Geologists have a bit of awkwardness in their language in talking about the deep past: distinguishing dates from durations. None of us has a problem with the weirdness of historical time--in 2012, we can easily say that an event in 200 BCE happened 2211 years ago, and that an object made back then is 2211 years old today. (Remember, there was no year 0.)
Among geologists, a widespread practice has arisen in the last few decades that gives dates (not ages) in the format "X Ma"; for example, rocks that formed 5 million years ago are said to date from 5 Ma. "5 Ma" is a point in time that is 5 million years from the present. Instead of saying that the rock is "5 Ma old," geologists use a different abbreviation like m.y., mya, myr, Myr or whatever. This is a little awkward, but context makes things clear.
Recently the International Union of Pure and Applied Chemistry (IUPAC) and the International Union of Geological Sciences (IUGS) convened a task force to decide on an official definition of the year to go in the Système International or SI, the "metric system." The exact definition isn't important here, but the symbol they chose, "a," would happen to override geological custom by requiring everyone to use "Ma" (and ka and Ga, etc.) everywhere. That would make writing geology papers somewhat harder, but I suppose we could adjust.
But Nicholas Christie-Blick of Columbia University has looked more deeply at the proposal and cries foul in GSA Today. He raises a question I'm sympathetic to: How can SI accommodate the year as a "derived unit" when SI rules require that these must be simple powers of base units? There's no room in the rules for a derived unit called the year, which would be defined as 31,556,925.445 s. Derived units are things like the gram (10-3 kilogram). If this were a legal dispute, Christie-Blick would be arguing that the year has no standing.
"Start over," he says, and get buy-in from geologists. I agree.