Today, the subject is death; not to worry, not human death but, rather, death of technical fields. In fields of technology, scientists and engineers are always interested in playing "doctor" and quick to assess the pulse and other vital signs of different technical areas. Especially in times of rapid change and funding pressures, they are very sensitive about getting caught with dead fields. As a result, some become voluntary undertakers who compete with each other in being the first to pronounce the final rites on a particular field.
Information Theory and its various subfields have seen their share of solemn health examinations, "final" pronouncements, even burials, and --- resurrections. Of course, as there are many watchful eyes that constantly scrutinize each area's vibrancy and health, there are also many that are eager to detect the birth of new fields or the possible rebirth of those declared dead. Unlike people, technical areas do in fact experience the phenomenon of rebirth or, if you will, resurrection.
As soon as our field emerged, it attracted the attention of many well-wishers, of the curious, and of some with the "evil eye." In fact, the enthusiasm and the hype of the early days gave rise to a litany of tributes and exaggerations that eventually resulted in Claude Shannon, himself, disavowing many of the claims of those who saw Information Theory as a scientific panacea that could be applied to everything from Physics to Biology and from Economics to History! In his famous article*, he cautioned the community to study carefully the newborn field and to be realistic about its usefulness and applicability. In a future column, we will review some of the early views on the breadth of Information Theory that were expansive and optimistic, and, sometimes amusing and/or wrong. The result of the founder's disavowal was a cooling of public attitudes and, in some cases, disappointment and even bitterness for having been vainly seduced. Thus, in the sixties, some people started ringing the bells of doom and decided that the field was turning pallid and flat. Where were the applications that Information Theory promised? Why should the world stay in the closet captive of a field that did not deliver on its promise? So, just as I was beginning my graduate studies I heard from several scientists and colleagues that Information Theory was deceased!
Undoubtedly there are people out there who may have encountered "grim reapers" of Information Theory earlier then I did. But in the late sixties it came to my attention that I was embarking on the study of a dead field! Not wanting to be accused of scientific necrophilia, I tried to obtain --- second opinions. To my relief I found many people (among them several of my teachers) who cautioned that death can be sometimes in the eye of the beholder. It was especially gratifying to hear at the 1972 ISIT in Monterey that new families of codes were constructed that were asymptotically good and that, therefore, the patient was not only alive and well, but also stirring with energy. In the early seventies, universal compression started being talked about and robust detection was developed by those who Sakrison affectionately referred to as the "det-est-ables." The decade of the seventies was definitely a period of intense life for the field. Euphoria set in once again and people started expanding the use of Information Theory into multi-user channels, cryptography, optical communications, and, even, system theory and signal processing. With mathematical certainty, the new wave of enthusiasm started moderating in the early eighties. Some problems were proving too stubborn and progress was not a fast as expected. And, despite the fact that coded modulation and magnetic recording germinated in those years and did eventually lead to techniques that almost achieved the capacity of the telephone channel and truly enabled the application of coding theory (respectively), the nagging "nay"-sayers of nihilism (to paraphrase a man whose long shadow was primarily due to a setting sun) announced once again the death of Information Theory.
As fate had it, I experienced that second death of the field just as I became President of the Information Theory Society. How was I ending up once again associated with an entity whose health was being questioned? It was a chilling moment. In 1987, during my tenure as President and in a year to be known subsequently as the dawn of one of Information Theory's golden periods, (indeed!) I happened to be participating in a meeting of the Board of Governors of the Control Systems Society (to which I had been appointed primarily because I had provided valuable service to that Society in the previous year). One of the officers of that Board was presenting a "state-of-the-Society" review for Control Systems. He admonished the members of the Board to endorse some of his proposals concerning the direction of the Society's transactions and conferences so that it wouldn't end up, and I am now quoting, "like Information Theory, which, let's face it, is now dead"(!).
I can assure you that at that moment I fulfilled the expectations of those who elected me to be President. I stood up and provided living proof that at least the members of the Information Theory Society were far from being dead. Essentially I said that the day of that pronouncement would live in infamy and that the future would vindicate us. Nonetheless, there I was, defending a field that some people genuinely believed to be dead.
The rest is, well, --- history. We are once again in the middle of a cycle of vibrant life of Information Theory. Attendance at our meetings is setting records, the number of proposals for workshops is exceeding --- capacity, the Transactions is getting better and "fatter," the Society continues to be in excellent financial health (certainly from that point of view the evidence of liveliness is overwhelming), and the fields of multi use detection and queuing networks, among others, are feeling the hot touch of Information Theory.
Out there, on the horizon, one can see already dark figures who are once again agitating about the needs of society, the status of technology (what with the growth of the Internet and the ubiquitousness of wireless), the reduction in funding, the tough competition, the pressures on curricula etc., etc. Undoubtedly, one of these days we will hear again that our field is dead. But having already died so many deaths, it is assured of perpetual life!
* "The Bandwagon," {\em IEEE Transactions on Information Theory}, Vol. 2, March 1956.