¿ªÔÆÌåÓýFor years, across multiple employers, I actually had a fair bit of success in evangelizing TDD.?I noticed two keys: 1. There was an objective measure of quality that everyone monitored and cared, and we were far more consistent than other groups in keeping quality high, and 2. I was able to work in person with developers who wondered how we were doing it.? Since we moved to remote, and in a process that makes it much harder to assign blame for failures, I have found it much harder to interest developers in learning. I have also found, in working with others, that the very idea of a test is, to most people, something that is inherently done to check a built product, not something that guides it. That has had me trying to think of another approach to selling it. In a Twitter conversation, I learned that some have suggested changing the terminology. One suggestion has been ¡°Example-guided development,¡± but a friend who has followed my suggestion, and used to teach math in college found that his students generally thought of examples as necessarily very simple. I have another idea that I am looking to write up, and would require some background to develop, but I agree that a basic problem is the way most people learn to program: syntax first and then play. It is not helped by university computer science programs which generally involve working on programs far smaller than those professional developers tend to work on - and where you can actually just keep the program in your own memory the whole time.?
|
