Interview - Ready for another ride

Sir Clive Sinclair is getting ready to re-enter the computer market. This time he's dreaming of cheap Linux PCs using parallel processors.

Sir Clive Sinclair doesn't suffer fools gladly and that includesket. This time he's dreaming of cheap Linux PCs using parallel processors. Intel's processor developers. He's not keen on customer choice being stifled and condemns over-priced, over-powered machines. That said, he's not accusing the rest of the IT giants of being incompetent.

It is with such strong opinions that the man who ignited the personal computer market with sub-£100 machines in the early 1980s is planning a return to the business. He claims he can bring out a portable machine within two years that will be less than half the price of whatever is on the market by then. And it will deliver the performance that corporates and consumers want.

The processor is key. Sinclair has very strong ideas about how chips should have evolved, and although it is doubtful that Intel will be called to account for not doing things his way, he has long thought that the chip giant has dragged processor development, and the desktop PC, off in a wrong, power-wasting direction. This view, combined with the appearance of Linux and leaps in display technology, has spurred him to consider re-entering the computer business.

He sees his return, or the arrival of someone else with a similar vision, as the salvation that the market needs. Sinclair is concerned by the dominance of Microsoft Windows and Intel processors. He hopes that the US Department of Justice case against Microsoft and its control of the software market will lead to PC suppliers unbundling Microsoft's products from their systems and allowing customers to choose software. Of course, this still leaves Intel processors dominant. He said the lack of consumer choice bothers him, and the overblown, memory-sapping software, combined with over-powered and over-priced processors, is a problem.

"It needs something like a dedicated Linux machine to break the mould," he said. "I think the situation is frightening. The manufacturers should be forced to unbundle. People shouldn't be obliged to pay for having Microsoft software. There ought to be a choice - one price for Microsoft and one price for Linux.

"Linux looks like one way in - a Trojan horse. Apparently, it's a good operating system and a lot of software suppliers are supporting it. They wouldn't do that if they didn't have a lot of confidence in it. I think it will be very interesting to do a Linux machine. The standard PC is expensive because of the Intel chip. It is elaborate and and consumes a large amount of power. The software is also very demanding of memory."

The machine Sinclair has in mind could be considered as a reworking of the Z88, the last computer he developed. The portable Z88, released in 1988, did not achieve the sales its inventor had dreamed of, but he's still fond of its concept. The technology he's now waiting for will give him the chance to re-visit the issues that made the Z88 fail: "It wasn't the success I'd hoped for, partly because of the limitations of display and because it was completely non-standard. That's still a possible route to take if it's good enough, but if you can use an OS that's out there, then at least you've got an audience that is familiar with it. Linux looks as if it might be the one."

Although Sinclair's well disposed toward Linux, he won't rule out other OSs. "Psion's is well-known and successful," he said.

Researching the problems

Sinclair's approach to researching news project is exacting. He enthusiastically hunts down the solutions to the problems he thinks exist.

He demands precision and accuracy in technology development. His knowledge of electronics lets him know what is and isn't possible. If you've not done or seen things his way, be ready to justify why not. If you tell him something new, you'd better give him the whole story.

Before Sinclair created the ZX80 and ZX81 home computers, he wrote about complicated self-build machines or the more pricey systems from Tandy and Apple. Sinclair's budget computers transformed the market, giving rise to a generation of people who were introduced to programming, touch-insensitive keyboards, and 16K RAM pack upgrades.

Sinclair has been looking at the market with a view to using current and up-and-coming technologies to create a low-cost alternative to Wintel machines. A suitable OS and display technology, combined with a low-priced, powerful processor, and he'd be in business. The ARM chip, forecast to be used in 70 per cent of all cellular phones produced next year, is an example of the kind of processor Sinclair hopes will smash Wintel machine prices. "ARM is an option as it's low-cost with a high performance. Processors are always coming along, but it looks attractive."

Sinclair was not immediately converted to Linux when enthusiasts started spreading the word. He gave industry analysts a grilling and has been evaluating it since. He won't commit to saying the Linux OS is definitely going to feature in his machine, but it's certainly on his mind.

Not subscribing to the retail scene

Sinclair will sell his new device by mail order, the way he's launched everything that he's invented, from the ZX80 to the Zeta bicycle motor and his miniature radio. He said his inventions create their own market and that isn't necessarily the kind of product range retailers want to stock.

However, he doesn't agree that it is retailers who are keeping the Microsoft/Intel power base strong.

"The retailers don't have much choice, they just sell what's provided. They don't determine the product. It's a Wintel-defined product and all computer-makers make clones of them. They don't give a hoot about the design, they just sell what's there. They don't know what's possible, what's not possible. They haven't got a clue.

"Intel is trying to keep people using expensive and complex processors. It's what they supply and it makes them money. It's a shame, but you can't blame them."

But what he can blame Intel for, he said, is leading computer development down a single processor design path. "Years ago, Sinclair Research was looking at parallel processing machines and it's the way things should have gone.

"The whole business of having one chunk of silicon as a processor and other great chunks of silicon as the memory is a desperately inefficient use of silicon. The memory and processing ought to be merged.

Instead of having one processor here, and having your memory there, with loads of wires connecting them and slowing everything down, you've got one piece of silicon. All over that piece of silicon, you've got blocks of processors and blocks of memory.

"Not only would they be faster, because they're all on the same piece of silicon, but there's 100 of them, so you've probably raised the processing speed of the machine 200 to 300 times," he said.

This would offer amazing performance for speed-absorbing problems such as speech input and complex display generation in real time, said Sinclair.

He added that Intel knows all about parallel processing machines because it produces them. He doesn't believe the direction it has chosen is a conspiracy to hold computing back just to make more money, but he's annoyed by it. "God knows what Intel is playing at. It's not a conspiracy, just incompetence."

Backtracking slightly, he said that he thinks Intel is making too much money from the way it is doing things and that it is going to take some external player to make it change its ways. The challenge might come from the games market and developments that console manufacturers have made with machines that can handle complex graphics in real time.

"Sony Playstation II is going to shake people up because the performance is so striking. When you've got a Playstation II, which makes a Pentium III look pathetic, people will say: 'This games machine makes my computer look weak. What's happening?' And that's just Playstation II, which is not really pushing boundaries.

"Because the games market is so huge, somebody could design the sort of silicon I'm talking about - multiprocessor silicon that will blow your socks off."

Sinclair is confident his machine will undercut the market when it arrives for these very reasons, and the manufacturers will be unable to chase his pricing because they're too locked into the Wintel way of doing things.

"Their costs are tied. The reason the machine I propose will be cheaper is because it will use less memory, use a much lower-cost processor, much simpler power supply and a lower-cost operating system.

"It will be lower in cost because of the fundamentals. People who make computers now work on very narrow margins, so they can't cut their prices without going out of business."

An academic approach

Sinclair's interest in creating a new computer appears academic as well as commercial. His knowledge of what Microsoft and Intel have created between them is probably based on exhaustive research as opposed to first-hand experience. He doesn't often bother using computers himself. "Computers are very frustrating. They drive me round the bend - they're such awful machines."

He laughs at this. He knows what computers can do - and what they would allow him to do - but it's still not enough for him. All his design work is mathematical, so he uses a calculator. He said it assists him with his sums far better than computers can and has no interest or use for the graphical display a PC would give him. If a computer needs to be used, he'll get someone else to do the work for him. His own creations haven't been quite so annoying, though. "They were nice and easy to use, but they were really only machines to learn computers on."