3 ms·
I disagree with the point about human validation tests being unsuitable for multiplication. Sure it’s a contrived example but even with more complex financial r
by exdsq 6y ago
I disagree with the point about human validation tests being unsuitable for multiplication. Sure it’s a contrived example but even with more complex financial rules, you’re building the system for a reason and generally know enough example inputs and outputs to have these tests. If I saw some similar code that didn’t have these types of tests because the ‘machine is more capable at solving them’ I’d fail the review. How do they know it works then?
- bluGill 6y agoMore than that, if I write in a test assert(10=2*3), eventually I will find that test fails and fix it. Thus getting the test wrong isn't fatal. Of course if this is a more complex case (which it probably is) the bug might take a while to discover since it made me write wrong production code, but at least once it is discovered and fixed (both the test and production) that will not be broke again.