The license contains the thought of 40-year-old data science

One 1985 Papercomputer scientist Andrew yaoWho will win the AM Turing Award, the best way to find a certain feature set, the best way to find an individual element or an empty space, is just a vision known as a single approach. At the same time, at the same time, the worst case you are looking for the last open point, the worst case, the worst case, never could be better x. Most of the computer scientists have assumed for 40 years.

Krapivin did not return a common wisdom because he was unaware of it. “I did it without knowing it,” he said. Investigations with small indicators caused a new species of hash table that did not rely on unified trials. And this new hash desk is the time required for the worst job requests and places (input) x(^2nd-Far faster x. This result contradicts directly to shame. Farach-Colton and Kuszmaul krapiv helped (input) x(^2nd Hash tables are an optimal, incredible limit for a popular class of Yao.

“This result solves and solves such a classic problem,” he said Guy Blelloch Carnegie Mellon.

“It’s not just their violation (yao), they found the best answer to his question,” he said Sepehr Assad Of the University of Waterloo. “We could have passed 40 years other than we know the right answer.”

Krapivin on the Royal College Bridge at the University of Kambridge. Its new hash schedule can be faster and more faster and compared to researchers who think so far possible.

PhotoRaph: Phillip Ammon for Quanta Magazine

Along with refuting Jaoh, the new paper also includes a more amazing result. Although a little different, it is a little different, the situation: in 1985, Yao was not only the worst inquiries, but also in the middle of all possible surveys. It was stated that those who have hadh tables containing certain features, including “greedy,”, new elements should be placed in the first existing spot x.

Farach-Colton, Krapivin and Kuszmaul, wanted to see that the same limit was applied to non-stingy hash tables. They met that a counterparty sample, an unparalleled hash table with average survey time, better than the log x. In fact, it is not dependent x usually. “You are taking a number,” said Farac-Colton said: “It’s not just a stable and no matter how much full of hash table.” You can achieve a permanent average survey time, regardless of the hash table, it was completely unexpected – the authors were unexpected for themselves.

The team’s results may not lead to an emergency application, but it is not important for everything, said Conway. “It is important to understand such information structures. You don’t know when it will be, it will open something that allows you to work better in practice.”

Original story reprinted with permission How many magazinesan independent editor of the editor Simons Foundation Whose mission developments and mathematics and physical and physical and physical and life sciences include the development of science and increase the understanding of science.