Agency officials say they will consider the recommendations in the coming months, as NASA re-evaluates the way it does business. But Goldin told lawmakers at a House Appropriations Committee hearing March 15 that "I probably pushed too hard" to reduce costs and tighten schedules. "I feel bad," he added, "But I won't apologize. We had to push to see what the limits are."

Typical NASA projects -- such as the Mars Viking landers -- used to cost $1 billion or more and take a decade or longer to plan, build and launch. When Goldin took over the space agency in 1992, he ordered his managers instead to conduct a mission which would cost as little as $100 million to $300 million and take three to four years. The fruits of that change have been successes like the Mars Pathfinder, an innovative spacecraft which captured the world's attention in 1997 as it bounced to a halt on the martian surface and sent out a tiny rover. But the recent twin loses of a Mars lander and orbiter, coupled with the failures of the Lewis and Wide-Field Infrared Explorer missions, have raised questions about NASA management.

The Baker panel found that adding small missions has been a boon to Earth and space sciences, and that the shorter cycle has lowered costs, allowed the use of more modern technologies and allowed scientists to provide more up-to-date instruments. And simply having more missions launched more frequently has energized the research community. "Scientific objectives can be met with greater flexibility by spreading a program over several missions," according to the study, which was written by seven university professors and industrial managers.

On the downside, the panel found that "heavy emphasis on cost and schedule has too often compromised scientific outcomes," while NASA has confused technology development with the scientific objectives of missions. In addition, the faster-better-cheaper approach can lead to risky designs, management practices and technologies. And "the present emphasis on smaller spacecraft is inconsistent" with the uncomfortable fact that small launch vehicles are hard to find.

To fix these problems, the panel urges Goldin to put science first in choosing a mission's objectives. NASA also should have a more vigorous technology program, and consider ways to use foreign launchers so that the bottleneck of vehicles is not so acute. And the panel also encourages increased international collaboration, which, it notes, "is sometimes considered within NASA to be detrimental" to a mission.