Do You Trust Computer Mathematicians?

from the prove-they're-wrong dept

A fascinating article in the NY Times looks at the controversy over the publication of a new mathematical proof that relies on some complex computer calculations to prove the theory. Since the computer is needed for such complex calculations, it's incredibly difficult for others to verify that the proof is valid. Many have looked over this particular proof and haven't found anything wrong with it, but they haven't been able to check it all conclusively. The journals that publish such things weren't quite sure what to do - not wanting to publish what could be a very good proof by suggesting that it was somehow less valid because no one was able to effectively review it all. In the end, they came up with something of a compromise solution: the mathematics part was published in a mathematics journal, while the computer part was published elsewhere in a computer science journal. Still, as mathematical proofs become more complex, this problem is going to show up more often. Computers can be helpful in solving some particularly tricky mathematical questions, but it's going to depend on how much people trust the computers to do the math properly.