do-tools-actually-matter.md (879B)
1 # Do Tools Actually Matter 2 3 I think about fuzzing / pbt and cool outcomes. that's great. Is this just on the edge? Are the kinds of ppl who use these things the kinds of people who would find these issues even if they had shit tools? I was trying to find a snapshot based testing tool in C++ to validate serialized data models and realized I don't really give a fuck and can just use asserts, one write string to file function, one read file to string. 4 5 ## Does it matter 6 7 I think vet kinda matters because there are optimizations we performed to find issues. So in a sense, that's better than just using a LM and saying find shit. In a similar vein, if we compare something using a better complexity heuristic, like LEOPARD-V, it'll be a better (fuzzer) than if it was just using some sort of naive code coverage metric, assuming consistent runtimes and non-exhaustive search.