So far, we used the curl command to perform a suite of acceptance tests. That is obviously a considerable simplification. Technically speaking, if we write a REST web service, then we could write all black-box tests as a big script with a number of "curl" calls. This solution would be, however, very difficult to read, understand, and maintain. What's more, the script would be completely incomprehensible by non-technical, business-related users. How to address this issue and create tests with a good structure, readable by users, and meet its fundamental goal: automatically checking if the system is as expected? I will answer this question throughout this section.





















































