Lynn Conway was one of the computer engineers whose work helped shape the architecture of modern processors. In 1965, while working at IBM, she developed an approach to selecting multiple instructions that were ready to execute during each processor cycle. The idea became part of the history of dynamic instruction scheduling, although her career at the company ended abruptly in 1968 after she announced her gender transition.

The key facts about Lynn Conway in 20 seconds

Conway worked on IBM’s Advanced Computing Systems (ACS) project, where she developed an approach to selecting multiple instructions for execution in each processor cycle.
An internal memorandum dated 23 February 1966 documented the work, with Conway and three other authors.
IBM dismissed her in 1968 after she announced her gender transition.
Decades later, IBM apologised for the dismissal in October 2020. Conway died on 9 June 2024, aged 86.

The IBM research that anticipated modern processor techniques

Conway’s work at IBM focused on the Advanced Computing Systems project, known as ACS. In 1965, she developed a method for selecting several instructions that were ready to execute during the same processor cycle. The approach was documented in an internal memorandum titled Dynamic Instruction Scheduling, dated 23 February 1966.

Lynn Conway, the IBM engineer whose idea helped change modern processors | Lynn Conway July 2006 — Lynn Conway, the IBM engineer whose idea helped change modern processors

The document had four authors: Lynn Conway, Brian Randell, Don Rozenberg and Don Senzig. It described work on a technique that would become an important part of the development of processor architectures capable of executing instructions out of their original program order when the necessary conditions were met.

The memorandum was an internal IBM document rather than a public research paper. Its subsequent history, including its republication and eventual recovery by Conway, would become an important part of the story of her contribution to computer architecture.

In 1968, however, Conway’s career at IBM came to an abrupt end. After telling the company that she intended to transition, IBM dismissed her. A planned transfer was cancelled following consultation with senior management. Conway later believed that the decision may have involved IBM chairman and chief executive T. J. Watson Jr., although there is no direct documentary evidence establishing his personal involvement.

The dismissal forced her to rebuild her professional career elsewhere, at a time when openly discussing a gender transition could have serious consequences in the workplace.

From Memorex and Xerox to the recovery of the ACS archive

After leaving IBM, Conway joined Memorex in 1969 and later moved to Xerox. Her career continued in computing, but the work she had carried out at IBM remained difficult to trace through the company’s public technical record.

In 1969, IBM republished the Dynamic Instruction Scheduling memorandum as research report RJ 565. A related patent, US 3,718,912, was filed on 22 December 1970 and granted on 27 February 1973. The patent lists five inventors, but Conway is not among them.

That distinction matters when reconstructing the history of the technology. The existence of the patent and the earlier memorandum provides documentary evidence of the work, but it does not, on its own, establish that Conway was deliberately excluded from the patent or that the company appropriated her contribution. Such a conclusion would require evidence beyond the documents available.

Conway eventually began recovering the history of the ACS project decades later. In 1998, she came across a webpage maintained by Mark Smotherman about IBM’s Advanced Computing Systems work. The discovery gave her a way to reconnect the surviving documentation with the people who had worked on the project.

In 1999, Conway contributed an archive containing 392 pages of material. IBM authorised the historical distribution of the documents in 2000, making it easier for researchers and historians to study the project and its technical contributions.

The archive helped preserve a record of work that had remained largely inside IBM’s research history. It also allowed Conway to document her own involvement in the project and provide material that could be examined independently.

Recognition, IBM’s apology and her legacy

Conway received the IEEE Computer Society’s Computer Pioneer Award in 2009. The award recognised her contributions to computing, including her work on processor architecture and her later influence on the development of very large-scale integration (VLSI) design methods.

Her story also became part of a broader discussion about the treatment of transgender people in the technology industry and the historical exclusion of women and minorities from technical recognition.

On 14 October 2020, IBM formally apologised to Conway for firing her in 1968. The apology came 52 years after the dismissal and acknowledged the harm caused by the company’s decision.

Conway died on 9 June 2024, aged 86. Her career left a complex historical record: an early contribution to processor architecture, a dismissal that interrupted her work at IBM, a successful return to the technology industry and decades spent ensuring that the history of the ACS project was not lost.

The surviving memorandum, patent records and recovered archive provide different pieces of that history. Together, they help explain both the technical work carried out at IBM in the 1960s and the difficulties Conway faced in receiving recognition for her contribution.

Frequently asked questions

What did Lynn Conway contribute to processor architecture?

In 1965, Conway developed a method for selecting multiple instructions ready for execution during each processor cycle. The work was documented in IBM’s 1966 Dynamic Instruction Scheduling memorandum.

Why did IBM fire Lynn Conway?

IBM dismissed Conway in 1968 after she announced her intention to transition. A planned transfer was cancelled following consultation with senior management.