Introduction to Software Testing

Introduction to Software Testing

• Software testing is a process intended to demonstrate that a program functions as intended and to uncover potential defects prior to deployment.
• Testing involves executing a program using artificial data, analyzing the results for errors, anomalies and information about non-functional attributes.
• Testing reveals the presence of errors but not their absence. It is part of a broader verification and validation process, including static validation techniques.

Topics Covered

• What is Software Testing
• Verification vs Validation
• Fault
• Software Inspection and Testing
• Test Cases
• Test Plan
• Stages of Testing
• General Testing Guidelines
• Requirements-Based Testing

Program Testing Goals

• Demonstrate to developers and clients that the software meets its requirements.
• For custom software, one test should be created for every requirement in the documented requirements. For generic software, tests should cover all features and their combinations.
• Identify situations where the software behaves incorrectly, is undesirable, or does not conform to its specification.
• Defect testing aims to identify undesirable behaviors such as system crashes, incorrect interactions with other systems, computation errors, or data corruption.

Validation and Defect Testing

• Validation testing aims to assess if the system performs as expected based on pre-defined tests that reflect expected use cases.
• Defect testing focuses on uncovering defects; designed test cases are not required to represent normal system use, and can be deliberately obscure.

Testing Process Goals

• Validation testing aims to show the client and developer the software meets requirements and that the system works as intended.
• Defect testing aims to discover faults or defects in the software where the program is behaving incorrectly or inconsistently with its specification; successful test cases result in a system performing incorrectly and exposing a defect.

Model of Program Testing

• A model of Program Testing involves input test data and output test results to determine if outputs match predictions or expected values. The input can cause anomalous behaviour.

Verification vs Validation

• Verification is ensuring the product is built as intended (product "right"). It assures compliance with the specifications.
• Validation confirms the product is the right product (building the "right" product). It addresses user needs and requirements.

V & V Confidence

• The confidence level in a system depends on the system's criticality for the organization, user expectations, and marketing environment factors. Early market releases often take preference over finding more defects.

What is a Fault?

• A fault (bug) is an error in software that causes unexpected or incorrect behavior.
• Types of faults include syntax, logic, and runtime errors.

Inspections and Testing

• Software inspections analyze static system representations to identify issues.
• Testing executes the system with test data to observe and record behavior.
• Inspections and testing are complementary verification techniques.

Software Inspections

• Software inspections involve people scrutinizing the source code for anomalies and defects to identify potential faults.
• Inspections don't require system execution, and can be conducted before implementation.

Advantages of Inspections

• They don't mask faults arising between interactions.
• Incomplete systems can be inspected, reducing extra costs.
• Help to ensure standards compliance, portability, and maintainability.

Types of Testing

• Development: Unit, Integration, System tests
• Release: Focused on validating features and meeting functionality and performance requirements
• User: Alpha (developer site), Beta (user environment), Acceptance (ready for deployment)

General Testing Guidelines

• Utilize inputs to generate error messages.
• Design inputs that cause input buffer overflow.
• Repeatedly use the same or similar input.
• Force invalid outputs.
• Use inputs causing large/small results in computations.

Testing Policies

• Develop policies defining the required system test coverage as exhaustive testing is impossible.

Requirements-Based Testing

• Examines each requirement to develop a test; verifies the software behavior matches what is expected.
• Ensure systems do what they're designed for.
• Help identify missing/incomplete requirements.

Key Points

• Testing shows errors, but not their absence.
• Development teams are responsible for development testing; separate teams handle release testing.
• During testing, try to "break" software based on experience/guidelines to find defects.
• Automated tests should be used when possible.