Visions

From ENIAC to anything?


After 50 years of digital computing, where is miniaturisation taking us?


On February 14th, 1946 throwing a power switch totally changed all our lives. "How?", you may ask. Simple. That switch was attached to one of the most innovative devices made this century - ENIAC. ENIAC stood for the Electronic Numerical Integrator and Computer and it was the first digital computer. To say that a lot has happened since then is the understatement of the century. Follow us down the 50-year path to 1997 and take a glimpse into computing's future.

Before ENIAC, computing was highly advanced. The only problem was that it was done by humans - roomfuls of men and women used to take three months to perform the ballistic trajectory calculations that ENIAC performed in 30 seconds. Powerful for its time, ENIAC, however, was a long way away from today's notebooks. With 0.05 MIPS (Millions of Instructions Per Second) it was a touch less powerful than today's Pentium -(about 100MIPS), it had just a little less memory (200 bytes, Toshiba's Tecras take up to 144 MB, that's 0.7 million times more memory!) and it was just a touch larger (you needed a fairly large living room to install it!). Still, it was enough to start the ball rolling towards the next major development to shake the world - the computer chip.

From room-sized to pocket sized

In the 20-odd years between ENIAC and the computer chip, computers made major progress. By 1969, 16-bit computers were available, and operating systems not too far removed from the basics behind those used today were already available. These computers moved from using valves to transistors in the late 1940s and then, when Fairchild shipped the first integrated circuit (IC) in 1961, they quickly adopted these due to their considerably lower cost and improved power consumption.

The computer chip was developed by Intel and first hit the market in 1971, with approximately the same power as ENIAC. However, where ENIAC cost over $400,000, the first chip, the Intel 4004, cost $200. And instead of taking up a room, as did ENIAC's 18,000 vacuum tubes, it was a mere 12mm2. Designed for the first ever electronic calculator, the 4004 made some compromises, mainly by using a 4-bit bus, but Intel soon saw that the 4004 had a bigger future than that. The microcomputer was born.

Before the 4004 was shipped, the 8008 was already in development in Intel's labs. This was a true 8-bit chip but was not reliable since both the instructions and data were carried on the same bus. Two years later, the Intel 8080 was born with an architecture remarkably similar to today's processors: it had separate busses for the data and the instructions.

In these early days, Gordon Moore, Intel co-founder and chairman, predicted that the number of transistors which could be placed on a chip would double every 18 months. This prediction, since named Moore's Law, has remained true ever since. Andy Grove, Intel's current chairman recently expressed his belief that it will hold for up to ten years more based on current technical knowledge. Moore's Law growth quickly developed the microchip, so that by the early eighties the stage was set for perhaps the most important computer ever, the IBM PC.

Computing made personal

The IBM PC was a real revolution at the time. It was equipped with a 4.77MHz Intel 8088 (capable of 0.33 MIPS) and an amazing 16kB RAM (most people thought that 64kB was an unnecessary luxury!). There was no hard disk drive, and storage was either on cassette tapes or on an optional one or two 160kB 5.25" floppy disk drives. With the arrival of the IBM PC, however, came DOS and with that another legend was born, Microsoft.

It was the triumvirate of Intel, Microsoft and IBM which was to shape the business and, to some extent, the home computing market. However, in 1982 Apple joined the fight for your desktop with their Lisa. Whilst this was not successful (it never really reached volume production), the Macintosh, launched in 1984, quickly swept the computing world by storm. It had a WIMP interface (Windows, Icons, Mice and Pointers) which was such a step forward when compared to DOS that it set the standard with which even today's market leader, Windows95 is compared.

Well before the advent of Windows95, however, mobile computing also became a viable option. After other companies' early attempts failed, Toshiba launched the first successful mobile computer, the T1100 in 1985. At just 4.1 kg in weight it was truly mobile, and the T1100 came with its complete operating system, DOS 2.11, in ROM (try to do that with Windows95)! We've certainly come a long way from those days. Compare today's Tecra 740CDT with the T1100!

Out onto the road

Miniaturisation is one of the greatest advantages of the current stage of the computer revolution, and it is an area in which we are seeing some fascinating developments. People need four basic things on the road: processing power (why compromise just because you're on the move), battery life (the whole point is not to be tied to a power point), easy access to the functions or information you need and, as always, a good price. Three existing breeds of mobile computer provide this, each in their special way, with a new breed, which Toshiba will soon bring to Europe, waiting in the wings.

The first breed is the almost ubiquitous notebook computer. These devices are getting smaller and cleverer each day - take Toshiba's Portégé 660CDT which offers full multimedia with 10-speed CD-ROM and a 150 MHz Pentium processor in a box considerably smaller than an A4 pad. Then, at the opposite end of the size scale, there are the power beasts such as Toshiba's Tecra 740CDT which has a massive 13.3" screen and a 3GB HDD, driven by Intel's latest 166MHz Pentium with MMX™ technology.

Digital assistance

The second breed now gaining popularity is the PDA, or Personal Digital Assistant. These small devices pack a lot of power into your pocket and are made to offer quick access to information and to allow you to enter information with the minimum of fuss. Two barriers to their widespread acceptance have always been foremost: price and functionality. But a new device from modem manufacturer U.S. Robotics has smashed these barriers.

This is the Pilot, a 150 g device which easily fits in your shirt pocket and solves the problem of data entry with an intuitive new data entry system called Graffiti™. It also offers quick synchronisation with your PC for no extra cost. And its price also broke an important boundary in this category - the $500 barrier. As the only reliable handwriting-recognition device selling for $250, the Pilot has taken the PDA world by storm, and is widely tipped as a sign of times to come.

For those who break out in a cold sweat at the idea of too long a break from Windows, Microsoft has created the third breed of mobile PC - the HPC or Hand-held PC. A new version of Windows called Windows CE comes with cut down versions of Microsoft Office programs and hardware requirements so reduced that lightweight hand-helds can be built for Windows junkies on the road to get their fix!

Acceptance of this new concept has been mixed, and some people feel that this current version of Windows CE may have some limitations. However, Microsoft never keeps us waiting with new innovations, so keep an eye on these machines, maybe you'll be pleasantly surprised.

Viability vetoed?

For a long time, mobile computer manufacturers have pushed to the limits the barriers of miniaturisation. So are there limits on how small a viable notebook can be made? With a range of computers under the Libretto brand name, Toshiba is currently destroying the barriers the market considered were fixed. Arguably here is a new breed of device - notebooks which are so small that carrying them need never be a chore, but which enjoy computing power in the same league as their larger brethren.

Imagine a Pentium notebook weighing in at 0.8kg with a 6.1" TFT screen, an 810 MB hard disk and all the functionality you are used to in your full-sized notebook. You've just imagined the Libretto 50, a Japanese featherweight notebook that will hit European shores in the very near future.

The trend is clearly more power per cubic centimetre. Yet at the same time there has been a dramatic drop in price against power. Remaining technical limitations will define how these trends develop. For example, battery technology has not advanced as fast as expected, so currently devices are limited by the amount of battery power they need. Also smaller PCs contradict the preference for larger screens (in notebooks) creating a challenge to find a new concept. This is what the computer industry loves most!

Human integration

One oft-reoccurring fiction is the degree to which man and machine can be integrated. An example of research work being done on this comes from British Telecom at their Martlesham Heath Laboratories. Connectivity is a growing theme in mobile computing, and understandably, this is a major part of BT's strategy in making what they call the Office on Your Arm. Based on a (highly modified) Apple PowerBook, the office fits on an arm and includes a GSM telephone allowing email, telephony and fax, wherever you are.

To create the desired large screen effect, instead of the small LCD display many find too hard to read, a visor projects a larger image in front of the user's eyes. In this vision small keyboards need never be a problem - the system is completely voice oriented. Whilst nowhere near market readiness, BT's research signals a new and highly interesting direction for mobile computing. Can this be a sign of times to come?

Noting the future

But before your PC is strapped to your arm, dramatic changes in the realm of notebooks are on the way. If screens need to be bigger than the box to which the are attached, then the solution must be for foldable screens to hit the market sometime. Flexible plastics and innovative liquid crystals might bring these early next century. Prototypes are already in laboratories of various research organisations, including at Cambridge University, England. Other researchers are looking at holographic solutions which do not require a screen at all!

Predicting that processing power will, as always, increase is not revolutionary. But a quiet revolution may be caused by an extension to the concept of Plug and Play, an idea which really came to the fore with Windows 95. Within 1997, USB, a cheap, simple and highly effective expansion system will become the driving force behind notebook connectivity. This innovation will allow you to connect up to 127 devices to your system with the minimum of fuss. No longer will a lack of bay or slot limit your options.

So what's the next step in terms of format? It's difficult to tell, but the notebooks we use today will probably be as far removed from the notebooks of tomorrow as they are of ENIAC! As devices do ever more things, in ever smaller packages, it is not difficult to imagine notebooks, PDAs and other mobile connectivity being combined.

Imagination is your limit

Why just do your personal management on a PDA, why not email, phone, and fax? The Nokia 9000 launched at CeBIT 1996 is a device which demonstrates this direction. But these are early days and the devices that follow will more than likely replace more and more functions that you now have separate devices for. In a similar manner, computers such as Toshiba's Libretto will adopt more functions, perhaps GSM telephony and PDA-like capabilities. Doing this with the right mix of price and functionality is a challenge for the market.

As we get cleverer at designing circuitry and as newer technologies come along to create the basic building blocks of computing devices, would an alternative be discrete devices performing discrete tasks? Imagine devices distributed around your home or work place, connecting together as a single computing device. Research is even being done on building devices into your clothing! Imagine how much processing power the heel of your shoe could eventually contain!

Just as the computing industry has changed radically over the past fifty years, so it will in the future. Fifty years took us from ENIAC to the Libretto, from a $400,000 room-sized device capable of 0.03MIPS to a $2,000, 0.8 kg device 2,000 times more powerful. What will happen in the next fifty years? However wild your imagination, the only thing ever certain in computing is that predictions almost invariably turn out false - the researchers just keep on surprising us!

The big stuff

Outside the PC world, some really big machines have been developed. While some companies were concentrating on shrinking the size of a computer, others were only interested in power. Such mega-machines ruled the computing world. In the dim and distant past, beasts such as the Amdahl 470s, famed for power in the 1960s (Amdahl continues to make high-end machines for specialised corporate use today) and DEC's PDP 7 and 11 series defined computing. They were the forerunners of today's mainframes. And although many people expected the PC to be the death knell of the mainframe, mainframe sales are now looking up again. Clearly as more distributed computing takes place, mainframes once again have a place in the computing world.

The most esoteric of large machines is the supercomputer, and these machines, few and far between, still perform the duties that the early supercomputers were designed for. Intel has recently developed the first teraflop computer (a teraflop is equivalent to one trillion FLoating point OPerations per second). This machine is based on 9,200 Pentium Pro processors working in parallel, and was designed by Intel and the US Government's Department of Energy's Sandia National Laboratory. It is part of the Accelerated Strategic Computing Initiative (ASCI) and will be used for tracking violent weather systems, mapping the human genome, imitating environmental pollution clean ups and simulating the effects of advanced medicines as well as virtual nuclear weapons testing, now that real testing has been stopped.

Cray computers, famous for many years, and now part of Silicon Graphics Industries, were also famous for the quest for the teraflop computer. Back in the 80s, its machines were already so fast that they needed to be round, to ensure that the data had to take the shortest possible path - relativity and the speed of light were already a problem. Before its founder tragically died, the company was looking at many highly innovative technologies such as biological conducting paths laid down in organic chemical reactions. No doubt these avenues of research will bring many more highly innovative supercomputers onto the market in the future.