Coding vs. Scripting: What is a Test Case?

摘要

Automation has its place in the test lifecycle, but when it comes to automation, the definition of a test case becomes a gray area. Is a test case an XML blob or other metadata that is interpreted by a harness; or is it scripted code (such as JavaScript or VBScript); or is it a function or set of functions within compiled code written in a language like C# or Java? A test case can be any of the above. One could define a test case as a set of instructions to execute a task and verify its results. These instructions can be described in a variety of languages. This paper presents our experience in defining and implementing a script language, its applications in our test scenario; as well as our experience in adopting the Microsoft .Net platform with the C# language to create a long lasting and reusable test model. We started with a set of principles and goals in conjunction with a number of technical challenges to solve. We initially created a test scripting language known as the Simple Language for Scenarios (SLS). We used this language for 3-4 years. As the C# language matured, we were able to find ways to implement the benefits of SLS as well as leverage the power of a fully functional development language. We concluded that overall each approach has its unique pros and cons and it makes sense to use the approach most suitable for a given situation. However, overall we found that moving towards C# was the solution that would work in most situations as it provided us with majority of the functionality and added the structure of an Object Oriented Programming Language.