Martin Banks, Personal Computer World 10/84 - checked

October 1984

Mother slaved away as mothers always have. The children wanted breakfast and, as usual, she had to prepare it.

As they sat around the table in expectation, knives and forks akimbo, she placed before them the products of her endeavours. 'Now remember to eat your chips, dear, 'she told the youngest. 'You'll never be brainy if you don't eat your chips.'

'Aw, but Mum,' whinged the small person, 'they taste awful. They don't taste a bit like the ones Grandma makes.'

'Yes dear,' Mother replied, 'but hers are made in the old-fashioned way - from potatoes. These new ones are much better for you. They're full of fibre and trace elements - and they're good for the brain.'

'Good for the brain ?' the small person enquired.

'Yes, eat enough of them and they might grow into one inside you.'

Yes, well. Ever since the national newspapers first discovered the microprocessor et al and found that they were generically referred to as chips, stories of the chip kind have been found everywhere. In time, all that clever stuff passed on, never, one hoped, to be seen again. But current research and development efforts are leading us inexorably towards the big-chip and more 'chip'-based stories.

More properly called molecular electronics, big-chips are expected to be around in workable form within the next 20 years. This may seem like a long time, and it is when one considers just how far the microcomputer business has come in under 10 years from the appearance of the early Altairs and Imsais. But it's no time at all in the great scheme of things, and when one considers what might be possible with big-chips, the progress made so far in electronics starts to look like the early struggles of the Wright Brothers.

It is first necessary to consider what molecular electronics is all about before thinking about its implications. In essence, what is happening is the development of a technology that can create circuit elements which are some 50 nanometers long: (the size of an individual molecule). In the world of microelectronics, it means the ability to cram into a single device much more than before. The `more' in question can be circuit functions like l/O systems and device drivers, or that all-important commodity - memory. Storage capacities thousands of times greater than anything currently envisaged could be possible, while the complexity of functions available in a single chip is nothing short of mind-boggling.

All this is well beyond the capabilities of any semiconductor technology currently used to process silicon-based integrated circuits. Indeed there are still two schools of thought on which is the best way to achieve such results - the biological route or the biochemical route. You will note that the same prefix appears in both cases - bio. Organic materials such as proteins are to be the building blocks from which these devices will be manufactured. It's quite possible that genetic engineering will be used to create microbes which will actually synthesize circuit elements, or even complete circuits.

Not surprisingly, there are many technical problems to be solved before molecular electronics appears as a working entity. Not least among these problems, according to one forecast report from Technical Insights in the USA, is just how connections between the molecules will be achieved. They will have to be brought together in some form of matrix if there's to be a working circuit, but achieving it with such small elements is going to require something completely different from good ol' copper wire.

Despite such apparently fundamental problems, the Technical Insights forecast is predicting that the first demonstration chip components, items such as simple switches and memory elements, could well be in existence within five years.

This is not very long to wait and see if the new technology does emerge. And if it does then be ready, for it's likely to bring with it not only major opportunities for new applications of electronics and computer technologies; it could also wreak havoc with existing applications and design approaches. The computer, as we currently conceive of the beast, may be doomed.

As an example of what might be possible, take that most typical of applications - the small business management system. Take also the area of portable computing, the convenience of which interests all of us. Many people and companies have, from Adam Osborne onwards, tried to combine these two into a single entity with varying degrees of success.

The Osborne had the essential applications but was somewhat hernia-inducing when it came to portability. The smaller the machine, however, the less it meets the applications criteria, until one ends up with pocket `computers' that are essentially calculators with ideas above their station.

Molecular electronics holds out the promise of Gigabyte-sized memories driven by 64-bit processors, all in a pocket calculator package. And a keyboard? Well, with processing power like that available, throw in speech recognition and synthesis for nothing: the pocket-sized business computer becomes quite feasible. The amount of processing power that could be squeezed into a small space would make it feasible to spread computing power around. There would be more than sufficient power to handle all the communications problems, and overcome the majority of protocol and operating system mismatches that are the great hindrance to inter-communications today. Instead of working on producing new standards that everyone should adhere to, we could work on producing systems which didn't really care too much what 'standard' was used. It could run everything from Kansas City audio tape, through CP/M and MS-DOS, then go on past Unix towards whatever comes in the future.

And with that level of power coupled with complex communications (all in a tiny little package, remember), why not have systems implanted in us? The sci-fi ideas of implanted computers to allow the blind to 'see', or to control artificial limbs, could all be possible applications.

And the opening scenario? Even that could happen. We could start eating our way to a better memory and see IBM buy Kellogg's as its new distribution company.

end