When asked to provide testing instructions for a new feature or bug fix, there are three types of developers: some are surgically precise, some are fairly precise while erring towards caution, and some give the widest set of instructions they can get away with, usually in an effort to protect themselves against accusations of insufficient test coverage.

Widely imprecise testing instructions mean two things: you’ll spend more time than necessary on testing, and you’ll find unrelated bugs.

The first problem is quite clear. If you estimated that a certain fix will take three hours to test, and the testing instructions lean more towards three days, that’s a serious change in your schedule. It also makes you look quite silly.

The second problem is more personal. Any bug you find is a good thing, as far as quality is concerned. But some developers will at this point yell at you for getting them blamed for unrelated bugs. It’s hardly your fault – you were following their instructions – but you might not want to have that fight. And in my experience, it’s exactly the developers that try to protect themselves with wild testing instructions that also can’t handle random bugs with grace.

The lesson here: insist on precision. Testing instructions should not amount to “test everything” unless the change really was that large.