Software engineers and timekeepers rejoice as the world votes to end the leap second
The rule begins in 2035 and remains in effect for 100 yearsBy Cal Jeffrey 11 comments
In a nutshell: Last week, an international coalition of scientists and government agencies voted to end the leap second, much to the relief of standards organizations and the tech industry. France's International Bureau of Weights and Measures (BIPM) and the National Institute of Standards and Technology (NIST) were just a couple of measurement authorities that chimed in on the vote.
The atomic clock was introduced in 1967 as a precise measure of the passage of time. It uses the vibrations of radioactive atoms and is accurate up to 1/15 billionth of a second per year, depending on the element used. Unfortunately, the Earth's rotation is not nearly as consistent as it varies and decays.
So in 1972, timekeeping authorities introduced the leap second to keep Coordinated Universal Time (UTC) and the standard day/night cycle in sync. It was an adjustment that asked devices to add one second roughly every 21 months. It seemed like a simple enough solution at the time, but it almost immediately proved problematic for computer programmers and the broader tech industry right out of the gate --- an issue that has continued into the new century.
In 2012, a leap-second bug took down several websites, including Reddit, Mozilla, Gizmodo, Lifehacker, and more. In 2017, Cloudflare had a DNS blackout caused precisely at midnight on January 1 when the leap second kicked in that year. This year, Meta wrote a lengthy argument about why we should abandon the time adjustment.
Unlike the leap year (technically a leap day), which adds a day every four years, always ending February on the 29th instead of the 28th, leap seconds are far less predictable. While the suggested adjustment is approximately every 21 months, actual changes are based on the Earth's irregular spin. Since one second is a minuscule time increment, it is hard to pinpoint when the synchronization occurs upon reaching the target year. It's like trying to accurately hit a dart board when the distance to it keeps changing.
Around the turn of the 21st century, timekeepers began realizing that changes in time can negatively affect computers and software designed to deal with immutable time. It's hard to say whether these thoughts were provoked by the Y2K scare, but that seems reasonable.
They all agreed that something needed to be done, but nobody could devise a good solution. It took over two decades of debate between concerned entities to decide to simply end the leap second. However, Resolution D will not take effect immediately or continue indefinitely. The rule goes into effect in 2035 and will remain valid until 2135.
The International Telecommunication Union (ITU), responsible for transmitting universal time, still has to vote its approval at the World Radiocommunication Conference in Dubai next year. However, it is considered a formality since negotiations between the BIPM and ITU, so far, have proven favorable, indicating it is on board with the change.
The New York Times, notes that many in the standards community were thrilled with the near-unanimous decision.
"Unbelievable!" Time Department Director Patrizia Tavella with France's BIPM said. "More than 20 years of discussion and now a great agreement. [I] was moved to tears."
The National Institute of Standards and Technology's chief of time and frequency, Elizabeth Donley, said, "It feels like a historic day. There's probably a lot of celebrating being done in style [in France]."
Russia and Belarus were the only countries that did not vote to pass Resolution D. Belarus abstained, while Russia cited complications involving GLONASS --- its system of global positioning satellites. They are hard-coded to account for leap seconds automatically.