INTRODUCTORY—BINARY DIGIT MEASURE OF INFORMATION
What is this "information theory"? Perhaps the best answer is to summarize the specific results which it has produced in the last decade—
(i) "Information is the negative of entropy." This means that the information content (or better the potential information-content) of any waveform, collection of code symbols, or any other pattern can be assessed mathematically by the same process used to define the entropy of such a system.
(ii) In communicating information over a channel which is liable to interference, the maximum possible rate of transmitting information is a function jointly of frequency bandwidth of the channel and of signal-to-noise ratio: between these two quantities there is a specific exchange rate.
(iii). Definite criteria can be laid down for the design of filters to obtain the maximum separation of signal from signal-plus-noise whether in a telecommunication system or in an automatic control system.
Much of this falls within the field of study sometimes called cybernetics; but the emphasis in this book will be mainly on the measurement and transmission of information, whereas cybernetics should imply rather the use of infi&iation to effect certain control actions. In particular, the relevance of the cybernetics of electrical engineering to the control systems of living organisms will not be discussed in this book.
What Kind of "Information" can be Measured?
Under the heading (i) above it was noted that our theory deals with potential rather than actual information-content, and this must be so because "information" in its usual sense refers to ideas which are not amenable to exact mathematical valuation. We are therefore forced to use a convention which implies, for example, that the information content of a book is proportional to the number of words contained between its covers. (One would hardly think so from some of the more...
Origine : Collection Mihadt J. GAZALE
Page d'accueil
Envoyer à un ami
Imprimer