Viewpoint

Forecasting -- an Impossible Necessity

Crystal ball gazing! Wherever we look we cannot escape the need for forecasting and long-range planning: overcrowded suburbs, anachronistic transportation systems and overtaxed health and education systems.

Yet unfortunately we agree with the famous French philosopher and computer scientist Jacques Hebenstreit who said that "any speculation beyond a time horizon of ten years is pure science fiction." Some reasons for this rather pessimistic view are obvious while others though still more important are less well known.

First of all, as everyone knows, science in key areas such as medicine, biochemistry and information science is developing very rapidly. It is a rule of thumb that the "half-life" of knowledge in these areas is six years. This means, for example, that knowledge in the area of computer technology will have doubled six years from now. In 12 years, it will be four times as much, in 18 years, eight times as much; and in 24 years, we will know 16 times as much as we do now. How can anyone think that reasonable forecasts can be made when we have only one-sixteenth (6%) of the knowledge that we will have twenty-five years from now? Perhaps you have heard the saying that if you have understood a new development in information technology then it is obsolete!

The second reason why forecasting is so difficult is that developments do not follow a smooth curve, but go in steps that correspond to inventions and discoveries. Of course we cannot forecast these because if we could they wouldn't be real discoveries. How could anyone predict nuclear warfare back when scientists thought that the atom was indivisible?

Back in 1975, 8-inch floppy disks held about 50 kilobytes of data. Predictions were made then that disks would evolve until they were half the size with one megabyte (about 1,000 kilobytes) of data. While these predictions were surprisingly accurate, they missed the mark entirely -- optical re-write disks can hold one gigabyte (about 1,000,000 kilobytes).

The third reason is more subtle. Even when a discovery has been made, forecasting its effect seems to be much harder that we think it should be. In 1877 Thomas Edison invented and produced the record player. He was quite a salesman and in a classic leaflet about the advantages of his phonograph he accurately predicted many consequences, such as "we can conserve the most beautiful voices, concerts and operas for eternity. We can carry music to the remotest farm house--". However the punchline at the end, written in very large letters, was "and above all we will finally be able to play music faster or slower as we wish"!

Another reason is that everything is blurred completely by factors on a global scale such as economic or ecological catastrophes or wars. Telephony is more than 150 years old. But it developed very slowly, partly because of the economic climate, but also because of the chicken and egg paradox. People asked, "Why should I install a phone when my neighbour doesn't have one? I can't call him and he can't call me." We see the same thing with technology today. We do not have wide-spread use of interactive CD players yet because we do not have libraries of CD disks to play.

Our last reason is more subtle and of even greater significance: if there is a jump in quantity, then there is a jump in quality, in the sense of new properties. If you take any arbitrary phenomenon and increase its size greatly, then you do not just get more of the same phenomenon -- you get something with completely new properties, a completely new phenomenon.

Consider the following analogy. If a Bedouin who has lived in the desert all his life is told about the Tasman Sea at first he will be incredulous. Then even if he does intellectually accept that such large bodies of water really do exist, when you challenge him to speculate on what this implies, he will come back with ideas like "Fabulous: I can water 10,000 camels once."

We believe he has no chance of predicting real implications  such as the development of swimming or boating. Or that if a wind comes up the water will start to undulate so that the boat may capsize and people may be drowned. For a Bedouin water has always had only lifesaving qualities, so that the idea that water can kill will be absolutely alien to him. There is no chance that he can make predictions such as these.

In the area of information technology we are going to witness tremendous jumps in quantity. In 10-15 years, everyone will carry small but powerful notebook computers. You will be able to talk into your notebook and have more commands, programs, and facilities available than we can imagine. For example, if you go overseas your notebook will translate for you. A global positioning system will display maps for you and pinpoint your location exactly. A mobile telephone in your notebook will give you access to all the databases in the world, so you can look up theatre programmes and bus and train connections. It will be your digital camera, and will replace your credit cards. Thus, we are going to witness a jump in quantity in the computer world, which will also introduce new characteristics and capabilities, and what they will be we can predict as little as the Bedouin can about water.

Where does this leave us? First of all, for the shorter term we can do careful linear extrapolations. As we saw in the case of the floppy disk predictions these may provide valuable data. However it is significant that the 1975 predictions, which have been completely surpassed, were thought at the time to be outrageously exaggerated.

We should not underestimate the effect of science fiction on society. Science fiction authors are frequently misunderstood. It is often thought that they are trying to predict the future. What they are really doing is painting scenarios, which may have either positive or negative connotations. We believe that these may have significant effects since mankind will react by purposely striving toward the positives while rebelling against the negatives.

It is because authors like Jules Verne placed powerful images of space travel into the minds of his readers, and men like President Kennedy were inspired, that space programmes were launched. Novels and films like George Orwell's 1984 inspired journalists and laymen alike. Almost certainly it helped spark protest movements as well as having profound effects in areas such as privacy and data protection laws.

Finally, computers are turning into the ultimate media machines and in doing so they will change the way we teach and the way we learn. We can store not only the text but also graphics, pictures, movies and audio clips. Interactive movies are available right now for which you are given the control needed to modify the actual plot -- to the extent of determining whether or not the hero dies!

With the aid of computerised helmets and special gadgetry like "datagloves" the user can manoeuvre through virtual worlds, swim through the body's circulatory system and become totally immersed in environments both real and imaginary. No one can really predict where this jump in technology will take us. Yet it is part of any computer specialist's responsibility to keep a lookout for new phenomena and to alert the public.

We feel that this sense of responsibility is an important part of the training which computer specialists should receive -- a need that few have appreciated yet. But only with this kind of approach will we be able to avoid the worst consequences of the fact that to forecast information technology is important, yet impossible!

Jennifer Lennon works in the Hypermedia Unit at Auckland University's Computer Science Department.
Hermann Maurer works in the Hypermedia Unit at Auckland University's Computer Science Department. Professor Maurer is also with the Institute for Information Processing and Computer Supported New Media at Graz University of Technology in Austria.