Although "safety first" was the watchword of shuttle launches, aerospace engineers have repeatedly complained that belt tightening and shifting priorities were denying Columbia and the three other shuttles the necessary upgrades and improvements.

As recently as last April, the former chairman of the Aerospace Safety Advisory Panel warned Congress that NASA's management of the shuttle program had drawn "the strongest safety concern the Panel has voiced" in 15 years. "I have never been as worried for space shuttle safety as I am right now," the chairman, Richard Blomberg, said.

None of this was supposed to happen. The last lethal shuttle disaster, 17 years ago, provoked calls for a revolutionary changes in the program's management. The agency promised that safety would henceforth be put far ahead of all other considerations, including budget constraints, the demands of its users, and any political pressures.

"We will never launch when it is unsafe," Fred Gregory, then NASA's director of space flight, promised the House Science and Space subcommittee nine months ago.

While none of those who issued warnings pointed specifically to a defect immediately known to be implicated in the new disaster, they warned repeatedly that safety was losing out in the battle for scarce NASA funds. The program's 40 percent budget decline over the last decade had undermined its ability to guarantee flawless performances, they said.

NASA's response was mostly to say it disagreed. The problems were not that bad; safety was still the top priority; the number of shuttle "anomalies" or defects was dropping fast. "NASA will continue to ensure that an adequate staff and shuttle workforce" is available to maintain a perfect record, Gregory promised.

But safety experts have long said NASA's claim that safety was improving stemmed from an illusion. The shuttle, they said, was an aging, balky, and delicate space truck that exceeded NASA's own risk limits for manned flight, and time was not its friend.

The ungainly glider was created in the 1970's through a marriage of adventurous design and well-known technology, and it was considered underfunded from the outset. By all accounts, the program has never really embraced the past decade's stunning advances in aerospace engineering and safety testing.

After the shuttle Challenger exploded on launch in 1986, for example, numerous safety advisors urged that a crew ejection capsule be added to save lives even in the midst of calamity. "There is a clear need...to develop a plan to address the absence of an escape system by either upgrading the space shuttle or initiating a program with a realistic timetable to replace it," the Aerospace Safety Advisory Panel concluded last year.

NASA has studied the problem for years, but the costs of retrofitting such a device kept it from acting. As a result, Columbia's crew had no choice but to follow the craft's fate as it broke up around the point of re-entry into the earth's atmosphere.

No new shuttle has been in development, and in fact, many of the most recent safety alarms stemmed from the agency's recent plan to try to extend the life of the current shuttles by another 25 years. Blomberg warned in particular that budget tightening compelled the shuttle program to spend most of its resources on current operations while planned improvements, including some that would "directly reduce flight risk" were deferred or eliminated.

"The concern is not for the present flight or the next or perhaps the one after that," Blomberg said last April. "One of the roots of my concern is that nobody will know for sure when the safety margin has been eroded too far."

"Repeated government and contractor hiring freezes" during the shuttle's operating life "have led to a lack of depth of critical skills" that become more troubling as the system ages, Blomberg said.

In an implied criticism of Congress and the White House, the panel said NASA's budgets were "not sufficient to improve or even maintain the safety risk level of operating the space shuttle. Needed restorations and improvements cannot be accomplished under current budgets and spending priorities."

In mid-2001, five of the nine members of the aerospace advisory panel and two consultants were asked to step down after NASA changed its charter and required rotating memberships. They were replaced but the year "was one of significant upheaval on the panel," the panel said in a report last March.

The most detailed independent assessment of shuttle safety in recent years, completed in March 2000, pointed to specific problems that analysts Saturday said may have played a role in the Columbia's breakup around the time of its re-entry into the earth's atmosphere.

It called for scrutiny of wiring problems in "difficult-to-inspect regions" of the Columbia shuttle, in particular, a problem that NASA said it had fixed. It also said that NASA was not using the latest scientific techniques to find and fix structural cracks and other consequences of routine aging. The panel said further that NASA was not working hard enough to find and fix corrosion beneath the tiles that protect the shuttle from intense heat during re-entry and that it was not working hard enough to find a way to probe or study portions of each shuttle's structure ­ on average one-tenth ­ that are entirely inaccessible.

"The large reduction in NASA Quality Assurance Inspectors for each shuttle is very disturbing," said the panel, which was chaired by Henry McDonald, the director of NASA's own Ames Research Center.

Some of the safety alarms stemmed from what experts have described as inadequate NASA oversight of those parts of the program that have been privatized. Just two days ago, for example, the Government Accounting Office described NASA's management of all its major contractors as "weak" and "debilitating," and accused the space agency of placing "little emphasis on end results (or) product performance."

The GAO report was the latest in a series by the congressional auditing agency faulting NASA's management of major programs, including the shuttle. Weak contract management had been ranked as "high-risk" problem at NASA since at least 1990, the report said.

The increased fiscal pressure on NASA is partly the result of steep budget cuts over the past decade. NASA and other civilian agencies involved in the space program saw their funding slashed by $1 billion in fiscal 2002, while Department of Defense spending on space programs rose by $600 million, according to a recent study by the Aerospace Industries Association, an industry trade group.

"The civil space program that NASA runs has been neglected for a generation and as a consequence we find ourselves flying increasingly aged technology," said Loren Thompson, a defense industry analyst for think tank Lexington Institute.

In 1996, NASA turned over space shuttle flight operations to United Space Alliance, a private firm owned by Boeing Co. and Lockheed Martin Corp. Under pressure from the Clinton administration and Congress to cut costs, NASA had gradually shifted many responsibilities to the private sector.

United Space is now considered the prime contractor for the space shuttle program and manages about a third of the program's budget. In addition to its role as part of United Space, Bethesda-based Lockheed also provides many crucial functions, including construction of the external tank that feeds liquid propellant to the shuttle's three main engines. It also developes the electronic systems that perform navigation, guidance and flight control for the space program and manages data collection, said spokesman Tom Jurkowsky.

While NASA managers have described their contractor oversight as adequate, NASA's Office of Inspector General disagreed. "The lack of systematic and well-documented contract surveillance is a particular area of concern," the Inspector General said in a report last June.

In response to such reports, NASA has sometimes sought increased funding, added backup systems and new safety routines, and taken other steps designed to bolster its already complex procedures for preventing accidents. But NASA has always acknowledged that the program would never be 100 percent reliable.

After the post-Challenger safety upgrades, NASA estimated there was a 1 in 250 chance of fatalities among its crew. "The shuttle is a wonderful machine, but it is at that left-hand bar of risk to humans," Sam Venneri, the head of NASA's Aerospace Technology office, said in congressional testimony last year.

John Pike, a director of GlobalSecurity.org and a longtime critic of the agency's shuttle management, said even this estimate is speculative. "NASA has been wildly unrealistic about shuttle reliability. If you go back and look at their reliability estimates, I think that they basically just made them up," Pike said.

"There's a statistical chance every time you launch that there could be an accident. This is just a very hard thing to do," John Logsdon, director of the Space Policy Institute at George Washington University, said Saturday. "It's only in retrospect that you can say that it wasn't enough. Second guessing is easy. Hindsight is wonderful."

"What people have done to keep an old system flying is just amazing. But it's an old system. At some point they had to expect something to go wrong," said Donna Shirley, a former Mars exploration manager at NASA and now an instructor in aerospace engineering at the University of Oklahoma in Norman. "It's remarkable that they've kept it going this long."