jaw@eos.UUCP (James A. Woods) (07/26/88)
a smattering of comments from an article in Computerworld, July 18, 1988. The protagonist: Bill Dunn, executive VP of Dow Jones & Co., a 1.4 billion dollar information services giant, who just spent a piddling $5.3 million on ... The machine: (two) Connection Machine(s), to supplement some IBM 3090s, 4300s, and DEC VAXen they'd rather junk. The initial impression: "It was a 12-sided hypercube -- I thought that was something you put in a drink," Dunn recalls. The problem: "We have this fantastic service," Dunn says, "a wonderful business amalgam of 70 billion characters of information on computers in networks, on packet switches, with terminals and personal computers. And if I walk up to this fantastic collection and say 'How is Boeing doing in its competition with Airbus?' I'll wait ... and wait ... and then I'll see 'Zero documents.' Then I play the machine's game. I type in 'Boeing' ... I get 8,763 documents. I try Boeing and Airbus ... 392 documents. And the meter is ticking at $2.40 per minute, and someone wants the answer and you're on the phone and sweating a bit. Then I say, 'The heck with it, it's not worth it.' Dunn is outspoken in his disdain for the unfathomable "Boolean" gibberish need to prompt today's computers. The solution: Thinking Machines Inc. relevance searching algorithm with feedback, programmed in C on a VAX frontend. or, as Bill puts it: "The Connection Machine is really an idiot. It runs like a son of a bitch, but it drools at the end of a 100-yard dash. You have to have another mechanism that wipes its mouth, pulls its pants up and takes it over to get the blue ribbon." The prize: Real-time access to six months of hundreds of publications like The Wall Street Journal, Fortune, Business Week, USA Today, etc. Viewing the investment as a "drop in the bucket" compared to $250 million the company spends just on printing presses, Dunn forsees the purchase of "many, many more" such machines. Summing up: "With Thinking Machines, we were the great white hope," Dunn says. "They were tired of blasting missiles out of the air or splicing genes in a mouse gonad. Here was something that could actually benefit mankind."