To celebrate the 700th birthday of the City of Berlin, Nixdorf Computer AG took a party to visit its operations in the modern equivalent of the mediaeval Walled City, and took the opportunity to spell out some of the company’s research and development aims and objectives. Mike Faden was there to listen. A discussion by Nixdorf board member Hartmut Fetzer on the perils and potential rewards of research and development – which included more than a snippet of news on the company’s activities – enlivened a rather lacklustre technology forum held by the company in Berlin last week (organised with tremendous precision, pity about some of the tired presentations). Fetzer provided a Nixdorf’s eye view on the company’s involvement with ill-fated Auragen, Pyramid Technology and other suppliers in the course of justifying some of the company’s more superficially surprising decisions on how to use its research and development cash – which amounted to DM426m, that’s about $240m, some 9.6% of revenues in 1986. Offer disks OEM And talking of surprising attitudes, what about the decision to go into disk drive manufacturing? At a time when many disk drive suppliers are going to the wall and prices are still falling, Nixdorf is ploughing cash into producing a range of slim-line 80Mb and 170Mb drives for its own consumption – and may even offer them OEM in future. Even given the foresight of late lamented founder Heinz Nixdorf – whose influence after death was emphasised by the number of times his name came up during the day’s presentations – it seems a surprising decision. But Fetzer took pains to justify it on two counts. First, economically it makes sense, he claimed – the proportion of the total cost of a system accounted for by disk drives is high enough to justify having the technology under your own control. Contrast that with semiconductors, the total cost of which in a computer system is equivalent to some 3% of the revenues derived from selling the system. Nixdorf had done short run chip production, but now just designs key parts and farms the manufacturing out, Ferranti Plc being a major supplier of custom chips to the company to Nixdorf’s design. Second, the highly demanding technology involved in disk drive manufacture – he compared the sophisticated mechanics involved in controlling read/write heads to flying aJumbo jet 3mm off the ground – would produce spin-offs in other areas. And if that sounds like a dubious assertion, Fetzer was ready with a nice example of the benefits. One result of the development was a very small drive motor – and the retail systems developers just happened to be racking their brains for a way to adapt a new range of under-the-counter bar code scanners to be small enough to to allow the person operating the till to get their knees under it. The disk drive motor, he claimed, turned out to be just the thing – and the scanner is now the basis of Nixdorf’s new range. Another rather less cheerful tale – but one that may yet turn out to have a happy ending for the company – concerned the Targon/32, based on the fault-tolerant Unix technology originally devised by US company Auragen Systems Corp in New Jersey, where Nixdorf invested some $3m or so. According to Fetzer, the impossibility of judging whether software will work as promised – in contrast to the relative ease of judging the state of hardware developments – meant that Nixdorf believed the future delivery dates being put out by Auragen, and rather rashly passed them on to its customers. The result, when Auragen finally threw in the towel, was that Nixdorf ended up developing the thing itself according to Auragen’s principles, which made the Targon/32 horribly late. And the happy ending? The 68000 family Targon/32 is at last starting to ship – and was the basis of that huge contract shared with rival Siemens to supply systems to unemployment offices throughout Germany. The sad state of software development generally, in fact, was an area where Fetzer didn’t have any firm answers – or at least he wasn’t talking about them. Some 4,000 programme
rs, he said, are employed writing applications within Nixdorf’s sales divisions in Europe alone. In software development, he pointed out, we try to be very deterministic – the opposite of the changing requirements of real life systems. He could only comment on the need for developing more flexible software – Nixdorf’s key contribution in this area is the Comet business suite, which is designed so that the salesman can take the user’s order and come up with what looks like a custom package in just a couple of days. Gain time Fetzer also attempted to deal with what appeared to be a rather sore point when he emphasised that the primary reason for basing systems on other people’s technology – such as the Pyramid Technology-based Targon/35 – was not, contrary to common thinking, because Nixdorf doesn’t have the development resources to go it alone, but rather to gain time in bringing products to market. Using Pyramid’s products, he said, saved the company two years’ development that would have been needed if it had designed its own high-end Unix minicomputer – but typically, the development cost of bringing a system to market initially is only 10% of the expenditure on development over the life of the product line. Fetzer also outlined three areas where Nixdorf has been expanding its research efforts – communications, following Heinz Nixdorf’s wish to extend the company into a major communications supplier; office automation, and retail systems. Although the majority of the company’s research and development efforts are still based in Germany, he claimed that there were some 240 developers in the US, a small group in Ireland and a hardware group due to be set up in Singapore where Nixdorf already manufactures and does software development. The company also operates technology centres in Santa Clara, California and in Tokyo, whose brief includes scouring local markets for likely emerging technologies that the parent firm might license.
