Rand Simberg has a little bit on NASA's X-34 white elephant.

There's more shame in that than he notes, considering all the interesting technology that was to be flight-demonstrated in that program. The same is true to a greater degree for X-33, though that program was much further from completion when it was cancelled.

Looking back over NASA's history, however, it should come as no surprise that these two programs were cancelled in the manner they were, with hardware fully or partially built, most of the money spent, a comparatively small amount of money needed to complete them, etc. It isn't a post-Challenger phenomenon, per se, but an institutional defect which only grew worse after that accident.

During Apollo days (as part of that program or other concurrent projects), the agency developed numerous interesting and potentially useful technologies -- which it shelved due to risk and cost. The "revolutionary" linear aerospike engines built for X-33 were based on a design developed and tested in the early 1970's. An annular aerospike version of the J-2 engine used on Saturn upper stages was developed, but never flown. An entire facility was built outside of Miami for the production of monolithic (no joints) SRBs with a diameter twice that of the design used on Shuttle -- big enough to replace the S-Ib first stage -- but it was mothballed and then scrapped after only a few motors were manufactured and (succesfully) static test-fired. Add to this the cancellation of Apollo 18 and Apollo 19, ostensibly on cost grounds (the cancellation saved a total of $46 million dollars but wasted the billions already spent), but in reality due to post-Apollo 13 fears that the program's luck couldn't hold out indefinitely.

But that is the problem inherent in government programs -- beyond the impact of the loss of valuable equipment and irreplaceable individuals, the political element itself makes any visible failure potentially lethal to the organization itself. The slightest failure gives the agency's enemies ammunition to use against it in stripping it of funding to use elsewhere, so the only alternative is to "study" and "research" and "plan" and "test" and always report "progress" of some sort -- until the day of reckoning comes and the project is terminated. Even termination is preferable to failure, however, as it can be claimed that "lessons have been learned" from the experience, and the controversy over the money spent can be spun into funding for future "paper queen" projects through the magical mantra of "the technology just isn't there yet" -- with emphasis on the "yet" and a hand held out for new funding for a new project that will most assuredly take us there the next time around.