Was the Open Software Foundation’s choice of DEcorum for its Distributed Computing Environment rigged? Foundation chief David Tory, in answer to a pointed question from the press conference floor last week, told reporters the Foundation’s selection was not based on a political process whatsoever. Saying so, however, doesn’t remove the suspicions, and one of the reasons they linger is the sensation that the tide turned in favour of the DEcorum contingent months ago – despite the claim that a formal decision was reached only a couple of weeks before the announcement and that the contracts or letters of intent were only wrapped up in the last 12 hours before the press conference. As it was, at least a month before it publicised its decision, the Foundation was briefing special interest groups and selected outside consultants as to what the thing would look like. Moreover, Netwise, the Colorado start-up whose strategic RPC technology was effectively rejected by the decision, complained in writing to David Tory as early as last September that it had direct feedback from some Foundation members that the decision was wired for Network Computing System.
Compensate for weaknesses
The complaint noted that some DEC employees had made similar allegations to customers and expressed alarm that the group that wrote the evaluation criteria for the Request for Technology was composed of Network Computing System supporters or firms directly involved in the joint submission, and protested that the flow of data favoured member firms that, knowing what competitors were submitting, could tune their submissions accordingly to compensate for their weaknesses. In his replies, Tory wrote back that the Foundation’s Request process would be unbiased, was not wired for Apollo and that the credentials and integrity of the evaluation criteria were beyond reproach. He also added that it has been suggested that Netwise may be trying to protect itself by construing a fallacious claim of bias in the event that its submission is not selected for the DAE technology. I trust it is not. Netwise’s position, as of last week, was that technical merit was not the deciding factor in Foundation’s Distributed Computing solution, at least regarding RPC technology. The Foundation’s own evaluation, the company claims, revealed the Netwise product superior in many respects to Network Computing System. Netwise believes it may have been knocked out of the running by the vocal position it took over the terms and conditions the Foundation offers independent software vendors. Netwise officials said the Foundation told them it was annoyed by Netwise’s behaviour and that it was not helping its position in the Distributed Computing Environment selection. Although Netwise contends there are three active components in the Request for Technology process – technical, business and political – and wonders openly about the possibility of a connection between the way the Request went and the major infusion of cash the Foundation received from its sponsors at a pivotal moment in the DCE selection, it has so far found no smoking gun or direct evidence that the decision was politically motivated. However, it said, the decision was clearly not influenced by a desire to unify the Unix industry. Eric Schmidt, vice-president of the General Systems Group at Sun, the big loser to the DEcorum lot, told our sister paper Unigram.X last week, I hate to use the word rigged but it’s my impression the decision was made very early on. It’s cultural. Schmidt, who readily admits he is not part of the Foundation’s internal process, claimed the Distributed Computing Environment results showed a clear pattern on the Foundation’s part. –
By Maureen O’Gara
The sponsors have undue control, he said. Schmidt contends that the differences between Sun’s Open Network Computing/Network File System and what the Foundation chose do not justify the disunity and incompatibility that will be spawned. For him it is clearly just a result of politics and a concerted effort to undermine Sun’
s position. The powers that be at the Foundation, companies like IBM and DEC, Schmidt said, which have significant non-Unix business to protect and cultivate, want to slow Sun down as much as they can while they get their product line in shape. The apparent big machine bias of the Foundation’s Distributed Computing Environment technology and the way it seems to fail to integrate personal computers and 25m Novell NetWare nodes tightly – not to mention Banyan or 3Com – into its scheme begins to lend the notion some credibility again. Schmidt reads it as an interpretation of the workstation as an engineering and technical box but there are suspicions in some quarters that it could run deeper. There are rumours floating through the industry that the Foundation’s technical team was highly divided on which technology to chose and would have preferred a less divisive solution but that it was taken out of its hands. But Phil Shevrin, a vice-president at Locus Computing, whose technology was part of the DEcorum submission but did not make the final cut, worked closely with the folks from DEC and Hewlett and pooh-poohs the idea that the Distributed Computing Environment was rigged for the founders. He maintains that all along they were very concerned that they would lose out to Sun and Netwise. Peter Weinberger, AT&T Bell Labs Department head of research and one of the AT&T Unix Software Operation’s chief scientists, was in the unique position of being in the enemy camp as one of 10 consultants overseeing the Foundation’s selection, charged with keeping the submitters honest.
Inequities and biases
In his opinion, the process was not wired in the small. But while the Foundation’s technical people appeared open and fair, the process as a whole was not as open as the Foundation would have us believe. Like Netwise and for pretty much the same reasons, he is critical of how the Request was originally written and in fact notes that there are a lot of different models for a distributed environment. In fact he is not even convinced that what has been produced is actually the stuff of distributed computing. If the process is so open, there was a lot of technology that was overlooked, he contends, wondering why Xerox Corp for instance which had this kind of thing for 10 years didn’t make a submission, or for that matter why AT&T didn’t come forward with Remote File Sharing. For him, the whole process is not particularly designed for compromise and there was a heavy, if not undue, emphasis on short-term technical readiness. One of the reasons the DEcorum submission got an edge, he feels, might have been that it was simply an integrated consortium and meant less work for the Foundation. However, it may be faced with considerably more work than it bargained for since Weinberger, who is simply appalled by the amount of code the Distributed Computing Environment represents, reckons it will have serious implementation problems down the road. Like Schmidt, however, Weinberger reckons the cultural component played a very heavy hand in the decision – making it almost a foregone conclusion. Unlike another institution with much practice ironing out inequities and biases, he said, the Foundation’s procurement policies are not up to the standards the government uses, perhaps because it is a purveyor as well as a consumer.
