3 ms·
I'm not even sure testing always scales better -- it depends on your formal tool -- but it's definitely different. For example, some properties -- especially of
by pron 9y ago
I'm not even sure testing always scales better -- it depends on your formal tool -- but it's definitely different. For example, some properties -- especially of concurrent/distributed systems -- are nearly impossible to test.
Also, the level of confidence offered by formal methods is often configurable to lower costs. Some formal methods can affordably scale to very large systems, with the tradeoff that the confidence is lowered, as it only covers the general algorithm design rather than actual code.