Meteorologist, oceanographer, Python solvem probler, experimenter, Raspberry Pi tinkerer, geek dad, manager, space enthusiast, ham radio lover, science columnist, ADHDer, synthetic intellect curious, linux admin cosplayer.

An Wang

Born: February 7, 1920, Shanghai, China Died: March 24, 1990, Boston, Massachusetts (aged 70) Cause of death: Esophageal cancer

A Chinese-American physicist, inventor, and entrepreneur. Co-inventor of the pulse transfer controlling device (the destructive-read-with-rewrite mechanism that made magnetic-core memory practical). Founder and lifelong head of Wang Laboratories, which by the mid-1980s briefly dominated the office word-processing market and reached $3 billion in annual revenue. One of the great Asian-American business success stories of the twentieth century, and one of the great parallel inventors in early computing – his Harvard work on core memory ran in parallel with Jay Forrester’s Whirlwind effort at MIT, and the resulting patent dispute was one of the formative IP battles of the early computer industry.

Childhood and Education

Wang was born in Shanghai on February 7, 1920, the eldest son of Yin Lu Wang, a teacher of English and a practitioner of traditional Chinese medicine, and his wife Z.W. Chien. The name “An Wang” means “peaceful king.” He spent his first six years in his maternal family’s Shanghai compound before the family moved to his father’s ancestral city of Kun San (Kunshan), about 30 miles west of Shanghai, where he started formal school.

Because the family elementary school lacked the first two grades, Wang began in the third grade and was thereafter always two years younger than his classmates – a permanent disadvantage he later credited with teaching him to swim rather than sink. His father began teaching him English at age four; his paternal grandmother tutored him in Confucianism, which Wang later described as “the practical philosophy that has profoundly influenced Chinese character” and credited with the principles of moderation, balance, and simplicity that he applied to business.

At thirteen Wang enrolled at Shanghai Provincial High School, about 10 miles outside the city, where instruction was conducted entirely from American textbooks. At sixteen he entered Chiao Tung University in Shanghai (often called “the MIT of China,” the institution his father had briefly attended), having posted the highest entrance examination scores of his class. He studied electrical engineering with an emphasis on communications, again from American textbooks, and later joked that he spent more time on the university table-tennis team than on his coursework. He graduated in 1940 with a Bachelor of Science in electrical engineering.1

His mother had died in 1936, broken by the Japanese aerial bombings of Shanghai. After graduation Wang remained at Chiao Tung as a teaching assistant. In the summer of 1941, with the Japanese occupation tightening, he joined a group of eight young engineers who slipped through Japanese lines from Shanghai to Hong Kong, then through the French concession of Kuan Chou Wan and overland to Nationalist-held Kweilin (now Guilin), 300 miles inside China. There the 21-year-old Wang was put in charge of scrounging parts and improvising designs to build radio transmitters for Chinese Nationalist troops. In late 1944, just before the Japanese took Kweilin, his group was evacuated to Chungking. Wang did not learn until much later that during this period his father and his older sister Hsu had also died as a result of the war.1

Harvard Years 1945-1951

In March 1945 Wang placed second in a competitive examination for a Nationalist government program sending several hundred young engineers for two years of advanced training in the United States. In April he was flown over the Hump from China to Ledo, India, took a train to Calcutta, and embarked on a month-long voyage to Newport News, Virginia, arriving in June 1945. The group was housed initially at Georgetown University in Washington, D.C.

While most of his cohort accepted apprenticeships at firms such as Westinghouse and RCA, Wang applied to Harvard for graduate school. The timing was fortunate: Germany had surrendered, but in early summer 1945 a long war with Japan still seemed certain, most American men were still in uniform, and Harvard had more openings than students. Wang entered the Harvard Graduate School of Applied Physics in September 1945. He earned an MS in applied physics in 1946 and a PhD in applied physics in June 1948 under a thesis in nonlinear mechanics that he chose explicitly because he knew it would satisfy his committee quickly. (His PhD adviser was not Aiken, with whom he later worked; the popular version that he was Aiken’s student is wrong – Wang only walked across campus to look for a job after submitting his thesis.)1

After receiving his degree in June 1948, Wang – daunted by the security paperwork required by Hughes Aircraft – walked across the Harvard campus to ask whether the Harvard Computation Laboratory had openings. The lab consisted of Howard Aiken, five or six research fellows, and a handful of assistants in a new but undistinguished two-story brick building still housing the original Mark I on the ground floor. Aiken was working on the Mark IV, his first all-electronic machine, for the U.S. Air Force. After interviewing Wang, Aiken hired him as a research fellow effective July 1, 1948.1

Pulse Transfer Controlling Device 1948-1949

On May 18, 1948 – within a day or two of Wang starting work, and even before his formal July 1 appointment – Aiken posed the central problem: how to record and read magnetically stored information without mechanical motion. Computer designers across the small early-computing community agreed that a toroidal core of high-remanence magnetic material could in principle store one bit by being magnetized in one of two directions. The problem was that reading the stored information disturbed or destroyed it.12

Wang struggled with the problem for about three weeks. The solution came to him while walking through Harvard Yard: it did not matter that reading was destructive, provided the information was rewritten back into the core immediately afterward. His notebook entry for June 29, 1948 records: “It is very possible that the information can stay there [in the core in the form of a particular magnetic direction] and be transferred many times before the information [is lost or muddied].”1 As Wang later put it: “The idea is that by destroying the information – I know it.”13

To realize the idea, Wang found a U.S. Navy publication describing a German wartime magnetic material called Permanorm 5000-Z (the U.S. equivalent, Deltamax, was made by Arnold Engineering, a subsidiary of Allegheny-Ludlum). Working with his Harvard colleague Way-Dong Woo – a fellow Shanghai native and Chiao Tung graduate who also worked at the Computation Lab – Wang strung the toroids in series to form a magnetic-core delay line, which became the memory in Aiken’s Mark IV (and, notably, almost no other machine).14 Wang and Woo published their work as “Static Magnetic Storage and Delay Line” in the Journal of Applied Physics on January 1, 1950.4

Wang first discussed patenting the device with his fiancee Lorraine Chiu in June 1949 – they had met at a Boston-area gathering for Chinese students in 1948 and would marry the following month. Chiu encouraged him to file. Harvard at that time only reserved patent rights to itself for inventions related to public health; the administrator Wang consulted advised him to file at his own expense and recommended Harvard’s own patent firm, Fish, Richardson & Neave. (The IEEE Computer Pioneer biography notes the irony that Harvard and Pennsylvania, both of which discouraged staff patents, “quickly lost and never regained their original computing eminence,” while MIT, which actively pursued them, has been “at the forefront of computing research” ever since.)1

On October 21, 1949, Wang filed his patent application, titled “Pulse Transfer Controlling Devices,” with 34 claims, most directed at “an information delay line.”15 The application was eventually granted as U.S. Patent 2,708,722, issued May 17, 1955.1 Claim 24 – the broad concept of destructive-read-with-immediate-rewrite in a magnetic core – ultimately controlled later patents in the field by virtue of its priority and breadth.

After filing, Wang suffered a sleepless night before telling Aiken. To his surprise, Aiken did not react at all and later gave Wang a substantial raise. On September 29, 1949, just before filing, Wang described the invention publicly at a Harvard computing symposium and subsequently published in both the Journal of Applied Physics and the Proceedings of the Institute of Radio Engineers, ensuring that the concept was widely known.1

IBM Patent Dispute and Settlement 1953-1956

By 1950 it was clear to Wang that Harvard intended to wind down basic computer research as a matter of policy. He left the Computation Laboratory in April 1951 and in June 1951 founded Wang Laboratories as a sole proprietorship in Boston with $600 in savings, no orders, no contracts, and no office furniture.1

One of his first business steps was to ask IBM whether they wanted to license his pending patent. After more than a year of correspondence and visits, in November 1953 Wang and IBM reached an agreement under which Wang would consult for IBM in exchange for a thousand dollars a month and IBM would receive a three-year option on a non-exclusive license to his still-unissued patent. Wang later wrote that he believed IBM was less interested in the specific consulting projects than in cultivating access to his thinking on magnetic-core applications.1

When the patent finally issued on May 17, 1955, IBM opened formal negotiations. Wang opened by offering to sell the patent outright for $2.5 million; IBM countered that “even half of $2.5 million is too high.” In October 1955, IBM offered $500,000 plus 70 percent of all royalties from third-party licensing.16 Wang was prepared to accept but objected to language regarding patent validity and interference proceedings. After Wang answered 58 questions attacking the patent from “every conceivable angle,” IBM revealed that it had located a 1947 patent application by Frederick W. Viehe, a Los Angeles public works inspector, which IBM believed would “certainly lead to an interference.”

Under this pressure, Wang assigned the patent to IBM in March 1956 under the October terms, with provisions allowing $100,000 of the payment to be withheld in case of an interference declaration. Two weeks before the deadline, in May 1956, the Patent Office did declare an interference with the Viehe application. In November 1957, mid-litigation, IBM bought the Viehe application – reputedly for $1 million – and hired Viehe as a consultant. The interference proceeding then continued as IBM versus IBM. The eventual decision awarded Viehe only one minor claim. Wang forfeited the $100,000 and harbored a lifelong grudge against IBM.1 (When Viehe died in 1960 he left a $625,000 estate; his son later confirmed that the buyer of his father’s patent had been IBM.)

A 1975 internal memo by J. William Hinkley of MIT’s Research Corporation, surfaced in later litigation, recorded a 1962 meeting in which IBM vice president James W. Birkenstock allegedly remarked “that they had probably underpaid Wang or Viehe for their patents… in view of the tremendous increase in the size of the computer industry.” According to the memo, Birkenstock cited this as an example of his negotiating skill but was “severely taken to task” by Thomas Watson Jr. for the comment.1

The figure most often cited in the trade press is the $500,000 lump sum, and that is the figure that appears in Computer History Museum records and the IEEE Computer Pioneer biography.16 (Some sources round up to $700,000 by adding the disputed $100,000 final tranche and other royalty income; the cleaner number, and the one to use, is $500,000.) Whatever the precise total, it became the founding capital that funded the early growth of Wang Laboratories.

In 1964 – separately from the Wang dispute, but in the same field – IBM also settled with MIT on Forrester’s broader array-architecture patent for $13 million, then the largest patent settlement in history.7 Forrester personally received $1.5 million; most of the rest went to MIT’s educational programs. The two settlements together fairly capture the parallel-invention story: Wang’s narrower, earlier single-cell patent was worth half a million dollars; Forrester’s broader array patent was worth more than twenty-five times as much.

Wang Laboratories 1951-1990

From the IBM money and a steady stream of consulting contracts, Wang slowly built the firm. The first major product was the LOCI-2 (Logarithmic Computing Instrument) desktop calculator, introduced in January 1965 – the first desktop calculator capable of computing logarithms and the machine that effectively created the desk-calculator market.8 Wang followed with the 700-series scientific calculators (announced 1969, shipping 1970) and, most consequentially, the Wang 1200 word processor (announced late 1971, available 1972) and the CRT-based word processors that followed in 1976.

By the mid-1970s Wang Labs controlled an estimated 80 percent of the office word-processing market.8 Revenue reached $1 billion in 1982, $2 billion in 1984, and approximately $3 billion at its peak in the mid-1980s, when the company employed more than 33,000 people. Forbes listed Wang’s net worth at $1.6 billion in 1984.2

Wang lived in Lincoln, Massachusetts, gave generously in the Boston area, and lent his name to several institutions: the Wang Institute of Graduate Studies in Tyngsborough (founded 1979, funded until 1987), the Wang Theatre in downtown Boston (the restored Metropolitan Theatre, renamed in 1983), and the Wang Building at Massachusetts General Hospital (a $4 million ambulatory care center). He was awarded the Medal of Liberty by President Reagan in 1986 and inducted into the National Inventors Hall of Fame in 1988.2

The Boston-Cambridge cultural geography mattered. Wang was based in Lowell, not Cambridge; his patent had been beaten down by IBM’s Boston law firm (Fish, Richardson & Neave, the same firm that would later represent IBM); his great rival in the core-memory story, Forrester, was at MIT, just across the river from Harvard. Wang was a Chinese immigrant in a New England technology corridor that was overwhelmingly white and Protestant. The chip on his shoulder against IBM was real, and it shaped his strategic decisions for thirty years.

Wang’s autobiography Lessons, written with Eugene Linden, was published by Addison-Wesley in 1986; the proceeds were donated to the Wang Institute.9 In the book Wang attributes his success to adapting technology to society’s needs and applying Confucian values of balance, moderation, and simplicity to his business practices.

The 1980s decline was steep. Wang Laboratories’ main business by then was its proprietary VS line of minicomputers, increasingly squeezed between commodity PCs running WordPerfect and Microsoft Word on one side and Sun, IBM, and HP servers on the other. Sales fell sharply in the third quarter of 1985 – the company’s first decline in ten years. In 1986, after taking leave for cancer treatment, Wang insisted on handing the company to his son Frederick Wang. Charles C. Kenney’s history of the firm called Fred Wang, a business school graduate, “by almost any definition… unsuited for the job in which his father had placed him.” Senior R&D and business personnel resigned. After Fred Wang’s reputation collapsed in the wake of a “vaporware announcement,” the elder Wang was forced to remove his son from the presidency in 1989. Losses reached $92 million in 1988. The firm filed for Chapter 11 bankruptcy on August 18, 1992, two and a half years after its founder’s death.8

Death March 1990

An Wang died of esophageal cancer on March 24, 1990, in Boston, at age 70. He had taken a leave of absence for cancer treatment in 1986 and had ceded day-to-day control to his son in the same year, but he never fully retired and remained chairman until the final months of his illness.2

He was survived by his wife Lorraine, two sons (Fred and Courtney), and a daughter (Juliette). His Confucian sense of family loyalty – and the insistence that his sons, whatever their qualifications, were the rightful successors – contributed both to the company’s late expansion and to its rapid collapse, and is a recurring theme in every honest account of the firm. Wang Laboratories filed for Chapter 11 in August 1992, emerging in 1993, and after several rebrandings was eventually absorbed; the Wang name today survives chiefly in the Wang Theatre in Boston, the buildings that bear his name in the Boston-area medical and academic landscape, and in his autobiography.

Sources

  1. An Wang IEEE Computer Pioneer biography, ed. T.M. Smith, IEEE Computer Society. PDF at history.computer.org/pioneers/pdfs/W/Wang.pdf – the principal source for Wang’s early life, the invention story, and the IBM negotiation. The IEEE biography draws heavily on Wang’s autobiography (Wang 1986) and the obituary literature. ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18

  2. An Wang – Wikipedia. https://en.wikipedia.org/wiki/An_Wang – Accessed: 2026-04-29. ↩ ↩2 ↩3 ↩4

  3. Emerson W. Pugh, Memories That Shaped an Industry: Decisions Leading to IBM System/360 (MIT Press, 1984) – the standard reference on the early history of computer memory and the IBM-Wang-Forrester patent landscape. Cited extensively in the IEEE Wang biography. ↩

  4. An Wang and Way-Dong Woo, “Static Magnetic Storage and Delay Line,” Journal of Applied Physics 21:49 (January 1, 1950). ↩ ↩2

  5. An Wang, “Pulse transfer controlling devices,” U.S. Patent 2,708,722, filed October 21, 1949, issued May 17, 1955. https://patents.google.com/patent/US2708722 ↩

  6. Computer History Museum, “March 4: An Wang Sells Core Memory Patent to IBM.” https://www.computerhistory.org/tdih/march/4/ – gives March 4, 1956 as the assignment date and $500,000 as the figure. Accessed: 2026-04-29. ↩ ↩2

  7. “IBM Paying MIT In Patent Dispute,” The Harvard Crimson, April 15, 1964 – contemporaneous reporting on the $13 million IBM-MIT settlement; archives. ↩

  8. Wang Laboratories – Wikipedia. https://en.wikipedia.org/wiki/Wang_Laboratories – Accessed: 2026-04-29. ↩ ↩2 ↩3

  9. An Wang with Eugene Linden, Lessons: An Autobiography (Addison-Wesley, 1986). Internet Archive copy at https://archive.org/details/lessonsautobiogr0000wang ↩