The mechanism for setting the sail was simplicity itself. Instead of a complex mechanical operation to raise and lower it, a series of gastight tubes were sewn around the exterior and connected to the main ship by much thinner tubes. Gas was pumped into the tubes to set the sail, pumped out while thin filaments were retracted to furl it. Heating elements within the tubes kept the gas from freezing and collapsing in the cold of deep space.

Other problems, microminiaturized electronics and an extremely lightweight spacecraft body, had been easier. Indeed, they had been almost natural outflows of ongoing, purely terrestrially oriented, research. It was a short step from nanotube body armor for soldiers to a nanotube spacecraft body, for example. The programming had been even easier if not precisely simpler.

Not to say that the ship was cheap. It had eaten up almost all of the United States' National Aeronautics and Space Administration's somewhat constrained budget for the better part of two decades. The less said about the scandals, the overruns, the bribes from various foreign subcontractors, however, the better.

The ship, if one could call a robot a ship, was named the Cristobal Colon. Many had held out for a different, generally more culturally sensitive and less eurocentric, name. These ranged from Saint Brendan and Leif Eriksson (obvious nonstarters) to Sinbad to Cheng Ho. Since the Americans were footing the bill, however, they got to choose. Moreover, they were, at the time, going through one of their periodic bouts of extreme nationalism. "Cristobal Colon" seemed good to them and the rest of the world could lump it.

The robot, or ship, was just under two meters in diameter and approximately nine long. Various projections-a radio telescope here, an antenna there-were attached to the outside. The computer that controlled it was deep inside, or as deep inside as one can get with a cylinder two meters across. The sail dwarfed the robot ship, though the sail massed very little and the ship several tons.