To Jupiter, Alice! To Jupiter!

With beautiful timing, the New Yorker is running a beautiful article on the incredible success of the Galileo probe. Launched fifteen years ago with technology that was a decade out of date at the time, Galileo discovered the first extraterrestrial ocean, holds the record for most flybys of planets and moons, pointed out a dual star system, and told us about nine more moons of Jupiter.

Galileo's story is the story of improvisational engineering at its best. When its main 134 KBps antenna failed to open, NASA engineers decided to have it send back images using its puny 10bps antenna. 10 bits per second! 10!

To fit images over that narrow a channel, they needed to teach Galileo some of the tricks we've learned about data compression in the last few decades. And to teach an old satellite new tricks, they needed to upgrade its entire software package. Considering that upgrading your OS rarely goes right here on Earth, pulling off a half-billion-mile remote install is pretty impressive.

(I'm also curious about the security issues. Does NASA use crypto? Could just anyone with a transmitter point it in the right direction and ask Galileo to searh for rings around Uranus?)

As if that wasn't enough hacker brilliance, design changes in the wake of the Challenger explosion completely ruled out the original idea of just sending Galileo out to Mars and slingshotting towards Jupiter. Instead, two Ed Harris characters at NASA figured out a triple bank shot -- a Venus flyby, followed by two Earth flybys two years apart -- to get it out to Jupiter. NASA has come in for an awful lot of criticism lately, but there are still some things they do amazingly well.

But let's talk a bit about that criticism. You know, about safety, and how they can't make honest estimates of odds when lives are at stake. I've expressed similar sentiments in this space before. The Galileo story points out, I think, that the problem is not that NASA is messed-up, but that manned space flight is messed-up.

Go back and reread the first sentence in the last paragraph but one. Those "design changes" involved stripping Galileo's liquid-fuel propulsion system; after the Challenger explosion, liquid fuels were considered too dangerous to carry as cargo. The whole triple bank shot was a brilliant hack to compensate for the externally-imposed design constraint of not using liquid fuel.

The incredible difficulty and riskiness of manned flight, in other words, dictated the conditions under which the unmanned operation had to function. Which, when you consider the "scientific" justification for the space shuttle program, is completely backwards. Instead of designing a space program to put up the spacecraft we want to launch, we're launching precisely those spacecraft that a shuttle designed and built twenty years ago can handle.

Manned spaceflight is, in the Ursula K. LeGuin sense, perverse. It's an act of pure conspicuous waste, like eating fifty hotdogs or memorizing ten thousand digits of pi. We do it precisely because it is difficult verging on insane. You almost don't need to put the people in the space shuttle: we've built, at enormous expense, a system which is perfectly capable of taking people up into space and bringing them down again safely, most of the time. And that's a nice feel-good show-off national achievement. I guess.

But something like Galileo puts things back in perspective again. We can learn about this amazing universe we inhabit and we can pull off the most stunning feats of ingenuity, and not everything worth doing comes with mortal peril attached.