by Don Kazak
HAL's Legacy, edited by David G. Stork; MIT Press; 376 pp.; $22.50
HAL may be the first computer that was also a serial killer. As the brain of the spaceship Discovery in Stanley Kubrick's movie "2001: A Space Odyssey" (and in Arthur C. Clarke's science fiction novel of the same name), HAL is certainly famous, as computers go, even fictional ones. So why should anyone care about a fictional computer from a movie released a generation ago?
Because, as HAL says in the movie, "I became operational at the HAL plant in Urbana, Ill., on Jan. 12, 1997."
Yes, we have reached HAL's actual (so to speak) birthday.
David Stock, a Stanford faculty member, remembers that someone came up to him at a dinner party several years ago and asked him, "How realistic was HAL?"
Since Stork is a computer scientist and engineering professor, this was an intriguing question, so he set out to find the answer.
The result is a book with a collection of chapters by noted scientists who examine many of the aspects of HAL in the movie. Famed "2001" author Clarke wrote the forward.
The time of HAL is upon us.
And the book, "HAL's Legacy," is much more interesting than one might initially suspect, especially if one is not a computer scientist.
Unlike many popular science fiction movies, especially the Star Trek and Star Wars efforts, "2001" made an effort to be realistic about the science being portrayed. Clarke, who co-wrote the screenplay with Kubrick and later wrote the novel, has a math degree from Cambridge.
Kubrick and Clarke also consulted with scientists at NASA and elsewhere to depict outer space as correctly as possible, unlike other popular science fiction movies.
One popular misconception that Clarke tries to clear up right at the beginning of the preface is that HAL was named in relation to IBM. (Think of those six letters in the two names and where they are in the alphabet). "I don't know when or how it originated, but believe me, it's pure coincidence," Clarke writes.
That said, what about HAL? And what about the science in "2001"?
Well, as several writers explain in great detail in the book, the science in the movie is very good, even when compromises have to be made in the interests of cinematography. For example, even though Discovery is hurtling through space, a real ship seen in real space would not appear to be moving in relation to distant stars, but that would have looked odd on a movie screen. In the movie, there is a slow drift to the stars.
HAL, the movie's cybernetic villain, is an attempt by Kubrick and Clarke to anticipate in 1968 what standards computer science and artificial intelligence would have reached by 1997.
According to "HAL's Legacy," they gave it their best shot, but they missed.
Several writers in the book, however, explain that the prevailing wisdom in computer science in 1968 was that artificial intelligence would be relatively easy to accomplish and would be right around the corner.
Well, artificial intelligence, or AI, has stumped computer wizards for generations and isn't around any corner within sight of today, it appears.
Thus, many of the things HAL does in the movie simply aren't possible and may not ever be possible, up to and including the computer's decision to commit murder. The last point, at least, should be reassuring.
The one huge miss Kubrick and Clarke make is not to be prescient enough back then to realize that the magic of computers would come as much from the software as the hardware, and giant supercomputers weren't the answer to solving the puzzles of AI. But nobody else knew that back in 1968, either, which was before the personal computer revolution and before Bill Gates became, well, Bill Gates.
In fact, "HAL's Legacy" is a book about a computer in a movie. But it is also a thoughtful and highly interesting summing up of the problems facing people working on AI issues, people who are testing the limits.
And while it is interesting to learn why we won't have HAL anytime soon, it is even more fascinating to get a glimpse of what we might have, and what computers will be able to do.
One real thing computers can do well today is play chess. For the first time ever, the best computer chess programs are beating the best players in the world. And HAL plays a chess game in the movie.
But one of the book's contributors explains what is wrong with the chess game in the movie. Murray Campbell, an IBM scientist and one of the creators of the Deep Blue computer program that beat chess champion Garry Kasparov last year, said that computer chess programs are programmed to make the optimal move and not take risks. In the movie, HAL wins the chess game by a brilliant but risky and very human move.
In turns out that Kubrick, a chess player, used a real but obscure 1913 game for the movie and "selected a clever checkmate but was careful not to employ one too complex for viewers to grasp," writes Campbell.
Writing of Kubrick and Clarke and their movie, Stork concludes that "the filmmakers knew and cared about getting the science right and made as few artistic exceptions to accuracy as possible."
They also made a movie which is breathtaking, with a sense of wonder and beauty. All that, and science, too.
If anything, "HAL's Legacy" increases the appreciation of the art by explaining the science.
Back up to the Table of Contents Page