This chapter presents an overview of current medical software applications and the factors that promise to drive growth in this area. We begin this chapter by defining and discussing some important terms such as digital biomarkers and digital health (Section 7.1).

This chapter provides background on the constraints imposed on software by the need to operate within a healthcare environment. We first present an overview of the environment and the constraints under which our software must operate (Section 3.1).

This chapter describes the process of software validation. We first begin with a brief overview of what validation is. We then review the regulatory guidance for validation (Section 15.1) in general, and then provide an extended discussion on the issues that affect the validation of software modules that use artificial intelligence/machine learning (AI/ML) techniques in particular.

Quality is an important issue for the software industry, though there is a balance between cost and quality. Starting with a few examples of software failures, this chapter discusses the need for testing, and uses an example to demonstrate why heuristics are needed to test software. A seven-step approach to testing software is introduced: analysis, identifying test coverage items and test cases, verifying the test design, implementing and executing the tests, and interpreting the test results. Key concepts are defined as used throughout the book, with a reference to the key IEEE/ISO software testing standard.

Random testing presents three main chanllenges: the test oracle problem, the test data selection problem, and the problem of when to finish testing. These are discussed in detail, and unit-test and application-test examples are worked using a simple but effective solution to these.Barriers to full automation are presented along with an overview of more advanced types of random testing. Some of the limitations are examined through the introduction of faults.

Testing with equivalence partitions introduces the reader to the first and simplest form of black-box and unit testing. First a worked example is used to demonstrate how to progress from a specification of the software to a fully automated test. The steps of the process are then examined in more detail, and the strengths and weaknesses examined through the introduction of faults into the software. The chapter ends, as do all the chapters which introduce new testing techniques, with notes for the experienced tester.

Testing object-oriented software is a significant topic in its own right, and this chapter presents the user with the essential underlying test techniques: testing in class context, and inheritance testing. As in the previous chapters, a worked example is used to introduce the reader to the concepts, which are then subsequently discussed in more detail. The chapter then summarises some more advanced techniques; state-based testing, UML-based testing, and built-in testing. Some of the limitations are examined through the introduction of faults into the working code.

Some forms of testing are significantly more time consuming to develop, and all-paths coverage is one of these. Is introduced to the reader as it is one of the most powerful forms of white-box testing, ensuring that every path from the beginning to the end of the code is exercised during testing. It is unlikely that a tester will use this technique in practice, but an understanding of this technique provides a baseline to compare other white-box tests against.

Building on the knowledge gained in chapter 1, this chapter explains the next form of black-box testing, based on the boundary values of each equivalence partition.

This chapter provides a more detailed description of the background of black-box and white-box testing. it then discusses a number of more advanced issues, covering testing with: more complex data types and data structures, more complex specifications, and more complex code. The chapter ends with an overview of some more advanced forms of white-box testing: condition coverage, decision coverage, decision-condition coverage, multiple condition coverage, and modified condition/decision coverage.

Test automation is used to make software testing must be reliable, fast, and repeatable. This chapter usesan exemplar test framework (TestNG) to demonstrate typical automation features. The handling of timeouts and exceptiomns is examined. A mode advanced look at inheritance testing is presented. The chapter ends with a further look at examining different types of application: web-based, desktop, and mobile.