When you read this column we have just recently finished ISIT
98, the Golden Jubilee Symposium. I am sure we will be able to
conclude that it was a great success. At the current moment, as
I am writing this letter in mid- July, we are just making the
final preparations. Many people are working hard in order to make
this symposium a very special one. It is, of course,
especially exciting that the Jubilee Symposium could be arranged
at the place which more than any other place is closely tied to
the pioneering achievements of information theory, namely at MIT.
One of the special features of ISIT 98 is the Golden Jubilee Awards. The idea with these one-time awards is to recognize some of the most important of the many scientific and technological contributions from the first 50 years of Information Theory. The awards are divided into two categories. The first category recognizes significant papers that have not been previously awarded. The second category recognizes technological innovations. The lists of awards in both categories are published elsewhere in this Newsletter.
Looking at these lists, one striking observation is how close to each other the theoretical and practical achievements have evolved within our field. Highly abstract mathematical ideas have very rapidly found their ways to practical applications, and new technological achievements have often inspired new theoretical development. It seems to me that one of the salient features of information theory is this close relation between theory and practice.
Another striking observation is the wide range of technology
and ideas. Among the awarded innovations we find Norman Abramson's
random-access communication protocol, Bob Lucky's adaptive equalizer,
the Lempel-Ziv universal data compression algorithm, the Diffie-Hellman
public-key encryption, Ungerboeck's trellis coded modulation,
and the Viterbi algorithm -- just to mention a few of them. As
we are all aware there are many further important contributions
-- both the remaining ones on the list of
Golden Jubilee Awards and many other ones outside this list. Together
they span over a very broad range of technologies, many of which
are of indispensable importance to modern society. The common
feature of these achievements is that they are all based on a
common family of theoretical ideas regarding representation, processing
and utilization of information --
information theory.
The paper awards also span a wide range. Here first of all
we find Golay's classical paper, presenting for the first time
the famous codes which by now are referred to as the Golay codes.
Very seldom are such beautiful mathematical objects discovered.
The paper in itself is extremely short - just half a page. In
contrast the number of subsequent papers it
has inspired is almost endless. As we know now there are beautiful
connections to group theory (Mathieu groups), lattice theory (Leech
lattice), design theory and many other topics. As codes, the Golay
codes are outstanding: they are the only non-trivial multiple-error
correcting codes. In addition to all of this, they are practical:
many efficient algorithms are available for the practical implementation
of Golay codes.
Among the other awarded papers we find Shannon's also by now
classical paper on the zero-error capacity. This beautiful paper
brings together information theory and graph theory and has had
an enormous impact on both
areas.
In order to illustrate the breadth of the ideas involved, let
us mention just a few of the other awarded papers. We have Gallager's
paper on low-density parity-check codes, Kailath's paper on a
general likelihood-formula, Forney's paper on the structural properties
of convolutional codes and the McEliece-Rodemich-Rumsey-Welch
paper on the linear programming bound. All of these -- and of
course also all of the other now and previously awarded papers
-- present fundamental theoretical insight with an enormous impact
on both the theoretical and practical
evolution of telecommunication and related areas.
One might reflect about the interaction between theory and
practice. Which comes first? One view is that theory always comes
first and that practical applications are all based on insights
first gained by theoretical investigations. The opposite view
is that technological evolution arises out of practical needs,
that solutions are normally obtained without any
theory at all and that theory is only an academic game which at
best is able to explain in retrospect certain practical results
already achieved.
I think that experience tells us that neither of these extreme
views is correct and that the relation between theory and practice
is extremely complex. Looking at the two lists of awarded papers
and innovations I think we must conclude that technological evolution
at its best occurs as a mutual interaction between theory and
practice, where new theoretical insights pave the road for new
practical applications and where practical experience and needs
inspire new theory and deeper insights. My personal
view is that the development within information theory during
the 50 years since the publication of the fundamental papers by
Shannon provides beautiful and very convincing support to this
view.