The Computer Industry

Motives and motivation

The interactions between universities and computing companies were beneficial for both. The universities' basic motive was straightforward: money. As a politician once said to me, "I know why you are here. Universities always want something for nothing." Over the years we bought computers at deep discount from Honeywell, Digital, Prime, IBM, Apple, Sun, Dell, and many more. We received gifts of millions of dollars of equipment, and cash grants of millions of dollars for specific projects.

What have the companies received in return? First there are sales. Universities are a major customer. IBM, Digital, and Apple were among the companies that cultivated this market with deep discounts and donations of equipment. In the early days of personal computers, IBM gave universities huge grants of equipment in the hope that the universities would standardize on the IBM PC and write leading edge software for it. But sales were not the only motivation. Many people in the computing industry are genuinely interested in education and are strong believers in technology for education.

Universities were early adopters of many types of computers and the first organizations to take networking seriously. As such, we were important for companies trying to introduce novel products. When Sun was a start-up, Carnegie Mellon was its biggest customer. Today it is hard to remember that the Apple Macintosh was a commercial flop when introduced in 1984. Without orders from a small number of universities it would probably have died within the first year.

Companies valued our insights. By briefing us on new ideas before they become products, companies could get early feedback. They presented their plans and we were not shy in telling them what we needed. When Bell Atlantic (now Verizon) released their first cell phone service, it was a multi-million dollar gamble. The vice president in charge told me that the only market research that he trusted was from an experimental installation at Carnegie Mellon. When NCR developed Wi-Fi networking, they came to us for an alpha test. When Steve Jobs showed his NeXT computer to a group of university leaders, we told him that universities would not buy a Unix workstation without the X window manager. He did not like it, but we needed it, and he believed us.

The companies also supported university research. Digital was particularly well organized in supporting external research. Most of their support was equipment grants, but for the right project they were willing to provide large sums of money. Intellectual property can be a problem with corporate research, but Digital allowed universities to keep the rights. Their benefits came from exchanges of ideas with researchers, and from leading edge software being developed on their systems. One of my major disappointments when at Carnegie Mellon was a large grant from Apple to port Unix and the entire Andrew environment to the Macintosh II, and to integrate it with the Macintosh user interface. This was brilliantly done, but when we came to deliver it, Apple had disbanded the engineering group that had sponsored us and the work was wasted.

Finally, companies are always looking for good recruits. Ph.D. and master students from the best universities are in enormous demand. Later in my career, at Cornell, I felt as though I was running a farm system for the Internet industry, teaching upper-level courses on information retrieval and software engineering, with teams of students doing independent research on an Hadoop cluster. Even in the depth of the recent recession companies were hungry for these graduates. Many times each year a student will ask my advice on which job offer to accept. One Carnegie Mellon student had to choose between a faculty position at Stanford and a central position at a fascinating start-up. He chose the start-up.

I think that the corporate people enjoyed visiting the universities. They would often ask to meet students, see demonstrations of research, or attend events on campus. A colleague at Penn State organized his requests for donations around the football season. Carnegie Mellon was at one time a part-owner of the Pittsburgh Pirates baseball team and we would entertain visitors in the owners' box. A visitor from Stanford University even went back to California and persuaded Apple to take a box at Candlestick Park.

Exchanges of information

Visits to the corporations were enjoyable and productive. We would sign a non-disclosure agreement, the company's engineers would brief us on their plans, and we would provide our feedback. Our local salesmen came with us and often learned facts about their companies' plans that were not usually known to the sales force. Each year a group of us from Carnegie Mellon went to California for a week, spending a day each at Sun, Apple, Hewlett Packard, NeXT, Adobe, and other companies. Other visits were to Digital's research labs outside Boston and the various IBM locations.

Several companies had formal university groups. The first meeting of the Apple University Consortium was a very special occasion. It was in San Jose in December 1983, before the official announcement of the Macintosh. We were given early information about future products and shown the famous 1984 Super Bowl commercial. The consortium later grew too large and lost its effectiveness, but IBM's university advisory group was always small. Each year it concentrated on a single topic. One year the topic was networking. IBM was throwing its main effort into OSI networking, while also contributing large sums to build the NSFnet, which became the backbone for the Internet. The IBM vice president for networking spent much of the day failing to convince us of the virtues of OSI and not understanding why we preferred the Internet.

Only a few universities had this privileged access to the companies. To reach a wider audience, the companies had booths at the EDUCOM conference and hospitably suites where we could talk informally. These grew increasingly lavish until, after consulting with his lawyer, the IBM head of university marketing sent a letter to his competitors, "Let's not have food fights. We should complete with our products and services, not with the lavishness of our hospitality." In contrast to this sensible attitude, I attended a couple of corporate seminars and was appalled by the lack of serious content and the extravagant hospitality. The worst was a so-called seminar by Bell Atlantic. It was a disgrace: minimal content, but expensive food, drink, golf, and sea fishing.

Salesmen

While at English Electric - Leo I had a first-hand look at how computer systems are sold and learned some useful lessons. One lesson was never to trust demonstrations. Multiprogramming was a major selling point for the System 4 computers but the operating system was incomplete. The company regularly ran demonstrations that showed several programs running at the same time. Card readers chattered away, magnetic tapes whirled, and the line printers churned out paper, but these demonstrations were faked. Actually these three programs were the only ones that could be run together.

A second lesson was that salesmen offer the configuration that will get the order, not the one that will do the job. We were asked to bid on a system for a company that ran off-track betting. They had difficult finding people to operate their telephone center on public holidays, when there are many horse race meetings. Consultants had recommended a computer system and provided a price estimate. Our salesman configured a bid based on that price and I was asked to write a section about performance for the proposal. My calculations showed that the configuration was hopelessly too small, but the salesman submitted the bid anyhow. Fortunately the contract went to a competitor, Control Data Corporation, who ran into all sorts of problems. Years later, another consultant solved the staffing problem by the simple expedient of moving the telephone center from central London to an area with high unemployment.

Years later, as a potential customer, we received a bid from Hewlett Packard that revolved around a suite of programs for business data processing. Not only was the memory in the configuration too little for the programs that were offered, the machine that was quoted could not be extended. All this information was carefully buried in the tender.

The most important lesson was always to know how the salesmen were rewarded. At Dartmouth we used Honeywell mainframes. Honeywell may have had a price list, but they paid their salesmen by the dollar amount that they booked. To buy a computer we first agreed on the price and then negotiated on what would be delivered. On one occasion, a salesman booked an option to buy a second machine and Honeywell recorded it as an order. Several years later a truck from Arkansas arrived at the Dartmouth computing center in New Hampshire with a large mainframe computer. Fortunately the operator on duty refused delivery. On another occasion Honeywell delivered the wrong central processor. On a happier note, we completed a complex negotiation with Honeywell just before Christmas one year. As soon as the contract was signed, but not before, the salesman delivered a Christmas present. It was a clock. Like many of Honeywell's products it was made in Japan.

The best time to buy equipment was the week before the company closed its books, as they were always trying to bolster their sales figures. As a computing director the most fundamental rule is that you stay within budget, but budgets are developed a long time before the funds are spent. Private universities have the flexibility to take advantage of opportunities. Twice at Dartmouth, when everybody else had left for the Christmas break, I sat in President Kemeny's office asking for extra funds to take advantage of a year-end offer. At most state universities every line in the budget is fixed and the universities do not have this flexibility.

If the Digital salesmen met their sales target they received rewards. One year our salesman came to me with a hard luck story. He was one computer short of having met his target for five years in a row. Please would I order a VAX system from him? I could cancel the order a week later. Naturally I refused, but he was an excellent salesman and I called his manager. He got his reward. As I remember it was a trip to Jamaica with his wife.

Aggressive marketing

The marketing strategy perfected by IBM for selling mainframe computers was highly centralized. The salesmen cultivated relationships with the computing directors and the customers' higher executives. In many organizations the purchasing of minicomputers and personal computers was also centralized, and the IBM method of selling directly to senior executives continued to be successful. In universities, however, as computing became decentralized, purchasing decisions were made by departments. The minicomputer salesmen understood this and developed relationships with the departments. As Barbara Morgan of the University of California at Berkeley quipped, "When the IBM salesman visits me he asks, 'What is happening at Berkeley?' When the Digital salesman visits me, I ask him, 'What is happening at Berkeley?'"

IBM had a reputation for bypassing the computing professionals and selling directly to the president of an organization. This was called an "end run". I could not understand why a chief executive would listen to somebody on a golf course rather than to the experts in his own organization, but it was very effective. When I worked for a strong president, such as John Kemeny at Dartmouth or Dick Cyert at Carnegie Mellon, this was no problem, but their successors were weaker. Kemeny's successor at Dartmouth had an IBM vice president as a buddy who caused enormous difficulties. He appeared to be ashamed that we did business with less prestigious companies such as Digital and Apple. Cyert's successor cost Carnegie Mellon a large sum of money by throwing out a carefully negotiated contract with NEC for a telephone switch because of an end run by the president of Bell Atlantic. He was a Carnegie Mellon alumnus who hinted that there might be grants to the university.

IBM and Digital were aggressive but they were good companies to do business with. The telecommunications companies were frequently dishonorable. The computing companies honored the spirit of an agreement, even if it was only a manager's letter, but the telecommunications companies were always trying to wriggle out of commitments.