5 ms·
Sadly, I share your pessimism about formal methods. I suspect one of the headwinds is that there's a greater demand for more software than there is for better-p
by DoofusOfDeath 7y ago
Sadly, I share your pessimism about formal methods. I suspect one of the headwinds is that there's a greater demand for more software than there is for better-proven software.
- tuvistavie 7y agoAt the end of the day I think it's just a matter of cost. If it's cheaper to let a software fail and simply fix it when it does, there will not be much interest into proving it to be correct. If a failure will likely be vastly more expensive than proving correctness, then proving correctness will make sense. There are simply way less instances of the second category. I think one of the few "new" instances, which might be worth mentioning is smart contract programming, where failures can cause millions of loss and proving correctness is not too expensive.
- crimsonalucard 7y agoWith the amount of infrastructure focused around testing I think the opposite is true. I think it's more ignorance and culture. Many people don't know about formal methods.
- Ididntdothis 7y agoAgreed.
- zozbot234 7y agoHaving better-proven software is a precondition to having more software. Especially if you expect that things like program synthesis are going to be increasingly important - who wants to "synthesize" buggy code?