Increased code coverage through unit testing can decrease the number of flaws in the final product, but that doesn’t come for free. Development time is spent working on and maintaining those extra tests. Sometimes the product would have been better for the users if that time had been spent on:
- Design
- Features
- Releasing earlier and more often
More time spent writing unit tests does not imply the product will be better for it.
Everything must be balanced against the product. There’s some level of testing that is optimal. After that point, pushing for higher code coverage isn’t ‘professional’, it’s criminal overengineering.
Discussion on Hacker News
-
telechargerz liked this
-
yieldthought posted this
