The first person to solve a multiplication question posed nearly 50 years ago is still terrified someone will prove him wrong.
Two academics from Australia and France announced last week they had found a new way to multiply massively large numbers.
“When it first came up, we were very nervous that we’d made a mistake,” said University of NSW Associate Professor David Harvey.
“So a lot of the work was convincing ourselves that it’s actually correct.
“I’m still terrified that it might turn out to be wrong.”
While long multiplication is okay for most numbers, it’s terrible for computers trying to multiply numbers with billions of digits, Harvey said.
Two German mathematicians called Schonhage and Strassen predicted in 1971 the existence of a particular algorithm that would allow huge numbers to be multiplied in a certain number of steps.
The pair got close to solving it themselves but Harvey’s solution is the first to prove the 1971 prediction.
He and his French collaborator, Joris van der Hoeven, realised they were on to something about 18 months ago and dropped everything to solve it.
“Everyone asks me how big do the numbers have to be and at the moment we don’t know,” the UNSW academic said.
“You’d have to start in the billions or trillions of digits.
“It sounds crazy but I assure you that in a lot of areas of science, people do need to work with numbers this big.”
It may not even be practical yet.
He compared the discovery’s importance to the early research on solar panels long before they were realistic alternative energy producers.
But for now, Harvey is just keen to stop thinking about it.
“I’m over it now,” he says while laughing. “I want to do some other stuff now.”
A paper proving the algorithm is now online and will be subject to peer-review over the coming months.Do you have an idea for a story?
Email [email protected]