Decision coverage requires that every decision point in the code is tested with both a true and a false outcome. The provided code has three decision points:

1. width > length

2. height > width (nested in the true path of decision 1)

3. height > length (nested in the false path of decision 1)

To cover the nested decisions, we must execute their parent branches. Two tests are needed for the first branch (width > length is true) to cover the true and false outcomes of height > width. Similarly, two more tests are needed for the second branch (width > length is false) to cover the true and false outcomes of height > length. This results in a minimum of four tests.

A. 3: Incorrect. Three tests are insufficient as they cannot cover all four unique paths required to exercise every decision outcome; one outcome of a nested decision would be missed.

C. 2: Incorrect. Two tests can only cover one path in each of the main branches (e.g., true-true and false-false), leaving two decision outcomes untested.

D. 1: Incorrect. A single test can only execute one path through the code, which is far from achieving 100% decision coverage for multiple decisions.

1. ISTQB® Certified Tester Foundation Level Syllabus, Version 4.0 (2023). Section 4.2.2, "Decision Coverage," states: "Decision coverage exercises the decision outcomes. To do this, the test cases are designed to execute both the true and false outcomes of each decision." This principle confirms that all three decisions in the code must have both their true and false outcomes exercised.

2. Ammann, P., & Offutt, J. (2016). Introduction to Software Testing (2nd ed.). Cambridge University Press. Chapter 2, "Graph Coverage," explains that decision coverage is equivalent to edge coverage on a control flow graph. The control flow graph for the given code fragment has two primary branches, each containing a further two sub-branches. To cover all edges, each of these four distinct logical paths must be traversed, requiring four tests. (DOI: https://doi.org/10.1017/CBO9781316091243)

3. Myers, G. J., Sandler, C., & Badgett, T. (2011). The Art of Software Testing (3rd ed.). Wiley. Chapter 4, "Test Case Design," discusses decision/branch coverage. The logic presented shows that for nested conditional structures, test cases must be created to independently exercise the outcomes of each decision. The mutually exclusive nature of the if-else structure in the example necessitates separate sets of tests for the nested conditions, leading to a total of four tests.