This is the rock-solid principle on which the whole of (IBM’s) galaxy-wide success is founded. Their fundamental design flaws are completely hidden by their superficial design flaws.
Lowe’s IBU task was to come up with products to move the brand into new sectors, to be a nursery for new mainframe opportunities.
The PC project sat outside the normal IBM corporate ethos. It was done in such a hurry, managed in less than a year and bizarrely no proprietary control of any of the component parts of the product was maintained.
Lowe reviewed what the Altair and others had achieved and the multi-million dollar businesses they had generated. In 1979 he produced a market analysis identifying both a business and a consumer sector.
He proposed either IBM should acquire one of the currently active companies (he went as far as suggesting this might be Atari) or that it should develop its own approach. The company’s Corporate Management Committee was receptive and asked for a prototype to be developed.
Bill Syndes was put in charge of twelve engineers to develop the prototype. The software lead, Jack Sams, had recently been involved in the System/23 Datamaster development and the internal creation of a BASIC. The lack of that software for the project had delayed progress heavily and it was largely for this reason that when the prototype was presented to the management committee it was proposed they acquire CP/M and BASIC externally. The planned development haste was another reason it was decided not to produce the OS and languages in-house.
Its bus specification mirrored the Datamaster format and was openly published to attract others to develop expansion cards for the IBM PC. In fact they would have had to publish it anyway because of an anti-trust resolution in the 1950s.
The prototype and approach were approved and a new IBU was formed in July 1980 at Boca Raton in Florida under the Project Chess code name and located functionally within the Entry Systems Division.
The team was given the right to breach all the normal IBM processes and told that it had just one year to come up with a PC.
Management books of the time took great interest in such clearly stated missions. President J F Kennedy had told NASA to put a man on the Moon by the end of the ‘60s. This directive helped to focus the mind, the team asking itself which course was the more likely to achieve that objective, and then selecting that route.
IBM’s team of twelve was initially headed by Bill Lowe, but having originated the people and the approach he was promoted elsewhere within IBM before the project finished. He was allowed to select his own replacement and he chose Don Estridge, a Floridian who had been in the army and at NASA before joining IBM.
Project Chess set out to do the impossible in IBM terms, create a new product in around a year. The IBM PC or 5150 would launch in September 1981.
The challenge was judged as fair, but a one-year deadline made it essential that they acquire technologies and software rather than take the time to develop them in-house. Here was another breach of the IBM norm. In the past it had always tended towards internal development, just as it preferred to promote from within too.
The product took on the codename ‘Acorn’, clearly implying the great oaks that this entry-level PC might grow for IBM. Acorn Computers in the UK was already founded back in 1978, but its Electron was not released until 1983 so there was no evident clash at the time in this project-name selection.
For speed a pre-existing monitor already developed by IBM Japan was selected together with an appropriate Epson printer.
The monitor was from the earlier IBM Datamaster project. It was assessed as a product that would not require a specialist installer; it was virtually plug-and-play. Work on Datamaster had been running for several years before ‘Project Chess’ and the ‘Acorn'; it came to market just a month before the IBM PC.
The team almost selected a processor developed internally, one used for the IBM 801. The 801 was named after the building in the Thomas J Watson Research Center where John Cocke experimented to find improved performance for IBM computers; he would win a Turing Award and a Presidential Medal of Science for this work.
The 801 processor was a very early RISC (reduced instruction set computer). RISC used primitives to work with compilers and make the most of the architecture of the computer. This CPU would perhaps have been a good choice for them. It was certainly more powerful than the eventual choice and had a more advanced operating system than the DOS that they would come to use; and it was their technology!
The need-for-speed for the planned development necessitated an ‘open-architecture’ approach so third party developers would be encouraged to populate the required range of optional extras which would expand sales of the PC; without IBM having to do this for itself.
In an early conversation with Bill Gates at Microsoft, Jack Sams indicated that they were looking for Microsoft to supply an operating system for an 8-bit MPU. Gates recommended that instead they leapfrog the current 8-bit microprocessors and go straight for a more future-proofed 16-bit approach.
It would appear that the design parameters were still a tad fluid in that they proved readily willing to be swayed in a single conversation with a third party supplier. The Intel 8088 16-bit processor was selected.