Martin Banks, Personal Computer World 05/88 - checked

June 1988

It all seemed so innocent at the time. There was I, in my local hi-fi shop, attempting to buy a couple of good quality audio cassettes for a special little recording job I had in mind. No, I was not going to pirate any games tapes.

I made my selection from the rack, went bravely towards the check-out and waited while the young man exercised his right bicep extensively on a fearsomely complicated cash-till. Needless to say, the shop was one of a national chain which, like many of them, demands that their staff key in all sorts of extraneous 'management information' about the fact that you have just bought a can of baked beans. (I keep waiting for them to start asking me my age, colour, socio-economic grouping and sexual orientation, just to make the picture complete).

Anyway, I stood there for an hour or two while this guy frantically bashed the keyboard of the till. Just as I was getting comfortable, he stopped. He looked at me. It was a very sheepish look.

"According to this," he said, pointing at the till, "you owe us over £1 million." Now, I may be a bear of little brain, but this suggestion had even me doubting the veracity of the suggestion, just a little.

He started frantically re-keying the whole, tedious transaction. It was even more tedious this time, because he was dead keen to get it right. Eventually, we settled on the £4.49 it had said on the price tag all along.

This incident got me to thinking, so I rested for a while in the hope that the mood would pass. But it wouldn't, so I decided to put some of the thoughts on paper (the polite ones about information, not the impolite ones about a certain company's cash tills - this is a family magazine, after all).

I think we would all be in general agreement with the theory that the computer is the ideal tool for the management and manipulation of information. Actually, I guess the human brain is a good deal better, but it tends to forget this fact. Anyway, the computer is relatively well equipped for the task.

It is at this point that I then start to contradict myself, for I am not sure whether the computer is the ideal tool, certainly in its current form. The incident in the hi-fi shop gives some clue as to why I think that way. We are entering a time when we, as either direct computer users or the sufferers of their exigencies, are having to confront information as more than something which just 'is'. We are having to think about how it is constructed and what it is actually made of, so that we develop the best ways of manipulating it.

In our western cultures, of course, a goodly amount of information is primarily constructed of words, which are themselves made up of character strings. This is a very convenient way of constructing information, for a limited set of symbols can be combined in any number of ways to form larger symbols that 'mean something'. The fact that they are symbols can be seen from the way it is noticeable when we come across an unfamiliar word and have to stop reading freely to 'construct' the big symbol from its component parts.

This is something the human brain does naturally, and we maybe tend to forget the amount of processing that goes into its happening. Yet, the incident in the hi-fi shop shows what happens when we apply the same structure for information onto a computer-based system. To enter the information requires an amazing collection of keystrokes. As users demand more information, for example not just the price, but the product code, the product category the sales staff ID number the branch ID number and the customer's inside leg measurement, so these character strings get longer and longer.

This poses two questions: is the information collected actually necessary? and assuming that it is, is the character string the right component from which to construct it? It is easy to forget, for example, just what sort of moron the computer is. Working with single characters is meat and drink to the beasts, for that is the way their internal workings are structured. Yet, when compared to the dumbest human brain (possibly mine), the largest Cray supercomputer pails into processing insignificance.

These are the type of machines that scientists are now trying to teach to 'see', and are finding the problems enormous. The reason is that the machine is still trying to process these much larger 'symbols' as a pattern of 'O's and '1's, which is about as simple a symbology as you can get. It is a symbology which works, but only after a fashion. Currently, you can show a computer something, and come back three days later to see if it has made up its mind what it is looking at.

So, have we come to point where we really ought to start thinking seriously about changing the symbols that we use to define information? I know that people are starting to exploit the capabilities of symbolic processing languages like Lisp and Prolog, but I'm not sure such things go far enough.

I am reminded of something I read once in a book about Zen Buddhism. The author suggested that such a philosophy could not have developed in the west simply because of the linear nature of our written and spoken language. Because our language is constructed from small symbols, all thought has to be constrained by the fundamental nature of the serial bit stream from which it finds its existence.

The author suggested that the oriental pictogram was much better, much more free for expressing grand concepts in a single symbol. Is it now a valid argument to suggest that we consider the impossible - changing the whole nature of our own language so that we can grasp the concepts needed to make computers work in the same way? For without such a change in the long term, I suspect that the computer will eventually choke itself on the millions of Terabytes of serial data it will produce in the name of 'information'.

Now, anyone for Kanji?

end