The Silicon Jungle by David H. Rothman

questions to ask to find out the computer crew’s true attitudes about micros. And come up with the right answers of your own if the mainframers feel threatened by your interest in do-it-yourself computing. Good data-processing people, however, won’t be frightened—quite the reverse. The scary aspect of Frank Stewart’s consultant fiascoes is that he and his people were hardly lost amid high tech. “Some of us were inventors,” Stewart said, and his vice-president, Bob Hillard, even had a smattering of the FORTRAN programming language. Stewart’s firm didn’t rush blindly into computerization. Originally, it had been hand culling articles from 140 periodicals and 16 major daily newspapers. “We sat down,” Hillard recalled, “and worked out a mathematical formula to give weight to a particular article by mention of certain words, certain individuals, certain companies.” Existing data bases like The Source and Dialog just wouldn’t do. You couldn’t easily use them to search out the political patterns affecting the fortunes of Stewart’s clients. Stewart spent $31,000 on a computer and soon was suffering defective circuits and junky software. He paid a secretary to cram almost a million words of newspaper articles into the machine’s hard-disk memory, but it still couldn’t fetch facts as easily as he’d hoped. Then the data-base program, the one he was using to store facts and juggle them around, crashed—apparently overwhelmed by the volume of data. Micro data bases at the time, the early 1980s, couldn’t easily handle big hunks of text. Stewart couldn’t retrieve whole articles. It was time for a consultant. A good consultant (the “Who”) theoretically could write or install a new program (the “What”) in an effective way (the “How”), justify his actions (the “Why”), and ideally do everything at a reasonable price (the “How much”) before the data base grew too unwieldy (the “When”). Stewart wanted him working on his own premises (the “Where”) as much as possible so he could supervise him. “Our first consultant lasted all of four days,” Stewart said. “He was abrasive, nasty, and didn’t know what he was doing. One of our staffers was so frustrated trying to work with him that she wept. We’d hired him because he was a friend of a friend and said, ‘I’m a computer programmer. I can fix it for you.’ We never quite could pin him down about his background. We were so burned that for five or six months we put off hiring another consultant. “And when we did get one,” said Stewart, “we wanted a solid, established firm.” So he went to a government contractor headquartered in an imposing building, the size of a small public library. The consultants even owned the same-model computer that Stewart’s company did. Unfortunately, however, they were using different software for much different purposes. Stewart’s needs were new to them. “Look at MDBS,” Hillard recalls telling the contractor. “It really looks like a good package.” But he says the consultants told him they hadn’t even heard of it. That in itself might have been a tipoff that Stewart had hired the wrong people for the job. MDBS stood for Micro Data Base Systems. Ads and reviews about it had appeared in _Byte_ and other prominent publications of the micro world. With the clock ticking away at up to $35 an hour, the consultants searched for data-base software meeting Stewart’s detailed specifications. “Three weeks later,” Hillard said, “they handed me a bill for $2,500 and said, ‘The package you need to use is MDBS.’ I said, ‘Where is the work, where are the hours in relationship to the $2,500 bill?’ It went around in circles.” Stewart and Hillard must have remembered the old saying that a consultant borrows your watch to tell you the time of day and then charges you for it. According to Stewart, in fact, the consultants based most of the report on material his company had given them. By think-tank standards in his area, Stewart may have gotten a bargain. Dealing with consultants new to micro data bases, however, he was in effect paying their tuition. “The federal government must like them immensely,” Stewart said, “because they do good, documented work in the fields they have expertise in. But a small businessman had no business using them. After two and one-half months we had nothing but recommendations on the data base—and a stack of bills”—$2,500 high. But there was hope. Stewart’s corporate benefactors were sticking with him. If he could get his data base running, he could easily recoup his losses, offering clients a treasure trove of information about government policy-making affecting them. Although Stewart thought of himself as a truth seeker, his vice-president was bluntly a money man. “We would have sold our service for a million a year,” Hillard said, “and we would have netted $600,000. We were using four people to feed the articles into the computer, but that would have changed, because optical readers eventually could scan magazine articles and newsprint.” So now Hillard shopped around for a new savior to unravel the complexities of the MDBS software that he had settled on. He found three possibilities. One consultant, a micro expert, was snowed under with work, and another man had only mainframe experience. So Hillard and Stewart chose Fred Brown, a retired military officer who had worked not only on big computers but also little Radio Shacks and TeleVideos. In fact, Brown came recommended by the manager of a local computer store. The manager leveled with Stewart. He said he and Brown were together in a partnership developing software for lawyers’ offices. Hillard asked only two people about Brown. And both references came from the computer-store manager with the business relationship with the consultant. Inquiring about Brown, however, Hillard asked some good questions. Did Brown document the software properly? Did he, in other words, tell how to use or change it? Did his software work? Did he meet deadlines? Did he live up to all the specifications in his contracts? “I got affirmative answers to most of these questions,” Hillard said. One reference said a software project of Brown’s hadn’t come in on time but brushed off the problem as a hardware one. Unfortunately, Hillard didn’t ask _all_ the right questions. Brown had never before worked with the brand of computer that Stewart Research was using; nor was he fully comfortable at the time with the exact kind of software used at the company. Brown, however, struck Hillard as “a very nice professional.” He wore three-piece suits, lived in a $140,000 home with several computers in the basement, and held a masters degree from an Ivy League school. “No problem,” Brown told Hillard. “Listen, I charge——” He gave an hourly figure to the nearest cent; $30.02, we’ll say. “What?” asked Hillard. “Why 0._02_ an hour?” “I like to be fair about my salary, and that gives me $30,000 a year.” “Okay,” said Hillard, “that’s cute.” Brown wasn’t about to apply for welfare. On top of the $30,000 for consulting part of the time, he was enjoying income from sales of computer hardware. In short, he came across as successful and honest. And he may have been. Just the same, client and consultant were clearly mismatched. Their downfall was a familiar kind—software problems. You could customize the MDBS program in FORTRAN or BASIC. The earlier consultants, the ones with the large government contractor, had suggested FORTRAN. Brown himself was more familiar with that than with BASIC. And Stewart was convinced that FORTRAN would make the best use of software already purchased. So Brown went ahead with it despite three strikes against him: