Perhaps the most baffling aspect of the events that led up to the creation of the Open Software Foundation last week was how AT&T ever allowed itself to get into a position where three of the biggest computer industry heavyweights were prepared to sink their differences and share a platform to denounce it. And the answer seems to relate directly to the personalities involved, most notably Vittorio Cassoni, the former head of AT&T Data Systems, now back at Olivetti. Until the start of this year, AT&T made new releases of Unix available to every manufacturer that wanted them at the same time. True they were installed first on AT&T’s 3B line first, but AT&T was seen as such a pussycat in the market, and the 3B as such a dog of a product that nobody minded that very much. But in October last year, Cassoni had come to the conclusion that AT&T’s computer strategy needed a complete rethink if the company was survive as a serious player. Pussycat The parent’s research arm, Bell Laboratories, had its own RISC microprocessor under development, but Sun Microsystems had already announced the Sparc, got a string of chipmakers to produce it in a range of performances, and was offering it in a product, the Sun-4. Cassoni decided that the time it would take for Bell Labs and AT&T Data to reinvent the wheel was simply too long, and cut his deal with Sun, committing to build the successor machines to the 3B line around the Sparc. And, with what at first sight might have been a logical move, if naive, that meant that – as it had done with the 3B – AT&T would in future be putting new releases of Unix first on the Sparc – after all, Sun was doing most of the Unix development work anyway. But if competitors complaisantly saw AT&T and the 3B as a pussycat, they saw Sun, and with good reason, as a tiger. At a stroke, Sun would be in an intolerably privileged position vis-a’-vis its rivals, and the other four biggest players in the workstation market, Apollo Computer, Hewlett-Packard Co, DEC and IBM, not surprisingly exploded. Because Sun Microsystems was not reticent about its good fortune: on the contrary, say competitors, in February, after AT&T had agreed to cement their relationship with phased cash injections, the company’s representatives started telling them that if they wanted to get the latest release of Unix, they would have to sign to use the Sparc in future products. Xerox Corp had already signed for the Sparc and Unisys Corp and ICL agreed to do so as well. But that was as far as Sun got, and at the Uniforum trade show in February, AT&T luminary Jack Scanlon was confronted by a furious mob of the great and the good in the commercial Unix world, all demanding to see Cassoni for a showdown. They confronted Scanlon with the demands they had received from Sun, and Scanlon flatly denied that any such policy with regard to the Sparc was in force – and was handed copies of Sun promotional material containing the offending demand. They went away thoroughly dissatisfied, and the seeds of the Open Software Foundation were sown. The scene shifts to Bell Labs. The chip designers there, well down the track with development of their own RISC, the CRISP, were not surprisingly a little miffed to find that their efforts had been called into question and that the computer upstarts were instead betting the shop on a RISC bought in from outside the house. Miffed enough indeed that a report was commissioned from Dataquest for a detailed comparison of their part with the Sparc and the other then leading RISCs on the market: the Intergraph Clipper and the unnamed one developed by MIPS Computer Inc; one or two others may also have been included in the study, that is not clear, but at all events, the lads and lasses at Bell Labs got what they wanted: compared feature by feature with the other parts, the Sparc on balance is understood to have come bottom of the tree. Far from being a monolithic whole, AT&T these days is a collection of now very disparate businesses, but whatever it may be seeking to become, it is still at bottom what it always has been – a
telephone company, pure and simple. And the tecchies at Bell Labs weren’t the only ones to be offended at the snub delivered to the company’s own RISC development: it may be a little galling to have to get your personal computers from an upstart Italian, but personal computers are hardly leading edge technology, and for AT&T to have designed its own copy of the IBM machine would have simply been stubbornly reinventing the wheel. But AT&T justifiably prides itself on its own technology, and here was the computer division that was meant to be a key part of the future of the company saying that in future AT&T’s leading edge computers were going to be based on an alien technology, a technology, moreover that an independent consultant seemed to be saying was inferior to its own. The AT&T telephone chiefs put their collective foot down: Cassoni and Data Systems could do what they liked, but where AT&T computers were needed in telephone systems and products – PABXs for example – they were damn well going to be true AT&T computers, not hybrids based on a design by some wet-behind-the-ears start-up firm that sounded like it was Japanese out on the coast. Jetting While all this was going on within the company, Carlo de Benedetti was jetting back and forth across the Atlantic trying to persuade AT&T to put more money into Olivetti, and then changing his mind when AT&T said fine, but more cash means management control, and as his need for cash became less pressing. And only a few days after the partners had declared that there were significant differences between them, Cassoni was on his way back to Ivrea to take over as managing director of Olivetti. There has been much public comment that the loss of Cassoni was a body blow to AT&T Data Systems, and the appointment of an accountant in his place only accentuated the perception that AT&T Data’s days were numbered. And Cassoni was certainly responsible for giving a totally demoralised company a sense of direction and a new belief in itself, he did stem to a considerable extent the flood of red ink from Data Systems, and was able to get his own dedicated sales force where his predecessors had failed. But there is also a widespread feeling in the company that Cassoni’s agreement with Sun has left AT&T in a well-nigh impossible position, and Robert Kavner’s first task, before he can get to grips with running the business, has to be to try to repair the damage he has inherited. That the plotters went ahead with their plan for the Open Software Foundation is a very bad sign, but otherwise he has made an unexpectedly good first impression, and his statement late last week that AT&T would consider handing over Unix development and licensing to an independent body once it is strong enough to stand on its own in a couple of years perhaps – is encouraging. But in the meantime, the battle he faces for the soul of Unix looks forbiddingly daunting.