The Open Software Foundation is a deal its founders would have avoided if they could. It’s not something they particularly wanted to do – even if it is the biggest thing to have hit the computer industry in a dog’s age. It’s something they were simply driven to – by AT&T and its unholy alliance with Sun. Take Sun out of the equation and the Foundation simply wouldn’t exist today. No Sun, no Foundation. Even with Sun in the picture, the Foundation still might not exist today if AT&T hadn’t been so intractable – if it had been willing to address the issues so worrisome to Foundation’s founders. But it wasn’t. One of the things that chafed and troubled Foundation folks was the report that Sun was claiming that its alliance with AT&T meant that if vendors wanted the latest version of Unix, they had to go Sparc. Provocation According to industry sources, such claims fired the most heated of the confrontations between AT&T and the Hamilton Group back in late January, early February when the Group managed to get its hands on Sun documents that said as much (CI No 934). At least some of those documents happened to be copies of a Sun slide presentation given to a bunch of Japanese companies in which Sun reportedly maintained that buying the Sun 4 was the only way to get the latest Unix.
Bull wasn’t in the Hamilton Group of Unix dissidents, and Georges Lepicard, its man on the Foundation board, wasn’t party to that particular exchange but he says he himself heard Sun make such statements in other presentations. Despite the provocation, however, Lepicard claims AT&T could still have mollified the Hamilton Group and headed off Foundation’s formation if it had been willing to make one very little concession and warrantee that Unix would not become architecture-dependent. Such a committment was not forthcoming. How could it be, it wasn’t part of AT&T strategy. When the Hamilton Group took its six-point list of complaints – all of them directly attributable to Sun to AT&T’s erstwhile computer boss, Vittorio Cassoni, back to January 29 in New York, his response was, it’s said, uncompromisingly hardline. It was AT&T’s turn to redistribute the wealth, he told them, and none of the Hamilton Group, least of all its European hangers-on, was ever gonna get a piece of the Menlo Park process, an allusion to the California lab where AT&T and Sun would develop Unix – with the Sparc chip in mind. Like many a politician these days, AT&T is now trying to backpeddle on Cassoni’s January admissions and pretty up its public image by flatly denying the contention that Unix is being tailored to the Sparc. Janet Davis, marketing communications manager of AT&T Unix Europe, the Unix licensing arm this side of the water, who spoke last month in the absence of her vacationing boss Bill Elliot, said she thought those old claims had been put to bed by now. It’s not true. The 3B is our porting base. But the original heady insistence on Sun’s privileged status didn’t escape Nixdorf’s notice. It was stated publicly that Sun would have the competitive edge, when Sun and AT&T teamed to co develop Unix, declares Foundation board member and president of US-based Nixdorf Computer Engineering Corp Bernhard Wobker. He says that a joint AT&T-Sun announcement let it be known that they will provide users with binary-level – ‘binary’, he stresses transport, or object code compatible within the computer architecture. The Sparc bias was also maintained elsewhere in print, Wobker points out, quoting from an article that appeared in the Winter 1987 issue of Sun Technology, a Sun-produced quarterly that to his dismay is sent out to end users. In it Sun’s executive vice-president Bernard Lacroute wrote that as part of three-phase development programme with Sun, AT&T had agreed to adopt Sparc as the RISC architecture for which Unix will be tuned. To Nixdorf, AT&T’s Sparc bias was doubly unfortunate. The Paderborners had already done a detailed evaluation on Sparc, Wobker said, and decided against using it because there were other far superior architectures ava
ilable. Albrecht Doehler, executive director of Siemens AG and its representative at the first Foundation board meeting, says AT&T refused to discuss Sun in one-on-one talks with his company. As he put it, They tried to tranquillise us… telling us there was no need to discuss Sun… no need… Sun was just there… a deal they had and that was that… Herr Doehler is a good natured obliging man, full of Teutonic calm and patient explanation, straight out of the unemotional we’re-all-hard headed-businessmen-here school. On this one point his words tend to tumble out in an excited rush, so quick it’s hard to take them all down. He doesn’t intend identifying exactly who they were, calling them only AT&T’s highest-ranking decision-makers. Like Nixdorf, Philips and Bull, the Foundation’s other European members, Siemens was one of Europe’s Unix pioneers and has installed 20,000 systems and 100,000 workstations, which probably makes it the number one European Unix supplier. What troubled Siemens from the beginning was what was troubling its customers: AT&T’s deal with Sun meant Siemens would get any Unix enhancements or developments 12 to 18 months after Sun. Sun’s treatment as a first-born son wasn’t the only instance of AT&T favoritism. According to Wobker, there were a few other companies, whose names he wouldn’t mention (but presumably he’s at least thinking of Unisys and Motorola), with whom AT&T struck up early exchange agreements. No real influence They would get preliminary versions of future Unix releases so they could co-ordinate their announcements. Anybody who wasn’t part of this select coterie, Nixdorf came to feel, would have no real influence on AT&T’s development plans. And if that wasn’t enough, it became clear to Nixdorf that the Unix development programme, as conceived by AT&T, didn’t tackle some (crucial) areas of applications compatibility. There was still an awful lot left to do to make the system viable, Wobker said. There were long-standing problems with virtual memory and security and the unreliable (Unix) file system. Some of these problems may have already been solved in work done by Hewlett-Packard or IBM valuable extensions that are available today. But if they went along with AT&T, the bulk of Unix licensees would be saddled with an inferior operating system because with these kind of extensions it wouldn’t be AT&T’s Sun iteration.
By the time the outside world was getting its first glimmer of the historic grumbling going on inside the US computer industry, the Europeans were a disaffected lot, ripe for insurgency. In teaming with Sun, Wobker says flatly, AT&T showed they don’t have a good strategy… They should have known when they entered the relationship with Sun it would upset other firms. Moreover, it’s perfectly true what they say about Cassoni alienating his Unix licensees and laying the seeds of Foundation, Lepicard says. AT&T’s strategy and tactics were most ill-considered.
