On Computers Bob Schwabach Advancing in increments The most significant hardware innovation: I don't think there was one. As I reflect on the advance of products over the last 10 years it seems the movement was incremental. Each year you had faster processors, larger storage and memory capacities. Scanners and modems got smaller and cheaper. Digital cameras became a consumer item. But all of these things already existed more than 10 years ago, and what we saw was a steady advance. Even the Internet was in place more than 10 years ago. This is not to dismiss those incremental advances. Differences in price and efficiency lead to differences in use and quantity. Differences in quantity in turn lead to differences in quality. In other words, it makes a big difference whether a computer costs a million dollars or a thousand. Similarly, I can recall when scanners cost $25,000. Now they can be had for $100. Memory chips were $100-$200 a megabyte; now the cost is about $1 per megabyte. The first optical computers are being developed. That will be a significant advance. Optical cabling is moving fast, pun intended. Another candidate would be the large-scale integrated circuit. What is likely to be considered the most significant innovation of this period when looking backward from the next decade has barely appeared yet at the consumer level, and that is the use of multiple processors. Computers with more than one processor will soon be commonplace. Software innovation The most significant software innovation is that which has emerged to make use of multiple processors. The trick is to write programs so that they can be taken apart and those parts processed simultaneously by several chips. The next trick is to bring the pieces all together again at the end. We can expect to see so-called neural networks and expert systems, some of which are already in use by stock traders, fund managers and a few research hospitals. This is software intended to replicate the thought processes of the human brain, and still further, to utilise the knowledge of experts in reaching decisions. It can be done, and a possibly scary thing to think about is that it will be done. Recent tests have found that a computer diagnosis of patient's diseases was far more accurate than human doctors. On a trivial level one could point to recent software advances in image manipulation and virtual reality as the most significant advance. But here again, these programs have been around for many years and the advances, while impressive, are again incremental. In terms of use it would probably be the scanner. Once an expensive and exotic device, it is now found in almost any office and many homes. In terms of growing use it would be the digital camera, in many ways simply a portable version of the scanner. Digital cameras scan out streets, the hallways of our buildings and even intrude on private places. The most important peripheral waiting in the wings will be the sniffer. This would be an electronic nose, sensitive to a wide range of molecules and able to detect everything from poisonous gases to the movement and personal genetic signature of anyone passing by. This is the stuff of science fiction, but is definitely being developed in the labs and is almost ready for prime time. Significant industry event The most significant industry event was IBM making their computer design and operating codes an open system, so anyone could write software for their computers. Of course that happened more than 10 years ago. Information technology Everybody's going wireless. Satellite systems are bringing us global communication, and nearly unlimited talk and data transmission at low, low prices. What's a traditional telephone company to do? Frankly, I don't know; they might as well just turn out the lights and go home. Logitech Fotoman Plus The challenge is transmission speed, of course. It can be addressed in a number of ways, but the fastest and simplest is packet transmission. High-speed bursts of multi-frequency signals can carry enormous amounts of information in fractions of a second. Laser beams can be modulated in a manner similar to radio waves. Not all light beams need be in the visible range, so the total range of transmission vehicles, in other words, the carrier waves, is essentially infinite. The biggest current challenge is mass storage devices that can handle the traffic. Because at some point the huge amount of digital traffic traveling the airwaves or through solid conductors has to stop at a destination. It is not going to fit on your father's hard disk drive. Which brings us to the next item... What the future holds: Out with the new, in with the old. The future holds an infinite number of possibilities in theory but a finite number in practice. One of the most interesting solutions to the problem of mass information storage and retrieval is paper. Yes, paper. Only this is called "digital paper." The process has already been developed and been tested for several years. It is so close to mass market that we should begin to see it in a year or two. It is a process of laying down complex patterns of microscopic dots that can later be read by a scanner and translated by the computer. It is already done in a small way with bar codes. The dots can be printed by a fairly inexpensive high resolution laser or inkjet printer. A fine finish paper will be used, much like the kind now used for quality color prints with an inkjet printer. The paper solution is a bow to the past and a nod to the future. The problem of storage is seldom gone into but boils down to this: Magnetic storage, the most common type up till now, produces disks and tapes that have a probable life span of a few years to a decade. Even then the storage must be carefully controlled; too much heat, or stray magnetic currents, and the disk becomes unreadable. Optical storage on CDs would seem to be an obvious and permanent answer, but in fact it is estimated that CDs will last no more than 30 years. Few people realise it but acrylic plastics, the type used for disks, expand and contract with temperature changes. Unless they are kept in a carefully controlled environment, such changes over the years will distort the disks sufficiently to make them unreadable. Do you want this to be your library of essential information? Paper, on the other hand, has already demonstrated the ability to last for centuries with only minimal care. Put it between binders and set it on a shelf and you can take it down and read it 500 years later. Under controlled conditions there is no reason to believe it would not last for thousands of years. A digitally encoded page could hold the equivalent of an entire CD on a single sheet of paper. The first steps (at lesser capacities) are already in development at companies like Hewlett Packard, which may introduce such storage as early as next year. Is it hardware or is it software? Research on organic computers began about 15 years ago. Nothing very much has come of it but the idea still lives, so to speak. The idea was that a computer that was alive would have certain advantages. If it had insufficient resources to solve a particular program, it would have a genetic incentive to grow more. Nutrients would be provided and eventually, as the computer realised the limitations of being fed nutrient baths, it might be able to obtain its own food supply. This all sounds like one of those low budget science fiction movies from the 50s, but in fact some small progress was made along these lines. It was found that is was possible to grow an organism that could store and process very simple digital data. The direction of the research then pointed to ways the organism could grow more cells as its computational needs increased. The falling prices of electronic methods made further research unfeasibly expensive and I've heard nothing more of the matter for more than a decade. But perhaps some day ... NOTE: Readers can search more than three years of columns at the "On Computers" web site: <http://www.oncomp.com.> You can e-mail Bob Schwabach at bobschwab@oncomp.com or bobschwab@aol.com. Back to top