Upon taking the helm of NASA’s Science Mission Directorate in  March, Alan Stern promised a get-tough policy for projects and managers that fail to live up to their budgetary promises. Specifically, he said, missions whose costs spiral out of control will face termination, as will principal investigators who allow this to happen on their watch.

Mr. Stern has yet to take any such drastic action, but in holding the line against further cost growth on the planet-hunting Kepler space telescope, he served notice that he is serious about reform. In the meantime, he has preserved, at least for now, a worthy mission, while offering some hope that NASA’s Science Mission Directorate can still get much accomplished at a time when the agency’s budget remains too small to accomplish everything Congress and the White House have put on its plate.

The Kepler story is, unfortunately, all too familiar for one-of-a-kind space science missions: a variety of factors, both internal and external to the program, cause costs to swell beyond the initial estimates. The mission managers approach NASA’s leadership, plead their case, and typically are given what they say they need to finish up the project. This cycle is then repeated any number of times until the mission’s final cost bears little resemblance to the original estimates.

Until recently, Kepler, a Discovery-class mission whose costs already have increased 21 percent, to more than $550 million, appeared to be following the script. But when the Kepler team came to Mr. Stern this spring to ask for another $42 million, his answer was no; when they had the audacity to come back June 1 requesting $54 million more, he said, in effect, hell no.

And lo and behold, the industry and government team members donned the green eyeshades and came up with a way to complete the mission without any additional extra funding.

The new plan, which cuts six months from the tail end of the mission, reduces program reserves, and perhaps most significantly, shaves millions of dollars from the award fee of spacecraft and instrument contractor Ball Aerospace & Technologies Corp., is not without risk. And of course, there is no guarantee that there will not be additional cost increases as Kepler’s November 2008 launch date approaches, a time when it becomes incredibly difficult to make good on threats to terminate a mission.

Moreover, it is far too early to tell whether Mr. Stern will be successful in his broader cost-control initiatives, such as getting more reliable cost estimates before missions are approved and holding principal investigators accountable for what happens on their projects.

But Mr. Stern already has shown what can be accomplished with a little backbone, and with that, he is off to a good start.