He could make his observations and guide his actions on the basis of those observations, but the Project wanted his actions to be intricate and subtle, at least potentially, and they were not at all sure what observations he might make.
To prepare for all reasonable possibilities and to allow for all the intricacy desired, the robot would need to contain a computer (some at Dallas referred to it as a "brain," but Anthony scorned that verbal habit- perhaps because, he wondered later, the brain was his brother's field) sufficiently complex and versatile to fall into the same asteroid with a mammalian brain.
Yet nothing like that could be constructed and made portable enough to be carried to Mercury and landed there- or if carried and landed, to be mobile enough to be useful to the kind of robot they planned. Perhaps someday the positronic-path devices that the roboticists were playing with might make it possible, but that someday was not yet.
The alternative was to have the robot send back to Earth every observation it made the moment it was made, and a computer on Earth could then guide his every action on the basis of those observations. The robot's body, in short, was to be there, and his brain here.
Once that decision was reached, the key technicians were the telemetrists and it was then that Anthony joined the Project. He became one of those who labored to devise methods for receiving and returning impulses over distances of from 50 to 40 million miles, toward, and sometimes past, a Solar disk that could interfere with those impulses in a most ferocious manner.
He took to his job with passion and (he finally thought) with skill and success. It was he, more than anyone else, who had designed the three switching stations that had been hurled into permanent orbit about Mercury- the Mercury Orbiters. Each of them was capable of sending and receiving impulses from Mercury to Earth and from Earth to Mercury. Each was capable of resisting, more or less permanently, the radiation from the Sun, and more than that, each could filter out Solar interference.
Three equivalent Orbiters were placed at distance of a little over a million miles from Earth, reaching north and south of the plane of the Ecliptic so that they could receive the impulses from Mercury and relay them to Earth-or vice versa-even when Mercury was behind the Sun and inaccessible to direct reception from any station on Earth ' s surface.
Which left the robot itself; a marvelous specimen of the roboticists' and telemetrists' arts in combination. The most complex of ten successive models, it was capable, in a volume only a little over twice that of a man and five times his mass, of sensing and doing considerably more than a man- if it could be guided.
How complex a computer had to be to guide the robot made itself evident rapidly enough, however, as each response step had to be modified to allow for variations in possible perception. And as each response step itself enforced the certainty of greater complexity of possible variation in perceptions, the early steps had to be reinforced and made stronger. It built itself up endlessly, like a chess game, and the telemetrists began to use a computer to program the computer that designed the program for the computer that programmed the robot-controlling computer.
There was nothing but confusion. The robot was at a base in the desert spaces of Arizona and in itself was working well. The computer in Dallas could not, however, handle him well enough; not even under perfectly known Earth conditions. How then
Anthony remembered the day when he had made the suggestion. It was on 7-4-553. He remembered it, for one thing, because he remembered thinking that day that 7-4 had been an important holiday in the Dallas region of the world among the pre-Cats half a millennium before- well, 553 years before, to be exact.
It had been at dinner, and a good dinner, too. There had been a careful adjustment of the ecology of the region and the Project personnel had high priority in collecting the food supplies that became available-so there was an unusual degree of choice on the menus, and Anthony had tried roast duck.
It was very good roast duck and it made him somewhat more expansive than usual. Everyone was in a rather self-expressive mood, in fact, and Ricardo said, "We'll never do it. Let's admit it. We'll never do it."
There was no telling how many had thought such a thing how many times before, but it was a rule that no one said so openly. Open pessimism might be the final push needed for appropriations to stop (they had been coming with greater difficulty each year for five years now) and if there were a chance, it would be gone.
Anthony, ordinarily not given to extraordinary optimism, but now reveling over his duck, said, "Why can't we do it? Tell me why, and I'll refute it."
It was a direct challenge and Ricardo's dark eyes narrowed at once. "You want me to tell you why?"
"I sure do." Ricardo swung his chair around, facing Anthony full. He said, "Come on, there's no mystery. Dmitri Large won't say so openly in any report, but you know and I know that to run Mercury Project properly, we'll need a computer as complex as a human brain whether it's on Mercury or here, and we can't build one. So where does that leave us except to play games with the World Congress and get money for make-work and possibly useful spin-offs?"
And Anthony placed a complacent smile on his face and said, "That's easy to refute. You've given us the answer yourself." (Was he playing games? Was it the warm feeling of duck in his stomach? The desire to tease Ricardo?...Or did some unfelt thought of his brother touch him? There was no way, later, that he could tell.)
"What answer?" Ricardo rose. He was quite tall and unusually thin and he always wore his white coat unseamed. He folded his arms and seemed to be doing his best to tower over the seated Anthony like an unfolded meter rule. "What answer?"