Kim Said:-

High technology projects *always* run into problems, whether run by a
government contract or by some company's R & D dept. It's the nature of the
beast. Expecting one to come in on target is asking for an awful lot that has nothing to do with management. High tech costs money - lots.

I don't see it this way Kim. There are a couple of problems here:

Either, the project managers are REALLY bad at estimating technological risk, or the projects are being launched prematurely.

You would think that experience on all these projects to date would allow the budget people at least to establish that projects typically overran by a factor of 2 or 4 or whatever, and inflate the budget request to match. The fact that they don't is part of the political game.

You pretend you can do the job for $1 billion, and congress reluctantly approves because they judge the result to be worth that investment.

After a few years you reveal you've spent $1.5 billion and are only 75% complete, so you request another $1 billion or so.

Congress reluctantly approves because because they've already pissed away the $1.5 billion , and the result is still worth the additional investment of $1 billion.

After a few years you reveal you've spent $3 billion and you're only 85% complete...

(etc...)

What really ought to be done is to run some smaller scale and cheaper projects that actually demonstrate the technologies are viable before you commit to the whole project. I admit that this is done to some extent, but not enough.

I further admit that it's hard to really test something at small scale and that only building the final version at full scale will really tell you about the problems.

Beyond this, though, we need to do some more conservative engineering.

Saturn V worked, so we threw it away and designed a whole fleet of alternatives.

They each worked, but we threw them away and designed replacements in the goal of making each one incrementally better.

The money sunk in new designs is crippling the space industry. If we simple mass-produced something simple and stupid, it would be better than hand-building something really smart, really clever, really expensive, and really risky.

I still maintain that, for example, having proven pathfinder, we should have built 100 copies of it (with a rechargeable battery in the rover, though), and launch those to 100 target sites on Mars. Each one would cost 1/10 or less of the original pathfinder, and you could put out true contracts to stamp these out of tin like model T fords.

Every Mars researcher could get his or her own site visited.

We would learn an incredible amount about Mars, and could afford the mission risk of visiting exciting sites, since we had equipment risk under control.

*sigh*

<Rant mode off>

back to business as usual...

Apologies to Kim for picking his post to start from. It could have been anyone!