In March of 1994, the National Aeronautic Association announced that its 1993 Robert J. Collier Trophy would be awarded to the NASA Hubble Space Telescope Recovery Team "for outstanding leadership, intrepidity, and the renewal of public faith in America's space program by the successful orbital recovery and repair of the Hubble Space Telescope." Representing the more than 1,200 men and women directly involved in the mission, the seven-person astronaut crew of Space Shuttle Mission STS-61 and four ground managers were named as the recipients. It was truly, by all assessments of the participants and the observers and in the language of the Collier award, "the greatest achievement."

The Hubble Space Telescope had taken longer to build and launch than any other NASA spacecraft, including Apollo. It had cost more than any other scientific space program, and more than nearly any other space mission. Deployed in the Spring of 1990 with the express mission of addressing the most enigmatic and exciting questions of astronomy, it promised a revolution in understanding the origin and evolution of the universe and myriad other astronomical and cosmological questions. Within weeks, a horrible realization gradually emerged: the instrument bore a seemingly fatal and irreparable manufacturing flaw that would severely degrade or even scuttle its fifteen-year mission. Worse, various other systems and components began to act erratically or to fail. Soon the Hubble Space Telescope and NASA itself were the objects of anger, scorn, and ridicule. What began even before launch as a planned and routine servicing mission grew into a bold and comprehensive overhaul, a "rescue mission in space." The Hubble Space Telescope Recovery Team rescued more than just the telescope and its mission, however. By all accounts they rescued NASA and the U.S. space program as well.¹

This article is dedicated to Bob Bless, leader of the High Speed Photometer Team at the University of Wisconsin, who worked tirelessly for decades and in many ways to see the Space Telescope achieved, and then gracefully made room for COSTAR.

I am grateful to a very large number of people who graciously gave freely of their time by sending or giving me information, oral history interviews, and informal discussions. The Hubble Servicing Mission was a massive effort that drew a dozen institutions and thousands of people into its cast. I have tried to give representative weight to the various contributors, but it would have been impossible within the limits of this brief article even to mention all their names. My own prior knowledge of the program from having worked on the Space Telescope History Project and the restoration and exhibit of the Space Telescope Structural Dynamic Test Vehicle was an important foundation for writing this article, and I am grateful to colleagues and friends at the Smithsonian Institution's National Air and Space Museum, especially Robert W. Smith. Finally, as ever I am grateful to the NASA History Office, especially Director Roger Launius, for supporting this article. Needless to say, I am solely responsible for all conclusions and interpretations herein.

Astronaut F. Story Musgrave, anchored on the end of the Remote Manipulator System (RMS) arm prepares to be elevated to the top of the towering Hubble Space Telescope (HST) to install protective covers on magnetometers. Astronaut Jeffrey A. Hoffman (bottom of frame) assisted Musgrave with final servicing tasks on the telescope wrapping up five days of space walks. (NASA photo no. 94-H-16).

In the introduction to this volume, Pamela E. Mack discerns a number of trends that altered the character of air and space flight over the eight decades of Collier awards:
 

1.  "NACA and then NASA [became] increasingly caught in a web of bureaucratic and political obligations", individuals came to matter less and the planning process itself became more important.
2. "Research and development projects have become more complicated in fundamental ways, " requiring more diverse expertise than any lone inventor could muster and separating developers of technology from users.
3.  "Attitudes towards funding research have changed"; after Apollo there arose a "new emphasis on cost-benefit calculations but also more willingness to fund projects on the basis of popular support. "²

(1) Bureaucratic and political obligations: Between the initial planning in the 1970s for frequent routine shuttle maintenance of the telescope and the December 1993 servicing mission, replacing components came to acquire a heavy load of significance and implications. The servicing mission became a way to repay bureaucratic and political obligations that extended far beyond the telescope program itself, even beyond NASA. In late 1993, the then-tarnished reputation of NASA as a whole, the capability of its shuttle system and astronaut corps, and the viability of its most important next program, the Space Station, all rode with the crew of the Endeavour. It seemed they would all come back in the same condition: heroic and vindicated, or disgraced. The servicing mission was a resounding success on nearly all counts. The long and demanding spacewalks were among the most unproblematic ever done, and over several days press and media coverage of the mission was more intense and more favorable than perhaps at any time since Apollo. Live television of the repair activities was carried uninterrupted on many cable television channels, and scenes of the astronauts working on the telescope dominated the evening newscasts. Astronaut Story Musgrave recalled being stopped by people while doing his Christmas shopping shortly after the mission: "They were bleary-eyed from staying up all night. We were better off than they were.... they were very excited because they had lived it vicariously."⁴

Tellingly, the crew and ground managers were joined by beaming agency officials and politicians in postmission press conferences and events. When the telescope

resumed operations and the repairs were fully tested in January 1994, its performance was far better than anticipated, and the end-users were extremely pleased. To what degree the Hubble Space Telescope was restored to its original planned performance is a difficult judgment, and to some extent depends on the criteria one uses. It was, however, dramatically improved. Over the following year, it produced a string of significant results that would not have been possible in its original state.⁵

(2) Users and developers of complex technology: The complexity of the process and alienation of the users from the developers is amply visible in the history of the space telescope itself. This aspect is documented and illustrated in published scholarly history, government audits by such units as the General Accounting Office, congressional hearings, and memoirs. From the first organized campaigns by astronomers in the late 1960s, through feasibility studies in the 1970s, development in the 1980s, and operations in the 1990s, the Hubble Space Telescope has been sustained by a complex and protean coalition of diverse parties. If astronomers are the end users (but in no way the only beneficiaries) of the space telescope, then in part the history of its development is one of astronomers struggling to prevent their interests from being submerged and overrun by all the others. The need for the servicing mission at all arose because the users (astronomers) failed to protect fully their interests and allowed the developers (engineers and managers) to produce what looked like a space telescope but did not function as a space telescope should. This illustrates big science as what James Capshew and Karen Rader have called "at once a broadly diffused mode of cognition and a concentrated form of organized labor."⁶

Not that the astronomers did not try —they tried mightily over the decades, singly and in groups. But the scope of what Robert Smith has called "the biggest kind of big science," big engineering, and big management was overwhelming. In my view, the system that was building the space telescope was bigger than any putative system builder (in the sense of historian Thomas Hughes) could ever hope to manage. Nor should we assume there was bad faith or villainous subversion. The users (astronomers) were not the only, nor necessarily the most important reasons for building it at all. Many other interests (engineering, commerce, politics) wanted a space telescope, and the astronomers only got one because these other interests wanted it. These other interests, however, did not require that the telescope function in quite the same way as the astronomers wished. These other interests did not willfully ignore or subvert the astronomers' requirements; mostly they were just insensitive to the astronomers and far more powerful. Their needs for a successfully functioning space telescope were far more relaxed than

the needs of the astronomers. Even though all the different interests assembled around a single big machine, it was an instrumentality with many meanings and many criteria for success. One manager put the issue succinctly: "If the agency did nothing more than take one unaberrated picture from the Wide Field Camera in 1993, it would have been declared a success; so why add any risk to [the servicing mission by fixing other problems]? Well.... we promised a functional telescope, not a stunt."⁷

The servicing mission was successful in part because the users asserted themselves and interjected themselves forcefully into the process. The idea for the optical fix came from the user advocates. Implementing the fixes was possible because the other interests found their salvation through the needs of the users. If the development of the telescope illustrates the alienation of the users in big science and the dispersion of interests, its repair provides an illustration of how big science can work beautifully when interests merge. The servicing mission evolved from a diffused set of activities with many parties operating somewhat independently into a tightly focused and exquisitely choreographed group effort. It was well funded, had the rapt attention of management, and a working level espirit de corps that transcended institutional and other loyalties.⁸

(3) Attitudes towards funding: Some ill-chosen hyperboles aside, the Hubble Space Telescope's benefits were as pure as the driven snow. It is hard to quantify the value of knowing better the nature, history, and origin of the universe in order to compare it to the cost of the telescope. The general public and its representatives might not comprehend fully the subtleties of astrophysics and cosmology, but they did expect at least entertainment and edification for the investment. While the several problems of the telescope affected all the instruments, not just the cameras, it was not fuzzy spectra or muddy digital data that became the public scandal, but rather "blurred vision." The myriad of misleading and horribly muddled similes and analogies used to convey the problems and remedies all came down to the easily and instantly understood pictures and clarity. Even Senator Barbara Mikulski, herself one of the most knowledgeable and savvy advocates, referred to her outrage at the "cataract" and her elation at the successful "contact lens." Hence, the Hubble servicing mission

was in some sense a redemption, because the defects in the extremely expensive instrument were immediately graspable by the public, who could also easily see the improvement after the repair. They could also easily grasp the magnitude, if not the specific dollar amount, of its cost. As one of the astronomers involved in the program teased in a 1984 lecture:

"That is a nice topic," said Alice, "I will put the project in its historical perspective and draw analogies with other great scientific projects."

"Nonsense!" said Humpty Dumpty, "The answer is very short. Space Telescope will revolutionize mankind's understanding of the Universe because it cost a billion dollars! That's all. They wouldn't have spent the money in the first place if it hadn't been so, now would they?"⁹

Initially, maintenance and refurbishment missions (including even return of the entire spacecraft for ground overhaul and then reflight) were part of the Space Telescope's routine scenario. The first official NASA telescope planning phase began around 1971, just as the Space Shuttle was being defined. With austere times ahead after Apollo, the Space Telescope and the Space Shuttle soon found common cause. An orbital telescope of that size and cost could not be justified unless it could operate for years or decades, and to that end the Space Shuttle promised routine access for repair and upgrade. Lacking the Space Station, deleted from planning due to cost, the shuttle needed a place to go and useful work to do that could not be accomplished by expendable boosters. To that end, the Space Telescope and Space Shuttle pair became an exemplar of a new and cost-effective way of doing Earth orbital science. Initially projected at about two-and-a-half-year intervals over the fifteen-year life of the Telescope, such service calls were expected to be unproblematic and routine. Indeed, the frequency of access to the spacecraft, predicated on the presumed regular and routine shuttle traffic, was to have removed urgency. If something could not be accomplished on one call, it could wait until the next. It also meant that expensive pre-flight design and testing could be relaxed somewhat, secure in the knowledge that the spacecraft would be accessible for servicing. This notion of routine maintenance remained with the Telescope through its planning in the 1970s, its new start approval in 1977, and through some traumatic budgetary times in 1980. However, several unanticipated circumstances emerged to cast the spotlight on maintaining the Hubble Space Telescope and to raise the stakes.¹⁰

First, early in 1983 NASA Administrator James Beggs was startled to be told, on the very eve of release of the President's budget for fiscal year 1984, that the program was once again significantly over budget and behind schedule. The full extent of the crisis emerged over the following few months, and provoked painful congressional hearings,

shakeups of the program and project management, and serious loss of political capital. By the time the new launch target date was in place, and the new budget understood, the program was operating under "last chance" understandings. It could not go back to the well again, and the program remained under scrutiny from within and without NASA. One salutory effect of the reorganization was that some orbital replaceable units, deleted earlier in the program for budgetary reasons, were reaffirmed, making the Telescope more easily maintainable than it otherwise might have been.¹¹

Second, after recovering from this setback, the spacecraft was waiting in storage and testing at Lockheed's Sunnyvale, California, plant when the Space Shuttle Challenger accident in January 1986 halted all shuttle missions for an indefinite time. The Telescope program had to keep its "marching army" idling but consuming money during the resulting hiatus, contend with more conservative rules for shuttle bay payloads, and worry about the potential effect on the spacecraft of the unplanned-for extended storage. Moreover, the more conservative shuttle schedule, and even tighter budgets that emerged in the post-Challenger era, meant fewer and less-frequent future maintenance opportunities. The program was forced to absorb the additional costs without ability to obtain more money from without. The maintenance and refurbishment budget was an easy target, and money was shifted by adjusting the maintenance schedule and that for future replacement instruments. Some worried about how the spacecraft and its components would fare, not having been designed to reside in storage for so long. Others worried that new instruments would not be ready in time to replace inevitably aging and failing ones.¹²

Third, the Hubble Space Telescope became the centerpiece for recovering from Challenger after effects. It was the largest and most complex scientific payload to be delivered to orbit, except perhaps the Galileo Jupiter probe. Unlike Galileo, which would take several years to reach its destination and return data, the Hubble Space Telescope would start providing results in only a few days or weeks of deployment. As launch neared, the press and media coverage was enormous.

When the Space Shuttle Discovery roared from the launch pad on April 24, 1990, occupying nearly all of the payload bay was an enormously complex spacecraft. Rather than the usual spacecraft with instruments attached, the Hubble Space Telescope was an immense collection of instruments enshrouded by a spacecraft and various appendages. At the heart, a cylindrical Optical Telescope Assembly held a 2.4 meter diameter primary mirror which would first receive the light from astronomical objects, reflect it forward to a 0.31 meter secondary mirror, which would then send the light back through a central hole in the primary mirror. just behind the primary mirror five scientific instruments and three fine guidance sensors, themselves large and complex assemblages of optics, motors, and electronics, would share the bundle of precisely focused light. Surrounding the Optical Telescope Assembly, the Support Systems Module contained dozens of electronic and mechanical black boxes to operate the ensemble, and sprouted two deployable antennas and two large solar arrays to generate electricity.

Designed to be serviceable in orbit, the Hubble Space Telescope was the most mechanic-friendly spacecraft ever flown. Dozens of bright yellow hand holds and sockets for the astronauts' portable foot restraints were strategically located around the spacecraft. Inside the multitude of hinged doors, all of the instruments and many of the black boxes could be removed and inserted by a space-suited astronaut using only a few tools. If motors or actuators failed, the solar arrays could be rolled up (like a window shade) or other appendages deployed or stowed using a ratchet wrench. Some of the deployment crew, notably astronauts Bruce McCandless and Kathy Sullivan, had practiced various emergency procedures for years. McCandless had worked, off and on, for twenty years helping to develop the maintainable features of the Telescope and the special tools required. They had spent many hours in Lockheed's clean room observing and working on the flight spacecraft itself, and in the water tanks at Marshall and Johnson working with various mockups.¹³

On April 25, astronaut and astronomer Steven Hawley grappled the spacecraft with the shuttle's robot arm and eased the 12-ton railroad tank-car sized vehicle out of the payload bay. Still attached to the arm, a carefully orchestrated sequence of commands instructed the spacecraft to deploy its antennas and solar arrays. There was a tense period when one of the solar arrays got stuck, and McCandless and Sullivan donned their spacesuits, waiting inside Discovery's airlock, ready to go outside if necessary. Ground controllers freed the array, however, and in mid-afternoon only one orbit later than planned, Hawley released the Telescope. Within ten minutes the freely flying spacecraft had locked onto the Sun, and later maneuvered itself to point the delicate optics, protected by the still closed aperture door at the front, away from the dangerous sunlight. As Discovery backed away, ground controllers began the complex sequence of commands to "wake up" the various dormant systems of the Telescope and prepare it for its fifteen-year mission. McCandless and Sullivan doffed their spacesuits, and with the rest of the crew continued Discovery's flight, the toolboxes never having been opened. Neither saw the Telescope again, and both left NASA shortly after the mission. Sullivan became chief scientist of the National Oceanic and Atmospheric Administration. McCandless retired from NASA, leaving the maintenance of the Telescope to other astronauts who might never have seen it in person-or so he thought.¹⁴

The euphoria of the successful deployment was mixed with mild concern as the spacecraft encountered some initial problems. These had nothing to do with the mirror, but rather involved a variety of glitches with the communications antennas and the control systems. The spacecraft somewhat spastically and repeatedly shut itself down into several so-called "safemodes" in response to the motions of various appendages, and had to be coaxed back into operation. When, a few days after deployment the fine guidance sensors attempted to lock onto stars, correctible errors were discovered in the programming but a more persistent oscillation prevented the telescope from keeping itself pointed with the required accuracy.¹⁵

While engineers had forewarned before launch that such a complex spacecraft and ground system was bound to take some time to achieve stable operations, these initial problems earned the spacecraft some bad press that, in the words of a Washington Post story, "has turned it into fodder for stand-up comics and prompted some citizens to murmur that it may be a $2.1 billion lemon." On May 15, 1990, late-night comedian David Letterman offered the "Top Ten Hubble Telescope Excuses:"¹⁶
 

10. The guy at Sears promised it would work fine.
9. Some kids on Earth must be fooling around with a garage door opener.
8. There's a little doohickey rubbing against the part that looks kind of like a cowboy hat.
7. See if you can think straight after 12 days of drinking Tang.
6 Bum with squeegee smeared lens at red, light.
5. Blueprints drawn up by that "Hey Vern! guy.
4. Those damn raccoons!
3. Shouldn't have used G.E. components.
2. Ran out of quarters.
1. Race of super-evolved galactic beings are screwing with us.

After stabilizing, but not solving these and other operational problems and taking the first test image a month after deployment, there was a brief period of almost giddy elation. By mid-June, however, some scientists had become more and more worried about subtle characteristics of the star images. This concern turned to near despair as scientists and engineers realized the reason for the Telescope's inability to focus. The heart of the instrument, either the primary or the secondary mirror, or both (it would later turn out to be the primary) had been ground and polished over several years a decade earlier nearly perfectly to the wrong specifications. That this error had gone undetected over the years of testing was something most people within the program found too startling to believe. just convincing them that the problem lay with the optics was difficult-it took nearly a month, from late May to late June. Scientists and managers responded to the suggestion with a mixture of bewilderment and outright denial. Once the analyses had finally become too compelling to ignore, NASA Associate Administrator for Space Science Lennard Fisk was told, and he responded that it might be space science's equivalent of the Challenger accident. A NASA press conference announcing the conclusion on June 27 showed a panel of somber faces, and the official making the first announcement literally stuttered over the words, "spherical aberration." Outside the program and NASA, the news was received with less grace, and more outrage and even ridicule. Congressional representatives were furious, influential Senator and space enthusiast Barbara Mikulski, calling the Telescope a "technoturkey," expressed outrage over Hubble's "cataract." Comedians and editorial cartoonists, from Herblock to Gary Larson's Far Side, could hardly believe their good fortune. Even filmmakers were quick to seize on the opportunity, and before the year was out: "An opening scene in the comedy film Naked Gun 2 1/2 features a dark lounge with a depressing atmosphere, downbeat music, and walls lined with pictures showing historically horrible disasters. There, between a picture of the Hindenburg and a half-sunk Titanic, is the Hubble Space Telescope."¹⁷

Initially, many feared that NASA or Congress would finally decide to write off the loss.¹⁸ The first servicing mission, already planned for 1993, then became much more than a simple scheduled service call: It became the only chance to save the program and the spacecraft from either euthanasia or perhaps resignation to living with its diminished performance. Scientists and managers organized themselves along several strategic lines of work. While fending off the various congressional reviews, NASA Headquarters appointed a "Hubble Space Telescope Optical Systems Failure Review Board," chaired by jet Propulsion Laboratory director Lew Allen. The Allen Committee, as it came to be called, began work in July and within a month had concluded the trouble lay with the main mirror. Investigating the records and the hardware that remained from the fabrication and testing of the mirror in 1980-1981, the Allen Committee found that a simple error in the test setup had skewed the measurement checks on the mirror, and that the computer-controlled polishing machine had dutifully shaped the mirror to the wrong curvature. Discordant test results at the time had been ignored under the schedule and cost pressures, and independent tests were not done. Eventually, the justice Department and Hughes Danbury Optical Systems (Hughes had earlier bought the portion of the Perkin Elmer Corporation that had built the Optical Telescope Assembly) settled out of court.¹⁹ While the Allen Committee was working to determine the cause of the spherical aberration, an HST Strategy Panel at the Space Telescope Science Institute under astronomers Holland Ford and Robert A. Brown began work to determine options for recovering from the mirror problem. The Science Institute had been established long before launch to be the focus for the academic community who wished to use the Telescope. With Goddard managing the mission and controlling the Telescope, the Science Institute would receive proposals from astronomers, manage peer review, and then work with the selected astronomers to schedule and obtain the desired observations. Institute staff included scientists who specialized in calibrating and understanding the Telescope's instruments, and engineers who specialized in merging thousands of approved observation requests into an efficient observing schedule. Under the leadership of the assertive and persistent Riccardo Giacconi, the Science Institute had made itself the watchdog for the scientists, at times much to the consternation of NASA managers. They interpreted their charge broadly, and since well before launch had been proactive without much regard for stepping on government toes.²⁰

At Goddard, the newly appointed Associate Director for Flight Projects for the HST Joe Rothenberg, and the Project Scientist, Al Boggess, met in late July to lay out a strategic recovery plan. First, they would do whatever science could be done with the Telescope in its current state; second, they would try to fix the Telescope's problems at the 1993 maintenance opportunity; finally, they would concentrate on extending the wavelength coverage into the infrared on subsequent scheduled servicing missions that would install replacement instruments. Rothenberg was careful to establish close working relationships with the relevant Headquarters managers and with the Science Institute.²¹

Planners were very lucky that the figure of the mirror was so precisely and uniformly in error. Concurrent with the Allen Committee and the Strategy Panel, a "Hubble Independent Optical Review Panel" under Duncan Moore, University of Rochester Institute of Optics Director, worked to determine the precise prescription for the aberrated mirror, based on records and artifacts at the manufacturer and data taken by the Telescope instruments. The 2.4 meter wide main mirror was too flat by about 2 micrometers, or 1/40 the thickness of a human hair. That meant that an optical element with the reverse prescription could correct much of the aberration. It had already become clear that future replacements for the existing scientific instruments, scheduled to be inserted every
few years beginning with the 1993 servicing mission, could incorporate internal optics that would reverse much of the aberration of the Telescope's mirror. However, the existing five instruments and the fine guidance sensors (used by the astrometry team as a sixth virtual instrument) would remain severely compromised for many years unless some other solution were found. In particular, the European Space Agency's Faint Object Camera had no follow-on in the plans, and it was scientifically as well as politically important to find some way to address the European concerns. The primary mirror defect had two serious consequences. First, because the light rays were not precisely brought to a single focus, the images would lack resolution and the other instruments could not pick out individual objects in crowded fields or very small features of extended objects. Second, because the light was diffused the Telescope's instruments could not reach the planned limits of faintness. These two desiderata, clarity and faintness, were precisely the reasons for putting a Telescope of this size above the atmosphere in the first place. While the Telescope was still capable of providing valuable data and addressing much of its observing program, it was compromised in the areas for which it was supposed to be uniquely suited. The existence of a very serious flaw that was nonetheless amenable to compensation was the foundation for an expanded servicing mission.

The Hubble Space Telescope Strategy Panel included many distinguished and experienced astronomers and engineers, as well as retired astronaut Bruce McCandless to advise on the on-orbit feasibility of various proposals. Co-Chair Robert Brown had served as Space Telescope Project Scientist for several years before launch. The Panel considered a wide variety of schemes, including: mechanically or thermally slightly deforming the main mirror; overcoating the main mirror to alter its shape; installing full-aperture glass or gas-filled corrective optics at the front of the telescope; replacing the secondary mirror.

These and other proposals were all found wanting, or downright dangerous (to the Telescope or to the astronauts) to various degrees. The Panel also considered how to incorporate changes into the planned replacement instruments scheduled to be installed every few years starting with the 1993 servicing mission.²²

Bits and pieces of the solution lay near to hand, but would require innovative modifications. As a precaution, work had begun at Jet Propulsion Laboratory in the early 1980s on a replacement for the main imaging camera, the Wide Field/Planetary Camera. Its internal optics would be modified by adding small mirrors figured to reverse much-of the spherical aberration introduced by the main mirror.²³ Early in the HST Strategy Panel discussions, optical expert Murk Bottema, of Ball Aerospace, suggested using mirrors similar to those in the replacement Wide Field-Planetary Camera II to adjust the incoming light for the other, axial scientific instruments. The problem was how to deliver such mirrors to the Telescope and insert them precisely into the light bundle behind the main mirror. The solution to this problem, due to electrical engineer James Crocker of the Space Telescope Science Institute is so remarkable as to seem apocryphal. One evening, during the Strategy Panel's meeting at the Space Telescope European Coordinating Facility in Garching, German, Crocker
stepped into the shower in his hotel room. The European-style fixtures included a shower-head on an arrangement of adjustable rods. While manipulating the shower, Crocker realized that similar articulated arms bearing Bottema 's mirrors could be-extended into the light bundle from within a replacement axial instrument by remote control: "I could see Murk Bottema's mirrors on the shower head."²⁴

In the early 1980s, work had begun on a device called STAR, Space Telescope Axial Replacement, an empty stand-in for one of the four axial scientific instruments, just in case one might have to be removed with no new instrument available for insertion. STAR was designed to be mechanically, thermally, and in other ways benign, so that the complex spacecraft would not "notice" the missing instrument. With Bottema's mirrors and Crocker's mechanical arrangement packed inside, STAR would become COSTAR, Corrective Optics Space Telescope Axial Replacement. Once installed as if it were a replacement axial instrument, COSTAR would deploy the tiny mirrors, each figured to intercept and then reverse the spherical aberration for that portion of incoming light directed to various apertures of the remaining three axial instruments. The three remaining axial instruments had a total of five precisely-placed entrance apertures, and the optical design required two mirrors for each. Thus, ten mirrors had to be inserted into the light path in such a way as to intercept the light, correct it, and direct it into the apertures, all

the while precisely shadowing the apertures from the flood of the remaining aberrated light. COSTAR was an exceptionally complex and delicate system of 5,300 parts including mirrors, mechanical components, and electronics all controllable from the ground. Some of the coin-sized mirrors were an optician's nightmare to figure, their shapes being "anamorphic fourth-order aspheres on toroidal blanks," painstakingly hand-made by Tinsley Optics in California. COSTAR would "fix" the European Faint Object Camera, the Goddard High Resolution Spectrograph, and the Faint Object Spectrograph, and the replacement Wide Field-Planetary Camera would restore the primary imaging camera abilities. Astrometry and pointing control would have to live with the diminished performance of the Fine Guidance Sensors, and the High Speed Photometer, the least problematic but the least-used instrument, would have to give up its spot for COSTAR. Having settled on the COSTAR approach, Ball Aerospace began design and proposal work in November, and was awarded a contract in January 1991. Ball had built the Goddard High Resolution Spectrograph for the telescope, and was designing two second generation scientific instruments for future servicing missions.²⁵

The first servicing mission now included replacing two scientific instruments and it would have to remedy a host of other emerging problems, some of which were becoming very serious. At the end of 1990, engineers operating the telescope had devised software changes to try to counter the oscillations of the solar arrays, which would respond sharply to passage between shade and the Sun that occurred twice each 90-minute orbit. The thermally-induced oscillations slowly damped out, but took longer than half an orbit to do so, when another passage would start them all over again. There was limited available memory in the onboard computers, and the new control laws to handle the solar array oscillations consumed a good portion of it. Solar array replacement looked as though it would have to join the other maintenance tasks, although the final decision to replace the European-provided solar arrays was not made until much later. After a spirited debate, they even edged out the optical fixes to assume highest priority.²⁶

The Telescope continued to be temperamental, and NASA finally decided in late 1990 that the engineering commissioning phase, called Orbital Verification, was complete —or that they had gotten the spacecraft operating about as well as they could and regular science operations should wait no longer. The Marshall Space Flight Center team who had been in residence since deployment returned to Huntsville, leaving Goddard Space Flight Center to operate the spacecraft and the Space Telescope Science Institute to continue its work coordinating observing proposals, scheduling objects to be observed, calibrating the instruments, and archiving the data. Between the degraded focus, solar array oscillations, the fine guidance sensors acting temperamentally, and a variety of other sporadic problems, all of the observing plans were continually in flux. Various observing programs, planned with the full capability in mind, had to be reassessed and sometimes deferred. The operational problems had further made scheduling difficult, both by degrading observations and also taking up scheduled observing time to resolve.²⁷

During 1991, the teams settled into doing the observing programs not preempted by the hardware problems, while developing COSTAR and the other major hardware for the servicing mission. Ball Aerospace began assembling hardware for COSTAR in July, receiving the first flight optics at the end of the year. As the servicing mission planning group considered how to include the new tasks (they had been meeting formally since August 1988), they found themselves under pressure to include still more. In May, a memory unit in the spacecraft's main computer failed, sending the Telescope into the deepest safemode available, nearly the equivalent of a coma. In June, a second of six gyroscopes failed, leaving only one spare since three were required for the spacecraft to determine its position and attitude accurately enough to conduct observations. In July, the Goddard High Resolution Spectrograph developed a problem in its power supply, eliminating half of its capability. By the end of the year, it was clear that the first servicing mission was going to be much more ambitious than anyone had expected. It would be more of an overhaul than a repair. Cost estimates are difficult, both for the servicing mission and for the Telescope as a whole, but the repair cost a significantly large fraction of the Telescope's initial cost. Had it been an automobile, it might well have been declared a total loss.²⁸

The close cooperation required of so many parts of NASA, combined with the high stakes and growing anxiety evident in Washington, encouraged a number of institutional clashes that had to be overcome. To the managers and astronauts at the Johnson Space Center, the servicing mission at first appeared as one among others, and it never occurred to them to treat it any differently. Confident and proud of their system for planning and executing Space Shuttle missions, with a host of other missions at various stages of execution, and already interacting with various institutional "customers" for those missions, Johnson was slow to change its approach. As Headquarters scrutiny increased, the complexity of the mission grew, and the need for more subtle interaction with other elements of the program became more apparent, Johnson managers began to feel the pressure .²⁹

The result was that planning for the mission at Johnson accelerated much earlier than was their typical procedure. For "customer driven" missions, a multidisciplinary and multiinstitutional Payload Operations Working Group would spend one or more years developing the outlines of a mission. Actual flight crews and ground controllers were seldom selected earlier than a year before launch. In the Spring of 1992, some twenty months before launch, Story Musgrave was the first crew member named. Musgrave, an Apollo-era veteran, polymath, and extremely accomplished in extravehicular activity, would be the Payload Commander, foreman for the repair crew and ultimately responsible for the Telescope. Musgrave had started working on making satellites maintainable and serviceable in orbit in 1976, and had been the astronaut most involved in developing the space suits and other equipment. Eleven astronauts began sixteen sessions in Marshall's Neutral

Buoyancy Simulator, working with mockups in simulated underwater microgravity to obtain rough approximations of the times required to do various maintenance tasks. All the while, new black box failures on the spacecraft put pressure on the roster of chores to be accomplished.³⁰

At about the same time, J. Milton Heflin was chosen lead flight director. With almost thirty years experience at Johnson, Heflin had been involved in ocean recovery of the Apollo astronauts, and had been in mission control since the very first shuttle landing tests in 1977. With ten years and twenty shuttle missions under his belt as a flight director, he was one of the most experienced and seasoned. As Heflin got organized, he was also supporting several other missions as well. He, like most at JSC, recognized that the Hubble Servicing Mission was challenging and needed more attention than some other missions. He would soon realize, however, just how much more attention and resources it would require. Initially, the core of his mission control team included payloads officer Jeff Hanley, robot arm engineer Sal Ferrera, and extravehicular activity experts Jim Thornton and Susan B. Rainwater.³¹

Meanwhile, events in Washington transpired that would have decisive and far-reaching effects on the agency and on the Hubble servicing mission. In early 1992, after escalating discord with the White House, Richard Truly was effectively fired as NASA Administrator. Truly, a career astronaut, was admired and respected throughout the agency but perhaps nowhere more than at Johnson. By March, a new, non-NASA face appeared and was confirmed on April 1. Not perhaps since James Webb had there been a NASA Administrator who was more of an outsider, more inscrutable, or more difficult to adjust to, than Daniel Goldin. Having worked in mostly classified space programs at TRW, Goldin was expected to bring a no-nonsense industrial agility and accountability to the agency. The Bush-Quayle administration had decided that Truly was too closely identified with the old guard, and were determined to bring a fresh approach to NASA. While Goldin championed what he considered to be NASA's strengths, and went out of his way to praise the achievements of NASA, he was determined to bring change to the agency. To some old hands Goldin seemed impulsive, abrasive, and fearfully insensitive to the agency's core traditions and values. To Johnson, anyone who replaced the beloved Richard Truly, under duress no less, boded ill. Stories of strife in "Code A," the office of the Administrator, were rife.³²

The strong signals of change grew more potent when, in May, Marine Major General Jeremiah W. "Jed" Pearson III replaced astronaut William B. Lenoir as Associate Administrator for the Office of Space Flight ("Code M"). Pearson had been deputy commander of Marine forces in Operation Desert Storm before serving briefly at Marine Headquarters, and he now commanded the very heart and soul of NASA —the Space Shuttle program. Goldin continued to replace many Headquarters and Center officials, while the old hands at the agency went through the anxiety of organizational change.³³

In mid-May Goldin traveled to Johnson to observe a Space Shuttle mission first-hand. The STS-49 Endeavour crew's mission was to capture and repair the Intelsat VI communications satellite and evaluate Space Station construction techniques. Intelsat VI's apogee kick motor had failed two years earlier, leaving it stranded in a low orbit. The crew was to approach the spinning satellite, insert a special tool into the central motor chamber about which it turned, and fire a grappling device that would seize the chamber, allowing them to slow it down. Once stabilized, they would repair it and then send it on its way to geostationary orbit.

Problems developed almost immediately. The four and one half ton satellite bounced away each time Pierre Thuot, in foot restraints at the end of the robot arm, tried to thrust the capture bar into position. Worse, it began to wobble, and fearing that further attempts might be too dangerous they decided to quit for the day and revisit their options. The seven million dollars capture bar had worked during numerous ground simulations. Something was wrong. Overnight, the crew worked with ground planners and astronauts in simulators to devise a bold contingency plan. The next day, after the capture bar again failed to work, for the first time in history, three astronauts were outside their spacecraft at once. Thuot, Richard Hieb, and Thomas Akers, "the gang of three," stationed themselves in a circle in the payload bay, their feet anchored in restraints, while pilot Dan Brandenstein eased the shuttle toward the spinning satellite. They used their gloved hands to reach out, grab, brake Intelsat VI to a stop, and then lower it into its repair fixture.³⁴

While the crew repaired the satellite and sent it on its way to geosynchronous orbit, they were hailed as "space wizards." Goldin praised the bold move as a "return to [the] can-do NASA of old," and editorials and commentaries gushed about the drama of the satellite rescue and its proof of the usefulness of astronauts. However, outside the limelight, it "set NASA managers scrambling to rethink their training methods and assumptions about handling large masses in orbit." The flight and ground crews had saved the mission, but they had taken what some considered to be undue, perhaps even foolhardy risks. Caught between deeply engrained and conflicting values, Johnson had been torn between conservatism and safety, on the one hand, and the driving desire to complete the mission, on the other. They had chosen deliberately on the ground and in space to be bold and complete the task. Tom Akers and Kathy Thornton ran into further problems assembling prototype Space Station elements in the payload bay later in the mission. This, too, did not go as it had in simulations and training. It shook Johnson's planning, training, and simulation groups to the core.³⁵

Worse, while Goldin praised publicly his charges for the daring satellite rescue, he is remembered by others to have commented on "those cowboys!" Barely six weeks into his tenure as NASA Administrator, this was not getting off on the right foot. While Johnson quietly looked inward to reassess their training methods, Goldin appointed a high-level group task force on satellite rescue and repair to study how NASA should handle future such situations. Implications for the Space Station and the Hubble servicing mission were not lost on anyone. Center Directors and Headquarters Associate Administrators decided they, too, needed to track the servicing mission more closely. By the end of July, four new review teams were looking at the Hubble servicing mission, and many more would follow.³⁶

At the end of August, sixteen months before launch, three astronauts were named to join Story Musgrave as the extravehicular activity contingent for the servicing mission, making it the most experienced and seasoned crew ever: Tom Akers, Kathy Thornton, and Jeff Hoffman. A lot of work was scheduled in Marshall's Neutral Buoyancy Simulator and Johnson's Weightless Environmental Training Facility to refine further the timelines for the servicing mission, and "EVAs of opportunity" were inserted into upcoming shuttle missions wherever possible to gain more experience. In January 1993, on STS-54, Mario Runco and Greg Harbaugh of the Endeavour crew improvised the first of these by carrying one another around the payload bay in a "mass handling" exercise and tested some new tools. Several other crews practiced aspects of the repair mission in orbit, spacewalks being added at every possible opportunity.

Meanwhile, the Telescope continued to be a problem child. In September 1992, the Faint Object Camera's power supply developed problems eliminating half of that instrument's capabilities. A third gyroscope failed, leaving the Telescope running on its minimum complement of three. If another gyroscope were to fail, the spacecraft would be safe but unable to collect any scientific data. A second flight computer memory unit failed. Most of these failures added relatively easy individual tasks to the servicing mission, but the timelines were already overbooked and confidence in them somewhat shaken by the new conservatism. They also caused a scramble among spacecraft engineers to determine which problems could be repaired at all, and of those which had highest priority.³⁷

Goddard and the Science Institute, resigned to operating what had become a positively cantankerous spacecraft, became quite adept at dealing with sudden hiccups and replanning observations again and again. The Goddard controllers and scientists had been operating astronomical spacecraft for thirty years. In particular, they could draw on their experience and institutional memory of the remarkable International Ultraviolet Explorer (IUE), which had survived well beyond its design life. They had even learned to operate IUE on a single gyroscope. It was befitting that in the IUE control room sat a stuffed toy "Energizer Bunny." Beginning at the end of the summer of 1991, after the

shock and denial had run its course, the astronomers and operators had resolved to make the most of the capabilities they had. The first science results were published in Astrophysical Journal Letters at the end of 1991. By the end of 1992, astronomers had accumulated a respectable suite of results, some of which surprised even themselves. The Telescope was performing at a level somewhat better than Earth-based telescopes, but far below its expected capabilities.³⁸

In December 1992, a year before launch and at about the time it would have been usual to first assign a crew, Richard Covey (Commander), Ken Bowersox (Pilot), and Claude Nicollier (a European Space Agency astronaut and expert on the robot arm) were named to complete the STS-61 flight crew. With these three, the servicing mission continued to enjoy the most experienced and seasoned astronauts available, now with sixteen previous shuttle flights among them. Also in December, a new face at Johnson was named to a position that had not been used since Apollo: Randy Brinkley was to occupy the newly created position of Mission Director.³⁹

Brinkley and his position were controversial from the first at Johnson, even though the post had been recommended by former Astronaut Thomas Stafford's review of the servicing mission. "Badged Headquarters," Brinkley reported to Washington but was located at and carried on his work at Johnson. The term carried special significance at Johnson, originally created as the Manned Spacecraft Center in the early 1960s. The Space Task Group, a small band of space enthusiasts among aeronauticists at the Virginia Langley Research Center, had virtually invented human space flight at NASA. As Apollo grew into a behemoth, Headquarters decided to create a new Center on some politically favorable bayside land south of Houston. The space pioneers from Langley had little choice but to leave the Hampton, Virginia, area that many of them loved, for to resist moving to Houston meant to be out of the action. They built the space center from the ground up, developing all the camaraderie and social structures of pioneers, and absorbing much of the rugged, self-reliant culture of Texas. They developed and flew Mercury, Gemini, Apollo, Skylab, and the Space Shuttle. They invented astronaut training, mission control, and spacewalking. At JSC, as they referred to their institution, they operated a veritable space university, with several associated colleges where they trained recruits in how to operate in space and support space flight from the ground. Through Apollo and into the shuttle era there had been a constant tension between Headquarters and JSC, a continually and dynamically negotiated balance of independence and subordination. From Webb onward, Headquarters administrators had placed personnel at JSC who reported not to the Center but to Headquarters.⁴⁰

Brinkley had been at Johnson for several months as a special assistant to Headquarters' space flight chief Jed Pearson before being appointed to the servicing mission. A Marine aviator for twenty-five years, he had served with Pearson during Desert Storm and then later

spent a brief time with McDonnell Douglas. Before taking on the servicing mission, Brinkley had studied the shuttle program and tried to get oriented, absorbing some aspects of the JSC culture in the process. He had been impressed by JSC's technical competence, but no less by its independence and his own difficulty in "break[ing] into that subculture." Brinkley researched the Mission Director concept and tried to figure out how to implement it, seeking the advice of many old hands. His charter was broad, to ensure the success of the servicing mission, and largely undefined: "I knew I had the responsibility," he recalled, "somehow I had to grow my own authority.'' This was a broad mission order commonly used in the Marine Corps, similar to others Brinkley had received there, and in keeping with Pearson's style. Brinkley assembled a small staff, representative of the various kinds of inside expertise he would need.⁴¹

He had also, early on, been impressed by how important this mission was to many beyond JSC. From meetings with Pearson, Goldin, and congressional representatives (especially the plainspeaking Barbara Mikulski), the message he received was: "this was a make or break mission for NASA ... Brinkley, don't screw this up, the future of NASA lies in the balance." To gain insight into the progress of the many phases of the mission, he turned to using the independent review groups that had already been set up by others, and created some of his own. The many reviews, some of them uncoordinated and motivated by growing upper-level anxiety over the mission, put a great deal of pressure on the working troops. If Brinkley added to this pressure somewhat by commissioning still more reviews and assessments, he also drew gratitude from the workers by managing and coordinating the reviews and serving as somewhat of a lightning rod. "Answering the mail from Headquarters," while initially not seen as much of a genuine contribution to the mission, eventually was seen as a valuable activity that allowed the people working on various aspects of the mission to function with minimal diversion. The various review recommendations and attention also gave the Hubble servicing mission team a good deal of clout in getting the resources they needed on a priority basis, clout which Brinkley was not afraid to exercise from time to time. This also caused some dismay among the crews of other missions competing for flight controllers, simulation and training time, and attention from the various technical service group.⁴²

In January 1993, Administrator Goldin appointed the most formal and highest level review committee of all. Tapping Joseph F. Shea, an Apollo manager, and several senior aerospace executives and experts, he formed "The Task Force on the Hubble Space Telescope Servicing Mission." After several formal meetings and numerous briefings at all the sites where the mission was being planned, the panel in May called the repair mission "achievable" but recommended continued close management attention, strongly endorsing the Mission Director concept. They were concerned about the "escalating nature of the mission," and the "instability" of the plans and on-orbit schedule, based on a "worrisome" trend of equipment failures on the spacecraft.⁴³

The servicing mission itself lay at the confluence of several streams of work involving NASA Headquarters, Jet Propulsion Laboratory, four NASA Centers, the Space Telescope Science Institute, and a half-dozen contractors. During the summer of 1993, these streams would truly converge, and any residual turbulence had to be dealt with. Choreographing the mission was easily as challenging as any other aspect. Before and after capture, myriad instruments and systems on the spacecraft would have to be turned off systematically, the aperture door closed, antennas retracted, and solar arrays rolled up. The shuttle crew would have to grapple the spacecraft with the robot arm, and gingerly berth it into the receiving fixture of the payload bay where electrical connections could supply "life support" during the repairs. The Goddard controllers would have to watch the health of the spacecraft, putting some systems into standby and turning others off as the repair crews began to remove connectors and components, and then turning the systems back on afterward to ensure the "aliveness" of the new parts. As various access doors on the Telescope were opened, the shuttle's attitude had to keep stray sunlight from entering the Telescope while still maintaining lock on the tracking and data relay satellites.

The "EVAs of opportunity" inserted into the shuttle schedule during 1992-1993 revealed crucial information that probably was decisive in the success of the servicing mission. For various reasons the areas of the Telescope being worked on had to be kept out of sunlight. If subjected to sunlight, for example, the black insulation that suppressed interior reflections might "outgas," exuding contaminants that might later deposit onto the optical surfaces. Solar heat might cause insulation to de-bond or expand. Over most of the extravehicular activity experience since Gemini, the problem had been keeping astronauts cool, since they were almost always in sunlight or reflected earthshine, and so the suits and gloves were very efficient at cooling. Mission designers, not taking crew temperatures into consideration, designed a trajectory optimum for the Telescope's and orbiter's needs. This oversight revealed itself dramatically and at a most inopportune time.

In May 1993, Story Musgrave began a series of human thermal vacuum tests, similar to those done by many astronauts since Apollo. After four hours in the airlock, breathing oxygen in his suit to rid his blood of nitrogen, he finally entered the vacuum chamber for an experience he described as, "the world's worst hell. That's the toughest day that you are ever going to have as an astronaut." Inside the black chamber, pumped down to the deepest vacuum possible, and dragging his counterbalanced 480-pound inflated suit "like a plough horse," he started several hours of tool fit checks. His job was to go through all possible combinations of tools —sockets, ratchets, extensions— and fasteners to see whether they would fit and behave at a hundred and seventy degrees below zero. Even Musgrave himself, a medical doctor, did not realize what was happening as he squeezed the tools harder to get them to snap together and apart.⁴⁴  It was, he says, the "insidiousness of going from pain to injury." Working in an inflated glove, and feeling numb anyway, he worked on for hours, occasionally pulling his fingers out of the gloves and up, the way a person in a parka might pull his hands into the sleeves, and occasionally trying to warm them in another part of the chamber that simulated sunlight temperatures. After finishing the fit checks, hours of decompressing, and emerging from the airlock, the metal on his suit was still too cold for anyone to approach. As he struggled to get his gloves off, and they dropped away, the damage was evident: severe frostbite, tissue death, in eight fingers.⁴⁵

"It was an essential thing which had to happen ... it redefined the entire EVA world," he said. John Young later noted that astronauts had "complained of cold EVA temperatures for years," and Musgrave recalled cold hands on his very first shuttle spacewalk on STS-6 in April 1983. But previous shuttle astronauts had only been in the shade for portions of their spacewalks, and the Hubble mission would require them to work for extended periods away from sun- or earthshine. With only seven months to go before the mission, Musgrave was flown to be treated by frostbite experts at the University of Alaska, mission planners turned to reconsider extravehicular activity equipment and procedures, and high-level managers fretted. Musgrave recovered fully in time for the mission, and was even in the water tank within days of the injury. But the incident produced "a whole new attitude" —quite literally.⁴⁶

John McCune, an experienced "pointer" who specialized in designing orbiter attitudes for missions, suggested flying an upcoming mission in the orbiter attitude planned for the servicing mission to learn about the thermal environment. Astronauts on STS-57 in June found that the payload bay was indeed far too cold for extended work. The mission design engineers went back to their computers to determine a series of orbiter attitudes that would keep the payload bay temperatures manageable, while keeping the direct sunlight out of the Telescope bays, accommodating the orbiter's needs, yet still conserving already tight maneuvering propellant. Others developed warmer overgloves and other techniques, and revisited the idea of storing the tools inside the warmer (but already full) orbiter cabin rather than in the payload bay toolboxes. Had the cold hands problem emerged on-orbit, in the middle of the repairs, it could have been very serious, even disastrous. As it was, planners had to pull out a thread lightly interwoven throughout the warp and woof of a complex and interlocking mission. It took considerable cooperation between Goddard and Johnson to work out a new flight plan that would accommodate the new thermal requirements.⁴⁷

Over the summer, as headquarters was preoccupied with President Clinton's decision concerning the Space Station, Goddard engineers and astronomers tested and evaluated the instruments and other components in preparation for shipping them, in August, to the Kennedy Space Center. As the instruments and other components flowed through Goddard, they were subjected to multiple, independent, and rigorous testing to cross-check the results and avoid the kind of error that had befallen the mirror. At Goddard and the Science Institute, planners refined the complex and interlocking sequences of instructions that would have to be sent to the spacecraft to prepare each component for replacement and test its successor. Houston was occupied with preparing for joint integrated simulations of the mission. A total of seven would be held, August through November, and represented the most complex total exercise of any mission plan in the history of the shuttle program. With various astronauts in the Neutral Buoyancy Simulator at Marshall, ground controllers at mission control in Houston, and various supporting engineers and scientists at Goddard, the Science Institute, and contractor sites around the world, they would rehearse various parts of the mission as well as numerous "failures" concocted by the simulation super-visors.

One recommendation from several of the review teams concerned increasing the fidelity of the water tank simulations. The robot arm at Marshall's Neutral Buoyancy Simulator was crude, and planning began on a higher-fidelity version. Johnson's tank was too shallow to accomodate an arm. Also, to allow longer and more realistic simulations plans were made to install a nitrogen-oxygen ("nitrox") breathing mixture, which would allow simulations of entire six-hour spacewalks, rather than having to break them up into smaller portions.⁴⁸ Here Brinkley ran into center parochialism, as the upgrades to the

Marshall tank were seen by some as competing with proposals to upgrade the much smaller and more-limited Johnson Weightless Environmental Training Facility to a full Neutral Buoyancy Laboratory, conveniently available to Houston flight crews. After paperwork "got lost" along the way, the changes were finally implemented barely in time for the October simulations.⁴⁹

The flight crew took advantage of the long water tank simulations to hone their bodies and their spirits as well. They would have only one chance on-orbit, and so prepared for that as if it were the Olympics. "We did things at Marshall like we were going to do upstairs ... at seven a.m. we would be in there and we would brief what we were going to do, and then we would position our tools ... then we'd get in the suit, and then we would go work, we'd do six or seven hours there, we would get out, we would debrief how everything went, we'd capture our lessons [learned]. Then we went off to the gym.... we would put in an hour and a half or two.... yeah, hard work ... then we would come back from there ... and start talking about and reviewing the next day's activities ... and then it's ten p.m. and at 7 a.m. we're back at Marshall and the whole day repeats, and we did this day after day after day without missing a day for three weeks. Now, you wonder why it worked?"⁵⁰

Brinkley "wasn't worried about the flight team," but was worried about upper level management's readiness for real-time decision making. Brinkley found that previous mission simulations had effectively concentrated on exercising decision-making among the crew, flight controllers, and the many "back room" technical groups that advise them. Generally, decisions that were too serious to be made within mission control had relied on a Mission Operations Director console position, the occupant of which would be the Flight Director's interface to the upper management. During the joint integrated simulations, a real-time mission management team comprising administrators from the highest levels of the agency participated and were faced with contrived situations that pitted the "safety of the orbiter versus the survival of the Hubble." Associate Administrators for Space Science (Wes Huntress), Space Flight (Jed Pearson), and Center Directors (John Klineberg, Goddard) worked with the team to practice how they would confront such situations during the flight. ⁵¹

In addition to the water tank simulations (738 hours), several other areas of ground training enjoyed renewed emphasis. Since the viscosity of the water tended to make handling massive objects less realistic, an air-bearing floor simulator was used to gain more realistic experience in the dynamics. Twenty hours of manned thermal vacuum tests were done. Computer graphics virtual reality simulators were developed, and were used to research positions for the crew and robot arm to use on the repairs. In the Manipulator Development Facility, Claude Nicollier and Ken Bowersox worked with a realistic robot arm, hoisting full-scale helium-filled balloons in the shape of Hubble and of a spacesuited astronaut (nicknamed "Gumby") .⁵²

At the High Fidelity Mechanical Simulator at Goddard, the crew practiced replacing components using highly realistic models and even flight hardware. Goddard had built the simulator, which reproduced the aft portions of the telescope, where the instruments fit, and the equipment bays that held the various electronics components, so that replacement and new components could be checked accurately on the ground. In addition to being responsible for the Hubble Space Telescope, Goddard had worked on satellite servicing concepts since the early 1970s, developing tools and techniques used on other earth-orbiting spacecraft. Rothenberg, Frank Ceppolina, and others had been sensitized to the minute detail required in testing and training simulators by many experiences, among which was the repair of the Solar Maximum mission spacecraft on STS-41C in 1984. There, astronauts had been temporarily halted by a small piece of insulation that had sagged out of place. Goddard also maintained a separate electrical simulator for testing the numerous data and control connections. Most of the components for the servicing mission would travel to the launch site via Goddard's clean room, where they would be tested and where the crew would have a chance to work with real hardware. "Goddard was just magnificent in knowing how to get a crew ready ... they were much more than just the customer ... they were EVA trainers, they swam with us all the time ...,"Musgrave recalled.⁵³

The astronauts also went to various other locations around the country and in Europe to see and handle the actual flight hardware. They even went to the Smithsonian's National Air and Space Museum, riding a cherry picker crane after-hours to inspect the old Structural Dynamic Test Vehicle. The only full-scale version of the telescope to have been built in a program that tried to save money by building only one full-fledged flight spacecraft and no prototypes, the vehicle had served as a wireform on which the 25 miles of flight wiring had been laid out. Not realistic in many respects, the carefully restored and preserved artifact nonetheless retained the many simulated connectors and black boxes where the cabling now in space had once been meticulously strung. Unlike the more aesthetic models built later to represent the flight spacecraft, this vehicle did not have a covering of multilayer insulation to get in the way and obscure details.⁵⁴

The Hubble Space Telescope program had eschewed building prototypes, and so there was no single place the crew and engineers could go to see a truly accurate and completely realistic version of the spacecraft that orbited more than three hundred miles up. There had arisen by necessity a panoply of mockups, models, and simulators, supplemented by thousands of photographs and video tapes taken of the flight spacecraft before it left the ground. Each was accurate in some important respects, yet each was dangerously misleading in most other respects. To get the whole picture, it all had to be synthesized in the mind. No one experience was adequate, and all had defects. The astronaut would have to pick out mentally the appropriate parts of each training exercise and suppress the rest. Musgrave described what it was like:
 

"In your imagination you extract from the manned thermal vacuum testing the apprapriate parts, so when you're in the water doing those marvelous things with those gloves, you have to say, 'this is not the way its gonna be.' You have to go back to the manned thermal vac, and think what my real gloves were like there, at flight temperatures ... only in the head does the entire mission exist.... You cannot go dumb in the

water tank and say, 'this is the way it's gonna be,'. . . you mentally pull from the water experience, you pull from the air bearing floor, . . . you pull from the vacuum chambers in your real suit, you pull from JPL, from Ball, from the clean room at Goddard, you pull from the previous zero-g experience, you pull from your previous EVA experience. Now you sit down and in your imagination you go through five days of work ... just like a ballet, where every finger every toe, I knew where every tool was throughout five days. So I built, just like the score of a symphony, I had the whole five days in my head. ... every single motion, every translation of the body, every worksite.... The reason Hubble worked was, number one, Hubble was incredibly friendly to being serviced by an EVA crewperson, and we were able to attack all the details in all of those environments and build a mission. "⁵⁵

Such an ambitious and expensive servicing mission might never have been seriously considered had not much more than the Hubble Space Telescope been riding on it. The Space Station had long been NASA's choice for the next logical step beyond the shuttle. As the Hubble Space Telescope's problems and their potential solutions emerged, and with the new awareness of the difficulty of spacewalking under certain circumstances, planning work on space station assembly revealed that it would require an unprecedented amount of extravehicular activity. With only the restrictive volume and weight capacity of the shuttle, assembling the Space Station seemed to call for spacewalks in a nearly implausible number, duration, and complexity and variety of tasks. Review committees questioned seriously whether such a scenario was reasonable. A successful Hubble servicing mission would provide dramatic proof that the Space Station assembly could be done. While the reviews of the servicing plans were going on in the summer of 1993, President Clinton was making an important and extremely contested decision concerning the future form of the Space Station. The Space Station "Freedom" of the Bush administration was replaced by Space Station "Alpha," a less-ambitious and less-costly version. That a space station at all emerged from the presidential decision process was cause for rejoicing, but it was now tied more firmly to the success of the Hubble servicing mission.⁵⁶

Later in the summer, several other unfortunate events served to tarnish further NASA's reputation and raise the stakes. A weather resources satellite, NOAA-13 failed shortly after launch in early August. The Galileo spacecraft, enroute to Jupiter after a perils-of-Pauline life of its own, had been unable to deploy its high-gain antenna, and attempts to free it seemed doomed, threatening the viability of its mission. The next shuttle mission, STS-51, was delayed from its late July launch date because of concerns about a particularly active annual Perseid meteor shower. When Discovery finally flew in mid-September, the crew tested Hubble tools on-orbit during a full seven-hour spacewalk September 16, 1993.⁵⁷

The most serious, and mysterious, mishap of those remarkable few weeks, however, was the Mars Observer spacecraft which, on the verge of entering Mars orbit suddenly went silent. Again the cries of critics rose to a din and NASA suffered the barbs of pundits. On September 6, 1993, NASA was honored with its second Dave Letter-man Top Ten list.⁵⁸

In addition, 1993 was a frustrating year for the shuttle program. By mid-August, four launch countdowns had proceeded to within twenty seconds of ignition and had to be scrubbed for various reasons. Two had actually ignited the main engines and then shutdown just a split second before the solid rocket boosters were to ignite-the point of no return in a launch. In October, the Landsat-6 satellite's kick motor failed after launch, leaving the Earth resources satellite stranded and useless. Some of these untoward events were not even within NASA's responsibility or control, and others were only remotely related to the Hubble mission, if at all, yet the servicing mission was acquiring a significance far beyond just repairing the Telescope. While upper-level managers worried about the larger picture, engineers and astronauts tried to maintain their focus on the details that would make or break the mission. As Musgrave recalled: "Did the pressure come to me? No. It's out there, and I know it's out there. But I am gonna go do my job, apart from the pressure and apart from the outside world. I'm the ballerina, and I know the opera company may be resting on my shoulders. OR, it's tough, but I'm gonna go do what I've got to do. It doesn't matter that the opera company is resting on my shoulders, or isn't. It's my art, and I'm going to perfect it to the best of my ability, and it's me and it's my art.... Yes, there's pressure, but it's not external pressure, it's internal."⁵⁹

By December, as launch approached, nearly every press story on the mission declared it to be do-or-die. It was a mission of superlatives: more spacewalks, by more astronauts, for more total time than ever-before attempted; one-chance to rendezvous; no second chance to rendezvous once the telescope was released at the end of the repairs. The complexity and high stakes of the mission were reflected in the unprecedented series of reviews, so numerous and extensive that some feared the mission was being reviewed to death, and that the preparation required for these reviews was taking resources away from the mission itself. By the November 17, 1993, Flight Readiness Review, 195 formal recommendations had been made by twelve review teams, of which only twenty-seven remained to be closed.⁶⁰

When the priorities for various repairs were put alongside the order in which certain tasks had to be done, it was not possible to do the various tasks one after the other using 6-hour spacewalks. In priority order, the tasks for full mission success were: solar arrays; two gyroscopes; Wide Field-Planetary Camera; COSTAR; magnetometer; and Solar Array Drive Electronics. Minimum success would be three reliable gyroscopes and an operational WF/PC II or COSTAR. Secondary objectives would include a fix for the Goddard High Resolution Spectrograph, a 386 coprocessor to replace the failed computer memory, a sccond magnetometer, and gyroscope control electronics. The final timeline scheduled these tasks to get the highest priority items done as soon as feasible, but not in strict priority

order. Each day's work had to leave the Telescope and the payload bay in a condition that, should the worst emergency arise, the crew could release the Telescope quickly, close the payload bay doors, and head home.⁶¹

On December 2 at 04:27 EST, following a one-day delay for weather, Discovery finally roared away from the Kennedy Space Center in an uneventful launch. After nearly two days of catching up with Hubble, and periodically firing thrusters to slow down the catchup rate, Pilot Ken Bowersox reported the most-often used words to describe the Telescope: "Houston, it's really big!"⁶² Claude Nicollier grappled the Telescope, and gingerly berthed it into the Flight Support Station where the shuttle's power would substitute for that from the solar arrays. The tiltable turntable would be used to orient the Telescope work areas so that Nicollier and Bowersox, operating the robot arm from the aft flight deck, would have the appropriate portions of the Telescope facing them.

On flight day four, December 4 at 10:46 p.m. EST, "the odd couple" of Musgrave and Hoffman began the first spacewalk. For five days one or another pair would exit the airlock at around the same time each evening and spend between six and eight hours in the payload bay. This was the real test of all their training, as Musgrave put it: "I like the heat of the kitchen; I live in it, I've lived in it for decades, and I thrive on it. And so, it gets me up-to where I've got to be. Goin' out the door for the first time, I was incredibly interested, concerned about, have we nailed it in terms of our imaginative process, are the simulators right, are the mockups right, did we approach this job right?"⁶³ After setting up the work site, Musgrave and Hoffman replaced the gyroscope packages inside the aft shroud of the Telescope, and gyroscope electronics and fuses in the modular bays around the "waist" of the spacecraft. They ran into some problems closing the aft shroud doors, but other-wise all went according to plan. "Endeavour, Houston; not to get you spun up, but we've got six good gyros on the telescope," Capsule communicator Greg Harbaugh punned to the crew after the aliveness test spun the new gyros up to operating speed.⁶⁴

On flight day five, Akers and Thornton removed the solar arrays. One array had refused to roll up into its cassette, a situation that had been the subject of one of the simulations.

Working during orbital night, when the array would not be generating electricity, Thornton removed the huge and cumbersome wing, holding it while Nicollier moved her into position with the robot arm. At sunrise, as Thornton let go and precisely spread her arms to ten and two o'clock, the array stood rock-steady for a moment, and as thrusters fired Endeavour began to back away from the shiny gold "bird." It rotated slowly, and then as a thruster plume hit the array it bent as if a bird or a pterodactyl flapping its wings. The replacement solar arrays went on without a hitch.⁶⁵

On flight day six, Musgrave and Hoffman removed the Wide Field/Planetary Camera, stationed it temporarily on a fixture, and prepared to install the replacement. Musgrave likened it to moving a baby grand piano, holding it by the keyboard, and trying to insert it into a gigantic dresser like a drawer. At the far end, however, to be inserted right into the heart of the Telescope just behind the primary mirror, was the camera's all-important " pickoff mirror." Musgrave had to remove the protective mirror cover, at which time the optics would be exposed. "That is the most powerful camera in the world, it sees the furthest out there in space and time, if you touch that mirror, however you touch it, that will be on every single image that comes down from the world's most powerful camera." The camera went in without incident, and Musgrave and Hoffman moved to the very top of the Telescope to install new magnetometers.⁶⁶

On flight day seven, mission control played the traditional wake-up music, appropriately the popular song, "I can see clearly, now." Working in the aft shroud, Akers and Thornton removed the High Speed Photometer and then installed COSTAR. Crocker got a message from a friend shortly after COSTAR was installed, "Congratulations ... there's a train leaving Baltimore tomorrow night at 8:30; if this thing doesn't work, be under it." All they could tell from the aliveness test was that the instrument was ready, but the mirror arms would not be deployed until several days after the servicing mission. In one of the equipment bays, they installed the 386 coprocessor to augment the flight computer's failing memory. Initially the new coprocessor seemed not to work, providing one of the very few serious problems in the mission. Planners considered adding another spacewalk to remove the new coprocessor and simply replace the entire computer, but it turned out that the problem was on the ground in processing the telemetry.⁶⁷

Flight day eight began with a boost from Endeavour's engines to get the ensemble to the highest orbit possible, 320 nautical miles. The last scheduled spacewalk was challenging, for it involved mostly equipment that had not been designed to be replaced or repaired. Musgrave and Hoffman installed a special cable to restore the power supply in the Goddard High Resolution Spectrograph, replaced insulation on the magnetometers, and replaced the electronics box for controlling the solar arrays. When the rolled up solar arrays did not extend, the crew had to crank them out manually. After that, the blankets extended from their cassettes normally. ⁶⁸

On flight day nine, December 9 at 11:43 p.m. EST, the Telescope was transferred to its own internal power. At 5:27 the next morning, Nicollier released the spacecraft and the

shuttle slowly backed away. While the controllers at Goddard began the carefully-orchestrated sequence of reactivating the various Telescope systems, the crew enjoyed a day off, and then prepared the shuttle for return. They landed at the Kennedy Space Center on December 13 at 25 minutes past midnight. The most remarkable thing about the mission was that nearly everything had gone according to plan. The few glitches that had occurred were minor, and relatively easily solved. Sky and Telescope magazine, in a relatively brief post-mission article, simply referred its readers to the preview published the preceding month, since things had gone so well. The telemetry signs were encouraging, but it was a tense Christmas season while everyone waited to see whether the new instruments would really work as well as they seemed to be functioning.

"It's fixed beyond our wildest expectations," Program Scientist Ed Weiler beamed at a mid-January press conference. Barely five weeks after the servicing mission, after numerous engineering checks, the Telescope had exercised both the new Wide Field-Planetary Camera and COSTAR with the Faint Object Camera. Jim Crocker said the performance was "as perfect as engineering can achieve and the laws of physics will allow."⁶⁹  The newly repaired Telescope "overshadowed everything else at the American Astronomical Society (AAS) Meeting" in January 1994.⁷⁰ The astronomical end-users of the technology were delighted. Eventually even skeptics had to admit that the Telescope was making good on its most extravagant claims. The dramatic images released only hinted at what was to come. Over the following months, as the rest of the mirror-arms of COSTAR were deployed and the various instruments focused and calibrated, the Telescope really got down to work.

In May astronomers announced the first generally convincing evidence for the existence of black holes, an object at the center of the nearby galaxy M-87 with a mass two-three billion times that of our Sun compressed into the size of our solar system. This led the Washington Post to editorialize that the "trials and tribulations" of the Hubble Space Telescope were, in the end, "worthwhile ." ⁷¹

In June, a team led by former Space Telescope Program Scientist Bob O'Dell announced that they had obtained images of protoplanetary disks around young stars in the Orion Nebula, and found the process of planetary formation going on around 56 of the 110 stars observed. These rotating disks of gas and dust had been inferred from other orbiting observatories' data and in one case even imaged, but not in the detail or in the numbers reported from Hubble.⁷²

In mid-July the fragments of Comet Shoemaker-Levy 9 hit Jupiter, and Hubble was ready. Day after day, for more than a week, images of the giant planet from Hubble showed the comet fragments as they went in, and the dusky spots where they had disturbed the giant planet's atmosphere.⁷³

In October, the astronomers were startled and the public bewildered by the first results from one of the projects that had been of prime importance since conception of the Telescope. Observations seemed to suggest an age for the universe of eight to twelve billion years, dramatically downward from the fifteen to eighteen billion years previously estimated. Paradoxically, this age was younger than some estimates of the ages of certain stars.⁷⁴

By November, confidence in the Telescope's abilities was so high that its failure to find stellar objects was considered a major discovery. For decades astronomers and cosmologists had been increasingly uneasy about the so-called "missing mass," some 90 percent of the matter expected to be in the universe that nevertheless does not show up in surveys. When two separate teams using Hubble Space Telescope observations failed to see anywhere near the expected number of red dwarf stars or other objects in certain fields, it deepened the mystery. The number of conventional places to look for the "missing mass" was dwindling, forcing theorists toward more exotic locales .⁷⁵

In December, three teams using the Telescope revealed that they had obtained images of galaxies from very early in the universe, perhaps only one tenth of the total time elapsed since the big bang. Surprisingly, the primeval galaxies were found to be of a variety of complex shapes rather than uniformly simple, clustered rather than evenly distributed, and to harbor apparently very old stars. At such an early age of the universe, they had expected to see more uniformity and less structure, and were somewhat at a loss to understand how so much evolution could have taken place in such a short time.⁷⁶

These and many other results flowed in a steady stream from the teams using the Hubble Space Telescope. Satisfied customers? Ecstatically so. The telescope carried a lot of baggage. But even if one takes into consideration the various ways in which the performance of the repaired telescope fell short of its original planned performance, it still was enormously useful. It had graduated from making-do to normal operations.

The political success of the mission can be read in who decided to appear on the dais at the press conference announcing the first corrected images from Hubble on January 13. Headquarters Acting Associate Administrator for Public Affairs Jeffrey Vincent, in a classic Freudian Slip, welcomed everyone to the "Goddard Space Flight Senator-ah, Center." He had good reason to be nervous. Instead of scientists and engineers, the first panel to speak included Administrator Dan Goldin, White House Science Adviser John Gibbons, and Senator Barbara Mikulski (D-MD), head of the Senate Appropriations Committee in charge of NASA's budget and in whose district was Goddard, as well as the

Space Telescope Science Institute. Goldin kept breaking a grin throughout his wide-ranging introduction that praised the servicing mission and likened it to great missions of exploration throughout history. Gibbons, referring erroneously to the formerly "astigmatic" Telescope, took the occasion to praise the vision of the Clinton-Gore administration as well. The remarks of all were rhetorical, ceremonial, political, and tinged with minor scientific and technical errors. Senator Mikulski got right to the point: "I chair the subcommittee that financed the manufacture of the most significant contact lens in American history, the fix on the Hubble Space Telescope, and then bankrolled this extraordinary space HMO that went out and gave Hubble Telescope a new contact lens, and I am happy to announce today, that after its launch now in 1990 and some of its earlier disappointments, [raising her voice] the trouble with Hubble is over!" Suddenly, she became a scientific briefer as she proudly held up two images taken by the Faint Object Camera before and after COSTAR and explained in detail how much better the performance was. "This shows what COSTAR can do, and, Mr. Goldin, I'm going to ask you to hold that, because there's more to come," drawing laughter from the assembled reporters as she turned the NASA Administrator into her chartholder. Holding up an earlier Wide Field/Planetary Camera image, she said it "looks like the way you would look at a road map New Year's eve," and then pointed out the much clearer image from the camera's successor. "I believe these pictures are tangible evidence that not only has Hubble been fixed, but NASA is well on its way to fix that culture that created some of these problems ... this was a high stakes repair for Hubble. . . . "⁷⁷

The success can also be read in what happened to the people who led the mission. Of the Collier awardees, two managers went on to significant positions in the Space Station program, the most important effort of the agency since it represented the future. Randy Brinkley became Space Station Program Manager at Johnson, and Milt Heflin began working on integrating extravehicular activity in the Space Station assembly. Brewster Shaw continued to lead Space Shuttle activities at Johnson. Joe Rothenberg eventually became Director of Goddard. The flight crew members went on a long public relations tour that took them to the White House, Congress, Europe, the hottest late-night talk shows, and even an appearance in an episode of ABC's hit comedy series 'Home Improvement.' There they played a bit of tape from one of the spacewalks that included two of the male astronauts in the cargo bay of the shuttle doing the primitive grunt popularized by Tim "The Tool Man" Allen. As Story Musgrave reflected, ". . . it was something so basic and primitive about it, humans and their tools, and the drama of whether it was all going to get done."⁷⁸

Commander Richard Covey retired from NASA and joined Unisys Space Systems to head their simulation and training activity. Shortly after the mission he said, "This would be a great mission to end my astronaut career on, and it would be hard for me as a commander to look to another one that would bring as much reward and joy. . . ."⁷⁹ Characteristically, the rest of the astronauts returned from their publicity tours to move on to the next assignment. Characteristically, when asked about their roles in the mission they would praise the team, and point to somebody else as responsible for the success.

The experience and lessons learned from the mission were carried far and wide by the many people who worked on it, and it became a paradigm for emulation. Brinkley, as he moved to head the Space Station planning, deliberately sought out the people and expertise from the Hubble Servicing Mission to incorporate them into Space Station. Even 18

months later, on June 25, 1995, after the first docking of the shuttle with the Russian space station, Mir, Will Trafton, NASA Headquarters Space Station Director said: "We're estimating now some six hundred plus hours," [of spacewalks; to assemble the Space Station, with the Russians doing an additional 200 hours.] "We look at EVA as a resource ... it's not just another way for a bad event to happen. Hubble has taught us a lot, we're using the Hubble crew and the Hubble experiences to look at the EVA work that's required to assemble space station, and we're pretty happy about where we are." From the Hubble servicing mission came confidence that more than a hundred days of Hubble-type work could be done in assembling the station.⁸⁰

The Hubble Space Telescope, amid the many mundane reasons for its existence, represented at least in part a transcendent and pure purpose: to explore and try to understand the deepest mysteries of the cosmos. It was also in serious trouble, yet not so serious it could not be saved. NASA, an agency that also represents, at least in part, transcendent and lofty goals, was also in trouble, and its fate tied to that of the Telescope. This brought out the best in people, in the samaritans who extended themselves far beyond their job descriptions to come to the aid of the machine and the idea. They worked long hours, pushed themselves, studied, became innovative and clever. They even were able to forget for a time business, Headquarters, Center, and divisional boundaries, and turf, a remarkable achievement in such institutions, Because people believed in the worth of the mission, and because they also feared in their very bones that failure would bring the most dire of consequences, they gave the mission attention and resources. From Congress and the White House all the way down to small divisions and work groups, they put in money, time, and attention. The anxiety of executives looking over their shoulders might have made more work for the executors, but that anxiety also gave them clout to override ordinary bureaucratic barriers. Amid all the other missions, tasks, and priorities swirling about, they put this one on top for a while. Thus the Hubble servicing mission represents the kind of infrequent and special kind of push that people and organizations do from time to time. "Not since Apollo," was a phrase that many people used to describe how they felt working on this mission. Like Apollo, or the Olympics, however, such a special conjunction of will, spirit, and effort probably cannot be sustained in the ordinary course of things. It was, quite literally, an achievement of focus.⁸¹

Like Apollo too, however successful the mission, it could not by itself create a rising tide for NASA. "That's a lot of baggage to carry with you," Covey said in answer to a question at the post-mission press conference, "we had a task to do, and that was to fix the Hubble, . . . and we knew that the best thing we could do was to do that job very well.... and that was all we were thinking about during the course of the mission.... We hope that someone else is able to translate that into NASA doing well."⁸²

Significant historical movements were afoot that were not controllable, indeed, hardly predictable. Aviation Week warned, shortly after the mission: "in political and public relations terms, Mission 61 probably presented more risks than 'up-side' potential. Failure could have been disasterous. But few space policy insiders expect success to give NASA a big, lasting 'bounce'."⁸³  Even while truly spectacular results were flowing in from the Telescope, and the Clinton administration reaffirmed its faith in the Space Station, the

NASA budget was in trouble. As an agency, NASA had been selected by the administration to be a showcase for "reinventing government," and from the White House came directive after directive to reduce its budget. Later in 1994, as a new Republican majority was elected to Congress, even more pressure was brought to bear on NASA. Many of the congressional representatives who had been friends to the agency and who had been impressed by the Hubble servicing mission found themselves in the minority party, having been replaced by other representatives who were either hostile or indifferent. Most ironically, the space science budgets were squeezed ever more tightly, Hubble's included. Scientists feared that the very mission that saved their observatory had lent credibility to the Space Station, which threatened to devour it.

Had the mission not been attempted, the Hubble Space Telescope would have been a constant, orbiting reminder of failure, even while producing very good science. With the successful repair came redemption-but only redemption. The agency's credibility, in this and many other areas, was saved but not boosted. Hubble was no longer an albatross around the neck of NASA, and it appeared the agency could indeed do what it said it would do. But what the public, the politicians, and others wanted NASA to do remained as it has for most of the history of the agency: uncertain, fickle, and contested.