ACM: The Association for Computing Machinery

Abstract

The Association for Computing Machinery was founded on September 15, 1947, before most universities had computer science departments and before “software” had a name. It created the Turing Award, often called computing’s Nobel, runs SIGGRAPH, SIGPLAN, SIGOPS, and dozens of other research communities, and publishes the journals in which much of the field’s foundational work first appeared. Its founders were mathematicians with a practical problem: they had machines but no organized way to share what they learned about them.

The ACM logo. Image: Association for Computing Machinery, public domain, via Wikimedia Commons.

Founding at Columbia (1947)

The ACM was established at Columbia University in New York City on September 15, 1947, two years after the end of World War II and nine months before the first stored-program computer ran at Manchester. Seventy-eight people attended, and they voted to found the Eastern Association for Computing Machinery; the Council dropped “Eastern” from the name in January 1948. The organizing push came from Edmund C. Berkeley, an actuarial mathematician at Prudential Insurance who had become fascinated by electronic calculating machines, and John H. Curtiss, who on July 1, 1947, had become chief of the National Bureau of Standards’ new National Applied Mathematics Laboratories (later the Applied Mathematics Division).

The group that gathered was heterogeneous: mathematicians, engineers, statisticians, and early practitioners who had worked on the ENIAC project, on ballistics calculations, or on wartime operations research. They shared a practical problem: they were using or building computing machines, had no professional organization, no journal, and no systematic way to exchange information with each other.

John H. Curtiss became ACM’s first president, serving from 1947 to 1948; John W. Mauchly was vice-president and Robert V. D. Campbell treasurer. Curtiss understood computing primarily as a tool for applied mathematics; his NBS work focused on numerical methods and the calculation of mathematical tables. This early orientation toward computation as applied mathematics shaped ACM’s character in its first decade: more mathematical than engineering, more academic than industrial.

Edmund C. Berkeley served as ACM’s first secretary and was its public voice in the early years. He was an unusual figure, a computing enthusiast who wrote Giant Brains, or Machines That Think (1949), one of the first popular books about computers, while simultaneously building simple relay computers in his workshop and organizing the ACM’s administrative infrastructure. His more democratic vision of computing, and his interest in its social implications, contrasted with the academic orientation that eventually dominated the organization.

The name “Association for Computing Machinery” was deliberately chosen to avoid the word “computer,” which in 1947 referred to a human being (a person who computed) rather than a machine. “Computing Machinery” meant the machines themselves. Within a decade, as electronic computers became the dominant meaning of “computer,” the original distinction faded, and ACM’s name became slightly paradoxical: an association for the machinery, run by the people.

Communications of the ACM (1958)

ACM’s first journal, the Journal of the ACM, appeared in January 1954. Its flagship general publication, the Communications of the ACM (CACM), began in 1958. Alan J. Perlis, then at the Carnegie Institute of Technology (which became Carnegie Mellon in 1967), was the first editor-in-chief, with Saul Rosen as managing editor; he made CACM the field’s main general-interest journal.

CACM under Perlis and his successors published papers that are still cited today: Dijkstra’s “Go To Statement Considered Harmful” (1968, as a letter to the editor), C.A.R. (Tony) Hoare’s Quicksort (Algorithm 64, 1961) and his axiomatic basis for programming (1969), and many others. CACM occupied an unusual position: rigorous enough for researchers, broad enough for practitioners, and widely read at a time when it was one of the few reliable sources of technical computing information.

The ACM Digital Library, launched in 1997 with full text back to 1991 and metadata back to 1985, has expanded continuously and has digitized CACM’s full history and ACM’s conference proceedings.

The Turing Award (1966–Present)

The Turing Award is ACM’s best-known creation. Named for Alan Turing, whose work on computability, the Turing machine, and the theoretical foundations of computing occurred before the ACM existed, the award was first given in 1966.

Alan J. Perlis received the first one, for his influence on programming languages, compilers, and computer science education. The next nine awards went to Maurice Wilkes (1967), Richard Hamming (1968), Marvin Minsky (1969), James Wilkinson (1970), John McCarthy (1971), Edsger Dijkstra (1972), Charles Bachman (1973), Donald Knuth (1974), and, in 1975, jointly to Allen Newell and Herbert Simon.

The award was worth little in cash for its first decades. Intel and Google together funded a $250,000 prize from 2007 to 2013, and Google alone raised it to $1,000,000 in 2014. That puts it on the scale of a Nobel Prize, though few people outside computing have heard of it.

Newell and Simon’s award, for their joint work on artificial intelligence and cognitive science, was the first to be shared. Forty years after the first award, Frances Allen became the first woman to receive it, in 2006, for her work on compilers and program optimization at IBM Research. Barbara Liskov followed in 2008 for data abstraction, the Liskov substitution principle, and the CLU language. In 2018 the award went to three deep-learning researchers at once, Yann LeCun, Geoffrey Hinton, and Yoshua Bengio, for the conceptual foundations of deep neural networks.

Info

Through the 2025 award (Charles Bennett and Gilles Brassard, for quantum information), three women have received the Turing Award in sixty years: Frances Allen (2006), Barbara Liskov (2008), and Shafi Goldwasser (2012). This record has drawn criticism. Most laureates worked in the United States or Europe, where computing research was concentrated and most ACM members lived.

Special Interest Groups

Much of the ACM’s working life happens in its Special Interest Groups (SIGs), each organizing a research subfield. There are 38 SIGs, ranging from the very large to the specialized:

SIGGRAPH (Special Interest Group on Computer Graphics) organizes the annual SIGGRAPH conference, which has been the largest venue for computer graphics and interactive techniques since the 1970s. SIGGRAPH conferences are unusual among computing venues: they are simultaneously research conferences, artist showcases, and industry exhibitions, blending academic papers with animated films, real-time demonstrations, and commercial technology. The first SIGGRAPH conference, held in Boulder in 1974, opened a venue that carried graphics research (including the recursive subdivision and texture mapping that Ed Catmull pioneered in his 1974 Utah PhD thesis) into both research and entertainment computing. (Catmull’s better-known subdivision surfaces, the Catmull–Clark scheme, came later, in 1978.)

SIGPLAN (Programming Languages) organizes PLDI (Programming Language Design and Implementation), POPL (Principles of Programming Languages), and OOPSLA (Object-Oriented Programming, Systems, Languages, and Applications), where much of the field’s programming language research appears.

SIGOPS (Operating Systems) runs SOSP (Symposium on Operating Systems Principles) and co-sponsors USENIX’s OSDI (Operating Systems Design and Implementation), where many of the best-known systems papers appeared: the Unix paper (SOSP 1973), the exokernel (SOSP 1995), MapReduce (OSDI 2004), and most of the major cloud systems papers.

SIGCOMM (Communications) organizes the networking research community, with the annual SIGCOMM conference as its flagship.

SIGKDD (Knowledge Discovery and Data Mining) grew quickly during the data science era, with the KDD conference becoming one of the field’s largest.

SIGCSE (Computer Science Education) is unusual among SIGs in focusing on pedagogy rather than research; it addresses how CS is taught at all levels and has become increasingly influential as computing education has expanded globally.

ACM’s International Presence

ACM is nominally an American organization (founded in the US, headquartered in New York) but operates globally. Its international presence operates through three structures:

National chapters exist in most major countries, providing local programming, networking, and connections to national computing communities. The ACM India Council is one of the most active, reflecting India’s large computing population and the numerous Indian computing professionals who joined ACM through academic and professional exposure to American computing culture. The German Chapter of the ACM, founded in 1968 by Albert Endres, is one of the oldest national chapters. It predates the Gesellschaft für Informatik by one year, which makes it Germany’s oldest computing professional society. It has operated continuously since its founding, and since 2004 it has run 34 regional groups jointly with the GI. A full history is in German Chapter of the ACM.

Student chapters, more than 680 of them alongside over 180 professional chapters, put ACM into university computing departments from Lagos to Lahore. The first was chartered in 1961 at what is now the University of Louisiana at Lafayette. Student chapters organize programming competitions, technical talks, and career events; for many computing students in countries without strong national computing societies, the ACM student chapter is their primary professional community.

ACM-W (ACM’s Women in Computing initiative) organizes programs to support women in computing globally, including the Athena Lecturer Award for senior women researchers and regional “Celebrations” events that have grown into major gathering points for women in the field.

ACM Europe has grown in significance as European computing research has matured, and the ACM Europe Council advocates for European computing interests within the global ACM structure.

ACM’s relationship to the IEEE Computer Society is mostly cooperative: both publish, both organize conferences, and ACM/IEEE joint conferences (such as SC for supercomputing and DAC for design automation) are among the field’s most important. The historical division (ACM as academic, IEEE as engineering-and-standards) has blurred as computing has become both more academic and more engineering-intensive.

The Professionalism Question

ACM has argued for decades about whether software work should be a licensed profession, with certification required before practice, as for engineers or physicians.

It created curricula, published educational guidelines, and offered certification programs, but has never pushed for mandatory licensing of software practitioners. Computing is too diverse (the same activity encompasses children programming Scratch and teams building safety-critical aviation software), too fast-changing for any fixed curriculum to remain current, and too distributed across industries to be regulated through a single professional body.

The debate returns after each software failure that hurts people, and software now controls vehicles, medical devices, and infrastructure. ACM’s 1992 Code of Ethics and its revisions list a professional’s obligations but provide no way to enforce them.

ACM Today

In the 2020s ACM has roughly 110,000 professional and student members worldwide. An endowment funds research awards and fellowships, and the digital library holds millions of articles. Its tax filing for the year ending June 2024 reports $88.3 million in revenue, of which $74.8 million came from program services: conferences and publications.

Its problems are structural. Open access has eroded the financial model based on journal subscriptions and conference proceedings; online communities increasingly satisfy professional networking needs; and the most engaged computing communities organize around specific technologies rather than around a general professional association.

ACM has responded by speeding up its move to open access, running research programs with industry partners, and expanding its education work. An association founded so that users of card-calculating machines could swap notes now competes for attention with GitHub and YouTube.