Phoenix rebuilt IBM's BIOS from a spec to make clones legal

By the early 1980s the IBM PC had become the platform software vendors wrote for, but everything else in it — the processor, memory, disk controllers — was off-the-shelf hardware anyone could buy and assemble. The single component standing between a hobbyist clone and a fully IBM-compatible machine was the BIOS, the firmware that let the operating system talk to the hardware, and IBM held a straightforward copyright on its code.

Any engineer who had read IBM's published BIOS reference manuals and then wrote replacement code risked producing something a court could find was copied, consciously or not, from IBM's copyrighted implementation — a legal minefield that had already begun to produce lawsuits against clone makers who cut corners.

Split teams: one wrote a spec from IBM's manuals, the other coded from it, proving independent creation.

The spec described behavior, not code, so the clean team never saw IBM's expression.

The documented handoff created a paper trail proving independent creation.

The resulting BIOS was functionally identical but legally distinct, so IBM couldn't sue.

When copyright blocks compatibility, separate the people who study the original from those who write the replacement, and hand over only a behavioral spec.

The clean-room BIOS held up in court; IBM sued other clone makers but not Compaq or Phoenix, which then licensed it industry-wide. Clean-room reverse engineering, formalized and legally validated through this case, became the standard method the industry still uses to build interoperable products without infringing copyright, and by removing the BIOS bottleneck it opened the door to the price-competitive, commodity PC-clone industry — including the low-cost machines running Microsoft's software that ultimately eclipsed IBM's own PC business.