第9章ソフトウェア工学

By the end of this section, you will be able to:

• Explain the importance of testing in software engineering

• Describe the types of tools used by software engineers

• Describe ways that software reuse is made possible

• Explain the role of ethics in software engineering

• Discuss the future of software engineering

The role of a software engineer is wide-ranging, and the factors that impact software engineers are many. There are a few areas, however, that are worth delving into deeper. These areas include software testing, refactoring, design patterns, software tools used in software development, software reuse, free and open-source software (FOSS), software engineering ethics, and legal aspects.

The purpose of testing is to verify that the software being developed delivers on the requirements that were set for the project without any unintended errors or side effects. In simple terms, it is to make sure software does what it was expected to do. Through this process of confirming that the software meets expectations, any issues that are found can be resolved to ensure the integrity of the product

Testing should happen within nearly every phase of a project, with each phase having a specific testing focus. One axiom of testing is to test early and test often. As can be seen in Figure 9.17, the earlier issues can be found when developing software, the less costly they are to fix or resolve. Issues are generally found as a result of testing.

Software goes through various levels of testing to ensure quality:

• Unit testing: Individual software components (functions or methods) are tested in isolation to verify they work as intended.

• Integration testing: Focuses on interfaces between components, ensuring they work together seamlessly.

• System testing: Tests the entire software system as a whole, verifying it meets overall requirements.

In addition to the various levels of testing, there are other crucial testing approaches that are used to ensure that software delivers to expectations. These approaches include:

• Acceptance testing: verifies that the software meets the expectations and requirements defined by the client or end users

• Usability testing: evaluates how easy and intuitive the software is for users to interact with

• Stress testing: assesses how the software performs under heavy loads or unfavorable conditions

• Performance testing: evaluates the speed, responsiveness, and scalability of the software under various conditions

• Security testing: identifies and addresses security vulnerabilities to protect against unauthorized access or data breaches

Some testing is done by software developers, including software engineers who write code, but a majority of testing is done by quality engineers whose primary focus is testing. End users can also participate in acceptance testing or user testing scenarios.

Microsoft uses millions of testers before deploying a new update of Windows. These testers are a part of the Windows Insider program and have access to the prerelease versions of Windows. Anyone can sign up for this program; they just need to agree to test the new version and provide Microsoft with feedback. Because testing is done on a prerelease version of the software, the testers know that the software may be unstable and will likely have issues. If you are curious about testing—or are interested in becoming a tester yourself—visit Microsoft Windows Insider (https://openstax.org/r/76WdwsInsider) or Apple’s Beta Software Program (https://openstax.org/r/76BetaSoftware) for more information.

There are three primary testing methodologies, and they are categorized by the knowledge the tester has (or needs to have) to conduct to the tests. The three methodologies are:

• The manner of testing where the tester is aware of the code, so they can test that the internal structure of the item being tested works properly, is called white box testing. Other names for white box testing include glass box, clear box, and structural testing.

• Tests based on requirements and the functionality of what is being tested without the need to focus on the code itself is called black box testing. The tester does not need to have any knowledge of the code. Other names for black box testing include input-output testing, specification-based testing, and behavioral testing.

• The hybrid of both white box and black box testing is called gray box testing. The person who designs the test has a partial knowledge of code structure and understands the intended design of the software.

When testing at the testing system level, we distinguish between verification and validation. Testing that the software solution functions without errors is called verification. In organizations that have large development teams, verification is often done by a Quality Assurance (QA) team. Testing that the software solution conforms to the requirements and does what the user wants it to do is called validation. Both verification and validation require code execution and can happen in all levels of testing, with verification generally taking place before validation.

Testing is also distinguished by purpose into functional and nonfunctional testing. With functional testing, the functionality is tested. With nonfunctional testing, qualities such as performance, scalability, and usability are tested.

The attributes of a good test are as follows:

• A good test has a high probability of finding an error or issue.

• A good test is not redundant. (It does not test the same thing as another test.)

• A good test should be neither too simple nor too complex.