build an instruction network using Plato and its large 6600 computers. The Plato system used the first multimedia terminals, with computer interaction, slides, voice, and audio output. Although Plato has been successful with thousands of courses and millions of course hours in university training and applications, including teaching basic skills to prisoners, it and newer PC-based CAI programs have yet to deliver the promised revolution. Clearly, improving education is an important goal of all countries. Perhaps the CAI revolution awaits the revolution in ubiquitous, zero-cost, multimedia capability foreseen by some for the 1990s.

If Plato were located in Silicon Valley, some company would no doubt start up to develop a low-cost computer platform to utilize the vast array of courses. Corvus, located in the valley, is still waiting for the market surge. Apple serves the market, albeit in an ad hoc fashion. By making computers fun-and-game-oriented, Nintendo may have found the true pathway into the market.

A classic marketing flaw is to attack a large company's customer base with a competitive replacement product. Rarely is this approach successful, since customers would prefer to buy from a few suppliers that are also the leaders. The new product typically attacks a strongly held market by using a different or incrementally improved next-generation technology. Existing suppliers, particularly start-ups, are unwilling to give up their market position and can hold their share of the market by enhancing their products through evolution. Attacking the customer base (e.g., IBM, Lotus 1-2-3 clone) of a supplier that is unwilling to accept the loss of revenue (e.g., add-on disk memories) or loss of control (e.g., database) is a flawed approach. It will succeed only if the new technology is compelling and the competitor cannot move prices rapidly (e.g., plug-compatible IBM mainframes).

Autodesk successfully attacked the mechanical computer-aided design (MCAD) market by building a product that ran on a PC as opposed to the older minicomputer (e.g., Computer vision). It succeeded because the established companies neither saw a threat nor were able to lower their margins, since they had fixed costs and fixed ways of operating based on selling a few, expensive software packages.

Two major industries have formed through efforts to attack a large company’s customer base: plug-compatible computers (pioneered by Amdahl) and disks (pioneered by IBM alumni). Both industries originated at a time when IBM had extraordinarily high profit margins on computers and peripherals. Al Shugart, disk pioneer and IBM alumnus, described the opportunity as follows: "IBM's high profit was immoral. Any self-respecting engineer would start a company just to bring lower-cost products to the mass market."

However, by aggressive pricing and by increasing the complexity of the disk subsystems, the established firms have decreased the significance of the plug-