OBITUARY

Aaron D. Wyner, 1939 - 1997

Aaron Wyner, a world-renowned information theorist, died of cancer on September 29, 1997, at the Morris Hills Nursing Home, Morristown, NJ.

Aaron was born on March 17, 1939, and grew up in the Bronx, New York City. He received a BS from Queens College in 1960, and a BS and PhD from Columbia University. He served as an assistant professor at Columbia before joining Bell Laboratories where he remained until his death. Aaron was stricken with cancer over ten years ago, but worked with courage, enthusiasm, optimism, and often brilliance, until just one month before he died. Those of us who were privileged to know him will treasure the memory of his warm personality and his devotion to his family, colleagues, and profession. He will be sorely missed.

Aaron Wyner contributed more fundamental new ideas and results to information theory than perhaps any other researcher since Shannon. Moreover, he was consistently alert in pointing out the applicability of his results to the practice of communications and cryptography. His work spanned so many specialties (channel coding theory, multiuser coding theory, source coding theory, optical communications, algebraic coding, cryptography, stochastic process theory, mobile communications, etc.) that one is again reminded of Shannon's work.

Aaron's contributions to information theory began with his Columbia doctoral thesis in which he worked out the first comprehensive algebraic theory for what are now called convolutional codes. He also found the convolutional codes that are the analogs of the Hamming single-error-correcting block codes. But his most distinguished research contributions are generally regarded to be those to channel coding theory, particularly to the theory of the Gaussian channel. His mathematically precise, but practically oriented, approach resulted in a series of outstanding contributions to the problem of signaling in the presence of Gaussian noise, both for the single sender and the many-senders cases.

Aaron's interest in cryptography led him to invent the ``wire-tap channel.'' In his 1975 BSTJ paper introducing this channel, he worked out virtually its complete theory, showing the then-surprising fact that one could obtain perfect secrecy without the advance exchange of secret keys provided that the legitimate receiver enjoys a better channel from the sender than does the wire-tapping opponent. This work, which preceded the introduction of public-key cryptography, has had a steadily increasing influence on developments in this field. Recent results by Maurer and by Csiszar and Korner showing the remarkable usefulness of ``common randomness'' in cryptography are direct consequences of Wyner's work. In 1979, Aaron invented and patented a secure voice-scrambling scheme that, unlike all previous schemes, did not expand the bandwidth of the signal. This is the only published voice-scrambling scheme of which we are aware that does not have substantial cryptographic weaknesses.

Aaron's recent papers on the redundancy of the Lempel-Ziv data compression algorithm, on the information theory of cellular mobile radio systems, and on the effect of limited memory in classification problems shows that he continued unabated to his death to produce innovative and important work over a broad range of topics within information theory.

It would be impossible to find anyone who worked as unselfishly as did Aaron Wyner for the advancement of information theory, in particular for the interests of the IEEE Information Theory Society. He held every possible service position (Editor of the IEEE Transactions on Information Theory, Associate Editor, President of the Society, Co-Chairman (twice) of the IEEE International Symposium on Information Theory, Workshop Co-Chairman). He was elected to many terms on the Information Theory Society Board of Governors. Aaron discharged every one of his professional duties extraordinarily well. His editing with N. J. A. Sloane of the collected papers of C. E. Shannon culminated in the publication of these papers as an IEEE book of about one thousand pages. Aaron's editing, which was a labor of love that spanned about ten years, was the sine qua non of this wonderful addition to the literature of information theory -- it includes many extraordinary papers that were not previously obtainable. Those who do research in information theory will always be greatly indebted to Aaron Wyner.

Aaron's many honors include all the awards of the IEEE Information Theory Society (the Claude E. Shannon Award, the Prize Paper Award, and designation as Shannon Lecturer). He was also a Distinguished Member of the Technical Staff at Bell Laboratories, a Fellow of the IEEE, and a Member of the U. S. National Academy of Engineering.

Aaron is survived by his wife, Nusha; his three daughters, Mrs. Tamar Wyner Herman, Mrs. Dena Wyner Glasgow, and Yael; his son, Abraham; a sister, Mrs. Sara Gootblatt; his mother, Mrs. Mary Jacobson Wyner; and six grandchildren.

It seems appropriate as we end this eulogy for a man who deeply admired and emulated Claude Shannon to mention that it was Aaron Wyner who coined the term ``Shannon Theory.''

James L. Massey (ETH, Zurich)
James E. Mazo (Bell Laboratories, Lucent Technologies, Murray Hill)
Jacob Ziv (Technion, Haifa)
October 7, 1997

Publications of Aaron D. Wyner