Computing Science 101

Capabilities

William Gibson said he was disappointed when he got his first computer. He found it couldn't do all the things he thought it could. This is one of the more subtle inaccuracies in the media. On Star Trek: The Next Generation, Data or Dr. Crusher would ask the computer to perform a complex calculation or analysis. Captain Picard will ask the computer a very subtle and complex question and get back the answer fairly quickly. In 2001: A Space Odyssey, HAL can carry on a conversation and pass a Turing Test without blinking an unblinking eye. We've seen computers do amazing things with amazing speeds, information and problem solving. If computers could do everything they did in movies and television, my life would be a lot easier. I could just ask my computer to research all the references I need for this essay!

I think the problem is society views the things computers do with awe and amazement and can’t wrap their minds around the idea that some things are very hard for a computer, especially the kinds of things we take for granted like seeing.

Computer Analyses Made Easy

The one thing I see in television and movies that I wish was true is computer analysis. In real life, if I wanted to analyze a large database, I would have to spend several days (sometimes weeks) writing a computer program to analyze the data. I have to figure out everything for the computer: how to fetch the data, how to compute the analysis and how to present the data back to me. This is the current state of the art. There are increasingly applications that makes things easier, but they are little more than summarizers. If you want to do any real, complex analysis, you have to start writing code.

In Star Trek: The Next Generation, Dr. Crusher says it will take her three hours to program the computer to take every star in the galaxy, figure out their relative locations several billion years ago and then find a pattern scattered amongst those stars. Man, what I wouldn’t give for a computer that could do that, but there are several implicit assumptions about computer capabilities in that scene. One is that you can just talk to the computer and the computer will figure out what you want. We are still a decade or more away from that. Computers know nothing. They can only do amazing things if some human tells it explicitly how to do it. You want to know what your computer really is? Take out a hand calculator. That’s your computer. It can make a few simple decisions (is this number less than this number?) and take different steps accordingly, but that’s it. It knows nothing of galaxies, orbital mechanics or mathematical analyses. A human must still program the computer to get it to do these things.

Know It All’s

Another assumption is omniscience: computers know everything. Every scrap of information has been digitized and stored in the computer in a format it can easily search, analyze, manipulate and use. In Babylon 5, John Sheridan wants to find out some information so he asks his second-in-command to do a computer search to find out about someone who lived at a specific address in London, England sometime during the 19^th century with the name Sebastian ("last or first"). It takes less than a day, mostly the communication lag time. Did Earth digitize every scrap of paper humanity has ever made into a computer, converted it into a searchable document and made it available on a global network that anyone can access? That would be very cool.

In Star Trek: Voyager, it has been mentioned several times that the Voyager’s computers contain the entire knowledge base of the Federation. Every major work of art, entire histories, holodeck programs. The whole ball of wax. This either says something about the storage capacities of the U.S.S. Voyager or the sum knowledge of the Federation.

In the real world as this time, very little information is available digitally, especially if it is old. If it is available digitally, I can almost guarantee you it’s not in a format a computer can easily read and analyze for itself. This may change, but not for a long time. So if you see someone in the 20^th century performing an exhaustive computer search without having to go to the library, you know you’re dealing with fantasy.

Artificial Intelligence

The final assumption is intelligence: the computer has an artificial intelligence that has been taught a great many things, especially about mathematics, science and astronomy. It can carry on and follow a conversation. That’s pretty good considering the Microsoft Word grammar checker can’t seem to tell the difference between a verb and a noun.

This might not be as far fetched as it seems. There is a lot of work going on to make computers voice activated. With the advent of mobile phones which don’t have keyboards, companies are suddenly interested in this neglected field. But how far can they go? Can they create a computer that can completely understand a human being and follow a conversation? Just from my gut feeling for the matter, I’d say yes, and we may see some surprisingly good applications of this soon, but I don’t think they could formulate an analysis on their own for quite some times. I don’t believe we’ll be able to create a general purpose AI that can create its own algorithms in the near future. If you think about the kind of education and experience a person needs to take a request to "scan a planet for the capitol city from orbit", you can imagine the programming challenge. I don’t believe it’s impossible; just not likely to be done anytime soon.