What does condition coverage require in terms of program decision testing?

Condition coverage specifically requires that each individual condition in a decision within the program is tested for both of its possible outcomes: true and false. This means that for every logical condition (such as an 'if' statement), the test cases must be designed to ensure that both outcomes are exercised at least once during the testing process. By doing so, it guarantees that the conditions behave as expected under various scenarios, leading to a more thorough validation of the software's logic.

This type of coverage goes beyond simply testing paths that the code could take; it's more granular because it directly assesses the logic of individual conditions without necessarily focusing on the entire path through the code that might result from evaluating those conditions. The emphasis is thus on verifying the correctness of each decision point rather than traversing every possible execution path or ensuring optimal performance under load, which are not the primary focus of condition coverage.