While Emily Dickinson wrote that:

Much madness is divinest Sense

To the discerning Eye. . .

the problem lies in the discernment. Distinguishing meaningful utterances from nonsense is not a trivial task. Confronted with a lengthy text in an unknown script, how does one determine whether those characters in fact contained a meaningful text, or were simply set using the equivalent of printer's pi or a lorem ipsum style text?

The problem is important in cryptography and other intelligence fields, where it is important to distinguish signal from noise. Cryptanalysts have devised algorithms for this purpose, to determine whether a given text is in fact nonsense or not. These algorithms typically analyse the presence of repetitions and redundancy in a text; in meaningful texts, certain frequently used words -- for example, the, is, and and in a text in the English language -- will occur over and over again. A random scattering of letters, punctuation marks, and spaces will not exhibit these regularities. Zipf's law attempts to state this analysis in the language of mathematics. By contrast, cryptographers typically seek to make their ciphertexts resemble random distributions, to avoid tell-tale repetitions and patterns that may give an opening for cryptanalysis.