Working Group 7A of the International Telecommunication Union's Study Group 7 may sound like an anonymous international committee like any other. But this is no quango of grey bureaucrats in greyer suits arguing over the desired colour of toilet paper. At the heart of this group's discussions is something of fundamental importance to anyone who has ever taken a second to fall in love or to score a goal: time itself, and how to define it.
Unbeknown to most people there is not a single accepted way of telling the time, but several different scales running concurrently. The differences are usually small, but the scales can be as much as 30 seconds apart and the gap between them is growing steadily.
Aircraft navigation systems tell a different time from the watches of passengers, pilots and air traffic controllers. Experts are warning that this could spell disaster.
"We should only have one type of timescale throughout the world," says Bill Klepczynski, a time expert who advises the federal aviation administration. "There's a possibility for danger."
The International Telecommunication Union - the global body that agrees time standards - is taking the issue seriously, and has set up the working group to advise it what to do. "We're trying to gather data on how people are using time, what sort of problems they have and whether or not a contiguous timescale would be beneficial," says Ron Beard, who heads the group.
But the plans have not pleased everyone, and arguments about the best way forward are rattling the usually steady world of timekeeping.
The problem arises because the Earth cannot keep time as accurately as modern atomic clocks, which count the steady shaking of atoms. These atomic clocks replaced the motion of the Earth as the world's official timekeeper in 1967. The pull of the moon is gradually slowing our planet down, so every now and then our clocks are halted for a second to let it catch up.
The first of these "leap seconds" was introduced in 1972, mainly as a favour to astronomers and others who still relied on the old-style celestial time. A further 31 leap seconds have been added since, most recently on December 31 1998.
And that would be that, were it not for the fact that the precise timekeeping offered by atomic clocks is now becoming widely available - most commonly through the satellite global positioning system used for navigation. To add to the confusion, GPS uses yet another timescale.
It includes the leap seconds added until the GPS clock was set in 1980, but has ignored those added since. This means GPS time is now running 13 seconds ahead of coordinated universal time - which includes all added leap seconds and to which most clocks on Earth are set - but is some 19 seconds behind international atomic time, which is based on atomic clocks and ignores leap seconds.
This multiplicity of timescales is increasingly dangerous as the systems diverge with every leap second added, warns Mr Klepczynski. "We need to go to a uniform timescale," he says. "When you have these planes navigating and flying around, what time system do you use to coordinate everything?"
Widening gaps between the GPS time used by aircraft navigation systems and the time used on the ground could generate confusion between a plane's reported and actual position, he says, and so increase the risk of a collision. A complicated situation will get more complex still when Europe launches its own GPS system, Galileo, which will be based on yet another version of time.
Computer software converts between the different timescales used. "But if anybody ever makes a mistake there's going to be a big problem," Mr Klepczynski says. His solution is to scrap the leap second, effectively merging atomic time and universal time. This is also the proposal being considered by the time lords of the ITU.
One group opposed to the scrapping of the leap second are astronomers, whose sensitive telescopes still rely on time set by the Earth's rotation. Switching to atomic time would throw their instruments out of kilter, and leave them facing costly upgrades. "The astronomers are up in arms about this," Mr Beard admits.
Changing the system could store up problems for our descendants. Without the braking effect of leap seconds, our clocks would steadily run faster and faster than the Earth's rotation, with the effect that the sun would rise later and later in the morning.
"I'm concerned that they're trying to implement a plan that will ultimately turn day into night," says Rob Seaman of the National Optical Astronomy Observatories in Arizona.
To stop this happening, we would need to introduce a leap hour every 700 years or so, in a similar way to how we change our clocks to account for summertime.
Since the debate began, the slowing of the Earth has become less pronounced and no leap seconds should be needed for several years. Experts are unsure exactly why this has happened -a number of factors can have short-term influences on its rotation, including earthquakes and even wind blowing on mountains - but they agree that the constant drag of the moon means the slowing will soon pick up again, and within a few decades we could be forced to add two or even three leap seconds a year.
As well as possible safety risks, that raises some intriguing legal issues, according to Dennis McCarthy of the US naval observatory.
"We face possible problems in the timestamping of electronic documents," he says. This is because a leap second is usually added at the end of the day, by asking clocks to change from 23:59,59 to 23:59,60 before going on to 00:00,00. But as most clocks don't permit the number 60, they show 23:59,59 for two seconds instead.
"I suspect it will happen eventually that someone says their 23:59,59 refers to a different 23:59,59 and lawyers will become involved," Mr McCarthy says. This could be important for legal or financial documents detailing the sale of bonds and securities at a specific time.
This may all sound terribly worrying to the people who stocked up on tinned food and fled to the hills over new year 2000, in case the millennium bug struck the world's technology. But there seems little immediate need for concern for computer owners.
"If computers share common files then some pieces of software could get upset if two computers differ by more than a second, but I couldn't tell you a common application that would," says Markus Kuhn, a computer scientist at Cambridge University.
Back in the committee rooms of Working Group 7A, the decision on the future of time is, well, some time away. Mr Beard says his group is unlikely to report before 2005, and the ITU will then have to get any suggested changes ratified by its member countries.
What time is it? It could be a while yet before we know for sure.
Watching the clocks
Coordinated universal time
The uniform timescale that forms the basis for most civil timekeeping in the world. UTC is based on atomic clocks, such as the one held by the National Physical Laboratory in Teddington, south-west London. Some 32 extra leap seconds have been added to UTC since it was officially adopted in 1972, to account for the fact that the Earth's rotation is gradually slowing down.
International atomic time
A statistical timescale mostly used for scientific reference. The Bureau International des Poids et Mesures in Paris sets TAI time by monitoring the regular vibrations of caesium atoms in atomic clocks around the world. Coordinated universal time is generated from TAI by adding leap seconds. TAI is currently 32 seconds ahead of UTC.
GPS time
An atomic timescale used by the US global positioning system. When it was set in 1980, GPS time was based on coordinated universal time, but GPS time is now some 13 seconds ahead because it does not count leap seconds. A Russian GPS system called glonass does account for leap seconds, but adding them has caused technical problems.
Greenwich mean time
A time standard established for British navigation in the mid-19th century. GMT has now been officially replaced by coordinated universal time, so Big Ben, the BT speaking clock and the BBC radio pips all mark UTC, not GMT as some people think -although the two are usually very close. British law still refers to GMT because a 1997 bill that tried to update it to UTC was never passed. It ran out of time.