On leap day, let us pause to consider the future of a much smaller time correction: the leap second. Last month in Geneva, the International Telecommunication Union, an arm of the United Nations that manages timekeeping, postponed for three years a decision on whether to abolish leap seconds altogether.

This much-needed delay will allow the world to assess exactly what it might cost scientists, computer engineers and the rest of us to lose a tiny but powerful mechanism for keeping track of real time.

Leap seconds have existed only since 1972, because that was the point when human clocks became so precise as to fall ever so slightly out of sync with the actual day -- the time it takes the Earth to spin once on its axis. Because a complete rotation lasts a couple of milliseconds longer than 24 hours calibrated to atomic transitions in cesium-133, an extra second is added to our master clocks every couple of years or so.

The U.S. government has argued that leap seconds are a bother no one needs. That adding them only risks disrupting the smooth running of air traffic, satellite navigation, financial transactions, power grids and other networks that rely on synchronized clocks.

Another difficulty in managing leap seconds is that the Earth’s rotation is so irregular -- influenced as it is by such forces as weather, sea level and changes in the Earth’s core -- it’s impossible to predict very far in advance when one will be needed. Official timekeepers are given only six months’ notice. (The next leap second is scheduled for June 30.)

But no one has yet looked into what problems might arise from allowing the disparity between clock time and Earth time to grow. The immediate worry is not so much that our clocks and watches would drift perceptibly out of line with familiar markers of day and night -- dawn, noon, sunset. It would take about half a century, after all, for atomic clocks to get ahead of the natural day by even a minute, and hundreds of years to be off by an hour.

But think how the drift could affect astronomers, who must be able to aim their telescopes at specific stars and galaxies, which cross the sky on Earth’s own time, heedless of atomic clocks. A slight asynchrony might be trouble, too, for controlling the navigation of satellites. It’s important to find out.

One thing we know is that, in the past 40 years, the 24 leap seconds have caused no trouble. As clocks grow ever more sophisticated, shouldn’t they only become better at adding the extra seconds seamlessly?

As Canada, China, Germany and the U.K. argued in Geneva last month, it would be a mistake to do away with leap seconds without first looking into the possible consequences. We should not be in a hurry to let go our ancient connection to astronomical time.

So we applaud the ITU’s decision to wait three more years.

Before making any decision to abolish the leap second, the ITU should also figure out how it would then reckon with the growing gap between atomic and astronomical time. To do nothing is not a solution. It would only leave a problem for future generations to solve.

Several years ago, some scientists suggested scheduling a leap hour for the year 2600. This idea was abandoned as impractical, given that the instructions would have to be left for people six centuries hence. But could there instead be, say, a leap minute every half century?

Let’s hope three years is enough time to figure this out; the clock has started.

Read more opinion online from Bloomberg View.

To contact the Bloomberg View editorial board: view@bloomberg.net.