The Causal Nexus

by Richard Gibbs

"If you destroy Logopolis, you unravel the whole causal nexus!" "Causal nexus! You insult my intelligence, Monitor."

All right, so I don't know what the causal nexus is either, but hopefully I won't insult intelligence in this article. Logopolis is all about computers, and unlike all the "megabyte modem" business in recent years, the technology in this story actually makes some sense. This is hardly suprising, since the author (Christopher H Bidmead) knew much about computers, having written for Practical Computing for some time. Within this story, he presented some remarkable concepts about the future of computers, but sadly they were not developed fully in the four episodes available. To analyse his ideas, I would like to compare Logopolitan mathematics to existing technology on earth, For anyone who knows little about computers, this may seem rather baffling, whereas those with some knowledge might find this a bit simplistic. Sorry.

Machine architecture. Sounds good, doesn't it? What it really means is the way all the bits and pieces inside a computer are conected together. (Obvious really!) Within your standard ZX Spectrum, BBC micro, IBM PC or whatever, there is one central microchip that does most of the work, called the Central Processing Unit, or CPU. This contains a few "storage areas" called registers, each of which can contain a number. Although the details vary from computer to computer, nearly all have an "accumulator", or A-register. Most operations are carried out on the contents of the accumulator.

in addition to the CPU in a computer, there is also a large area of memory. This usually contains both the program, and the data to be processed. At any one time, the other main register, the Program Counter, "points" to one of the numbers in memory, to show the point in the program to be executed. The CPU reads this number and translates it into an instruction, which it then obeys, Lastly the Program Counter is advanced, so that it points to the next instruction needed. For instance, an instuction might say "look at memory location 1 and copy the number you find there into the accumulator". The next could say "look at location 2 and add the number you find to the accumulator". Thus after obeying these instructions, the CPU would have added two numbers with the result left in the accumulator. Now you can probably see how the accumulator was named.

This description shows one of the main limitations of today's computers : all the processing of the data is done inside the CPU. Hence much time is wasted transferring the numbers from memory to the CPU and, in addition, only one instruction may be carried out at any one time. If Logopolitan methods are examined, though, a subtle difference can be noticed. The basic parts of a computer can be seen, with the Central Register building serving for a CPU, and the individual ihhabitants being used as memory locations. However, each of these "memory locations" has its own computer associated with it - namely the individual's brain. Thus when the program is running correctly, there would be hundreds of smaller computers each working on a small part of the problem. In this type of computer, the CPU would not be capable of directly accessing each memory location (that's why the Monitor and Adric had to go out onto the streets to find the error). Instead, a hierarchical system is used, with the Central Register splitting the main problem into a few easier tasks, which are each given to the most senior Logopolitans. They in turn delegate others to do smaller parts, all by word of mouth, so that eventually the entire population is involved. The answers to each of these smaller tasks are then passed back up the hierarchy to the Central Register.

This method is much more efficient, since all the different parts of the program are executed simultaneously. Only in this way could the Logopolitans keep the CVE's open, as conventional systems could never calculate the information in time. Looking from a scientific point of view, though, It is interesting to note that this story was written in 1981. It is only recently that "transputers" have been developed. These are small chips, each with its own processing power and memory. They are arranged in a hierarchical system, each executing a small part of the program. I wonder if Christopher Bidmead played a part in the most exciting breakthrough since the microchip, or perhaps he just recognised the way developments were going in the computer industry.