Arthur C. Clarke was one of the most respected writers of science fiction. Best known for 2001, he wrote convincingly about space travel, the colonisation of the Moon and other parts of the Solar System and how fundamental scientific principles might be used to overcome the huge hazards and costs of these ventures. But he didn't know much about computers, even though his HAL, the computer that runs amok in 2001 is the most famous example of its kind. I have been reading 2061, a follow-on to 2001 and its successor 2011. Clarke depicts his hero, Dr. Floyd, thinking of a line of poetry, trying to recall the author and musing that no more than ten minutes on the world's computer network would produce the answer. So I took that line of poetry and put it into Google and got the answer at once. In 2013, only 26 years after 2061 was written, we can do better than Clarke imagined would be the case in 48 years from now.
Actually Clarke was not alone. Isaac Asimov, no mean scientist himself and author of a huge number of SF books, was obsessed with the idea that one computer (which he amusingly called "Multivac", probably a jibe at the Univac series of early computers) would do all the world's computing and only men wearing white coats would be allowed anywhere near it. Oh, and all communication with it would be by punched cards and input tapes. In his most famous work, Foundation, set thousands of years in the future, he envisages his scientist heroes using "calculator pads spotty with age", although to be fair, they can control their monitor displays using brain power alone (but I suspect that we may be able to do this in a generation or so).
Cheap and fast space travel is as far away as it was when 2001 was published but the leaps in computing power enabled by networking and by the Internet have surpassed the imaginations of some of the giants in SF. Funny, eh?