Non-IBM personal computers were available as early as the mid-1970s, first as do-it-yourself kits and then as off-the-shelf products. They offered a few applications but none that justified widespread use.
Drawing on its pioneering SCAMP (Special Computer, APL Machine Portable) prototype of 1973, IBM’s General Systems Division announced the IBM 5100 Portable Computer in September 1975. Weighing approximately 50 pounds, the 5100 desktop computer was comparable to the IBM 1130 in storage capacity and performance but almost as small and easy to use as an IBM Selectric Typewriter. It was followed by similar small computers such as the IBM 5110 and 5120.
IBM’s own Personal Computer (IBM 5150) was introduced in August 1981, only a year after corporate executives gave the go-ahead to Bill Lowe, the lab director in the company’s Boca Raton, Fla., facilities. He set up a task force that developed the proposal for the first IBM PC. Early studies had concluded that there were not enough applications to justify acceptance on a broad basis and the task force was fighting the idea that things couldn’t be done quickly in IBM. One analyst was quoted as saying that “IBM bringing out a personal computer would be like teaching an elephant to tap dance.” During a meeting with top executives in New York, Lowe claimed his group could develop a small, new computer within a year. The response: “You’re on. Come back in two weeks with a proposal.”
Lowe picked a group of 12 strategists who worked around the clock to hammer out a plan for hardware, software, manufacturing setup and sales strategy. It was so well-conceived that the basic strategy remained unaltered throughout the product cycle.
Don Estridge, acting lab director at the time, volunteered to head the project. Joe Bauman, plant manager for the Boca Raton site, offered manufacturing help. Mel Hallerman, who was working on the IBM Series/1, stepped forward with his software knowledge and was brought in as chief programmer. And so it went. As word spread about what was going on, talent and expertise were drawn in.
Estridge decided early that to be successful and to meet deadlines, the group had to stick to the plan: using tested vendor technology; a standardized, one-model product; open architecture; and outside sales channels for quick consumer market saturation.
About a dozen people made up the first development team, recalls Dave Bradley, who wrote the interface code for the new product. “For a month, we met every morning to hash out what it was this machine had to do and then in the afternoons worked on the morning’s decisions. We started to build a prototype to take — by the end of the year — to a then little-known company called Microsoft.” The team beat that deadline. The engineers were virtually finished with the machine by April 1981, when the manufacturing team took over.
The manufacturing strategy was to simplify everything, devise a sound plan and not deviate. There was not time to develop and test all components. So they shopped for completely functioning and pretested subassemblies, put them together and tested the final product. Zero defects was part of the plan. [ … ]