Often in this column I have speculated on the evolution of Information Theory over the last half-century. As the fiftieth anniversary of the founding of the field is fast approaching, it is perhaps appropriate to continue and broaden this speculation. You may recall from previous columns that early-on, barely a few years after the publication of Shannon's famous paper, there was already controversy about the "reach" and the "content" of Information-theoretic concepts.
Of late, with occasional exceptions, the topics that are covered in the transactions and in the presentations at our workshops and symposia seem to center on a few principal coordinates of the field: that is coding theory and techniques, Shannon theory, and Source Coding and Compression. Nobody will doubt for a moment that these are indeed the areas of activity that lie closest to the core of Information Theory and that find increasingly important applications. Of course, once in a while there is an article and/or a session on other topics like complexity, cryptography, communication networks, neural networks, etc. But these have become infrequent and are typically considered by many to be outside the "mainstream".
Is this trend a natural consequence of the growing number of scientists and the increased specialization that encourages the clustering of activities around narrowly defined topics? Perhaps. Is it a healthy trend? I am not sure, but instinct urges me not to think so.
This column is not supposed to be a forum of preaching but rather of reporting critically on the past with the intent to entertain, inform, and, sometimes, encourage useful thinking. So, instead of expanding on the thesis of what should be the scope of the field, I would like to pull from the past the technical program of the 1978 workshop on Information Theory that took place at the Northstar Resort at Lake Tahoe, and let the readers compare it with the contents of more recent workshops and draw their own conclusions.
The year was 1978, just about after the mid-point in the history of the field to date. It was June and there was some truly glorious weather in the mountains of California. There were about sixty-to-seventy attendees and a single track of five half-day sessions. Each session consisted of four long talks. Without further ado, here is the complete program:
Monday, June 26,
Morning Session SOURCE CODING WITH APPLICATIONS TO SPEECH
Organizer: Robert M. Gray, Stanford University
1) A.H. Gray, Jr., University of California at Santa Barbara LPC Speech Compression and Speech Distortion Measures
2) John B. Anderson, McMaster University Sequential Source Encoding of Speech Waveforms
3) Steve Wilson, University of Virginia Tree-Searched ADPCM
4) Robert M. Gray, Stanford University "Fake Process" Speech Compression--Is there a rate distortion theory for speech?
Monday, June 26, Afternoon Session
APPLICATIONS OF DETECTION AND ESTIMATION THEORY TO OPTICS
Organizer: Joseph Goodman, Stanford University
1) Joseph Goodman, Stanford University Introduction
2) Richard Hudgins, Itek Corporation Estimation in Adaptive Optics
3) Bob Hunt, University of Arizona Detection and Estimation in Image Processing
4) John Walkup, Texas Tech University Estimation in Signal-Dependent Noise
Tuesday, June 27, Morning Session
ASPECTS OF ANTHROPOMORPHY
Organizer: Fred Jelinek, IBM Research
1) Ira Pohl, University of California at Santa Cruz On Heuristic Search
2) David Huffman, University of California at Santa Cruz Analysis of Polyhedral Scenes
3) D.B. Lenat, Carnegie-Mellon University Automated Theorem Formation in Mathematics
4) Fred Jelinek, IBM Research Self-Organization of Models of Speech Processes from Observed Data
Tuesday, June 27, Afternoon Session
MULTIUSER CHANNELS AND ALGEBRAIC CODING
Organizer: Jack K. Wolf, University of Massachusetts
1) Hisashi Kobayashi, IBM Research Modeling of Satellite Multiuser Communication Systems
2) Pierre Humblet, MIT Efficient Addressing in Multiuser Communication Systems
3) Thomas M. Cover, Stanford University Information Theory for Multiuser Communication Systems
4) Jack K. Wolf, University of Massachusetts, Amherst Error Control for Multiuser Communication Systems
Wednesday, June 28, Morning Session
CRYPTOGRAPHY AND COMPLEXITY THEORY
Organizers: Peter Elias and Ronald Rivest, MIT
1) Nicholas Pippenger, IBM Research Information Theoretic Arguments in Combinatorial Problems
2) Leonard M. Adleman, Garry Miller, and Ronald L. Rivest, MIT Cryptography and Computational Complexity
3) Andrew Yao, Stanford University Probabilistic Algorithms
4) Richard Karp, University of California at Berkeley Probabilistic Analysis of Combinatorial Algorithms
As you can see, there was a truly broad (you might say adventurous) range of topics and a rich selection of titles. Instead of focusing on the presentation of their most recent results, the authors dealt with in-depth, retrospective analysis that aimed at educating, at drawing analogies among seemingly diverse fields, and at pointing to new productive directions of research and to the discovery of interesting new problems.
Think what you may about whether such a format is preferable or whether the broad coverage is desirable today. The fact is that these two-and-a-half days were truly enriching and everyone left at the end with a strengthened conviction that this was a "fun" field to work in.
To help the readers transpose themselves mentally to that time and place, and to give them a sense of being there, I would like to add a few "vignettes" from the proceedings at the workshop.
First of all, this was where and when the Board of Governors formally approved the constitution of the Society (then Group) that, as it turned out, languished in the IEEE headquarters for over ten years until subsequent enlightened leaders of the Society took the last few formal steps required by the IEEE bureaucracy to put it formally into place. Of course, in the interim, the Society operated as if it actually had its formal constitution (which of course it did in the most real sense).
This was also where and when Bob Gray introduced the first speaker in his session (the opening one in the workshop), who happened to be his older brother, as "the man I've known all my life"!
Finally, this is where and when John Anderson referred to a method he had used to obtain a result as "brute-force cleverness"!
It was a warm environment in which a few dozen good minds interacted with the usual mixture of seriousness and levity and added one more nugget to the evolution of the field.