Lessons from the fundamental issues
The preceding examples are mostly modeled as end-to-end testing, where all the components are exercised as one system. In Chapter 4, we’ll discuss how to build end-to-end tests, which are easy to debug, structured to easily change, and even so readable to be a documentation aid, explaining how the software should behave. For now, we’ll talk about overcoming the other weaknesses.
It’s easy enough to throw up our hands and say that unassisted evaluation of the software by a tool is a bad idea and walk away. We have peers who have said similar things; some call them the “anti-automation brigade.” By now, you probably realize that at least some contexts exist where this debate is worth having.
This debate can also help us find ways to overcome these fundamental problems:
- Visual testing, for example, is the received term for a style of testing that compares window captures from just portions of a screen...
