suppose a system has 5,000 lines of code (LOC) and a defect density of 0.8 defects per KLOC after release. The expected number of defects to be fixed during corrective maintenance is:
\$\text{Defects} = \frac{5{,}000}{1{,}000} \times 0.8 = 4\$
Understanding such metrics helps plan maintenance resources.