Here is your preview of the story.
Everyone has secrets. Sometimes, those secrets have to be communicated to someone else, but there is a risk of the communication coming into the hands of an unintended third party.
Secret communication has three aspects. One is physically protecting the communication from interception while en route or theft after it is received by the intended recipient.
The second is hiding the secret communication so the interceptor doesn't know it's actually there. This is the art of steganography.
The third is encrypting the communication so that even if it's intercepted, and recognized as a communication, it cannot be decrypted within the time period during which it is important to maintain the secrecy of the communication. This is the art of cryptography.
A communication may be transmitted by radio, wire (phone, telegram, telex, and fax), visual signals, or hand (carrier pigeon, official mail, or private courier).
By their nature, radio transmissions are more readily intercepted than other modes of communication. The normal radio transmitter antenna broadcasts radio signals in every direction, and the higher the antenna and the more powerful the transmitter, the further they go. As a result of transmission just under the surface of the Earth (ground wave) and reflection from the ionosphere, they can even travel beyond line-of-sight.
The security of radio transmissions can be improved by using a relatively weak transmitter (if the recipient is close enough) or a highly directional antenna (emitting toward the recipient), transmitting on a frequency band that the enemy's radios cannot tune to or at least that the enemy doesn't expect is in use, burst transmissions, frequency hopping in a previously agreed manner, etc.
Wire communications are susceptible to tapping, so it is desirable to have a method of electronically detecting the tap, and also to periodically inspect the line for taps.
Messages being transmitted by hand may be heavily guarded or physically concealed—e.g., in secret compartments.
If we could provide absolute physical security during transmission and storage, there would be no need for steganography and cryptography. But we can't. Each person with access to the message might be compromised by bribery or blackmail. An intended recipient might be impersonated by an opponent. And messages can be stolen or captured.
Invisible inks. It is possible to write a message using an ink that, after drying or perhaps some chemical treatment, becomes colorless. Typically, the writing is revealed by either heating or by reaction with a particular chemical. The first description of an invisible ink is in Pliny the Elder's Natural History and several inks were known to the down-timers.
Microdots. A microdot, developed in the 1920s, was an extreme photographic reduction that was cut out from the emulsion, transferred with a syringe, and affixed by adhesive over a period or other masking character in an innocent message.
Semagrams are graphical concealments. Meaning is conveyed by the position, arrangement, shape or color of objects that constitute part of some larger scene or image. For example, Baden-Powell hid maps of Boer artillery emplacements as markings on the wings of butterflies he drew.
Textual concealment. In 1605, Francis Bacon devised a method of concealment generally called the Baconian cipher. It requires that some distinguishing feature of the characters or words of a decoy message have two possible appearances. If so, then one can think of each such feature as a bit, and each letter of the secret message is encrypted by five such bits. What sort of feature? It could be the size of the letter (small vs. large), the slant (forward vs. backward), the spacing between letters (small vs. large), the number of letters in a word (odd vs. even) and so on. Too obvious, and a Baconian concealment may be suspected. Too subtle, and the recipient may be unable to read the entire secret message. Note also that this method, if encoded on a character by character basis, requires a decoy message five times as long as the secret message.
A grille cipher (devised by Girolamo Cardano in 1550) is really a steganographic system. It requires a grille, a sheet with windows cut at specific locations. The sender and recipient must have identical grilles. The grill is laid over blank paper, the message (signal) is written into successive windows of the grille, and the intervening blank space is filled with irrelevant text (noise) to make it appear that no encrypted message is present. Again, the signal-to-noise ratio is low.
In a null cipher most of the message is irrelevant text and the true message is picked out by the position of the letters or words, relative to the beginning of the message or relative to some sort of signal. For example, the message could be the first letter of each word, every third word, every second word after a comma, and so forth. A null cipher message sent during the English Civil War told a prisoner "panel at east end of chapel slides" (Wrixon 405).
The art in textual concealment is in wording or formatting the overt message without giving rise to suspicion that a covert message is concealed therein.
Cryptography covers the methods by which a sender encrypts a message, and the intended recipient decrypts it. A message consists of units (bits, individual characters, fixed length groups of characters, words, or phrases) that are individually manipulated (replaced or permuted) by the cryptosystem.
Strictly speaking, a code is a cryptosystem in which the principal unit is semantic—a word or phrase—and individual letters are encoded only when there is no code for the word in which they are included.
In contrast, a cipher is a cryptosystem in which the principal unit is the individual letter or fixed length group of characters.
That said, in the past, in popular literature, and even in some modern government circles, the terms code and cipher are often used interchangeably. Code and cipher systems are discussed in more detail below.
A nomenclator was a cryptosystem in which a code book with a relatively limited number of codes was combined with a cipher system.
If a message is first encoded (using a codebook), and then enciphered, the latter is called a superencipherment, and this is a special case of a double encryption.
Cryptanalysis is the decryption of an encrypted message by one other than the intended recipient. Cryptology encompasses cryptography and cryptanalysis.
Most cipher cryptosystems and a few code cryptosystems can be described as algorithms (text manipulation processes) controlled by a parameter, the "key," whose value alters the output produced by the algorithm for a given input. The key itself might be a plurality of individual values (subkeys) that act together to completely control the encryption process.
If the cryptosystem is known, then the encrypted message may be attacked by "brute force," that is, testing all possible keys. Thus, one criterion for a good cryptosytem is that it have a large enough keyspace to resist brute force attack during the time frame in which secrecy is important. How large a keyspace is needed depends on the computing power available to the adversary.
It is important to recognize that a large keyspace isn't enough by itself. A cryptosystem may be susceptible to a more targeted attack.
The worst case scenario for the cryptanalyst is that in which the only material available is the ciphertext itself, and thus the cryptanalyst is dependent on statistical analysis of the ciphertext to reconstruct the message (and key), and if need be the cryptosystem.
In a known plaintext attack, the cryptanalyst is aware that the acquired ciphertext somewhere includes the known plaintext (crib). (A "probable word" attack is merely a weak form of the "known plaintext" attack.) Known plaintext attacks are common against official communications, which frequently contain stereotyped language, especially at the beginning or end of a message. A known plaintext attack might also be launched if an opponent makes the mistake of encrypting the same message in two different cryptosystems, one of which has already been compromised.
In a chosen plaintext attack, the cryptanalyst has the opportunity to request that chosen plaintext (e.g., a string of bytes with a zero value) be encrypted and then compare it to the enciphered text. This is common when the encryption software is publicly available, so the security of a given message depends entirely on the key. It is also achievable if the cryptanalyst expects that certain types of information published by his own side in the "clear" will be retransmitted by the enemy in cipher, or if particular activities of his own side (e.g., mining a particular harbor) will be so reported.
A side-channel attack assumes that the cryptanalyst is able to monitor one or more physical aspects of an encryption or decryption device, such as an operational characteristic (e.g., timing, power consumption, electromagnetic radiation, sound production) or a post-operation residue (data not effectively deleted). In general, this requires at least proximity to and possibly even tampering with the device.
In 1883, Auguste Kerchoffs formulated six maxims for the design of ciphers for military use. I have updated them for our present needs:
(1) the system should be, if not theoretically unbreakable, adequate in practice; i.e., able to withstand cryptanalytic attack for the period of time in which the secrecy of the message is important;
(2) its ability to so withstand cryptanalytic attack should not be dependent on the attacker's ignorance of the cryptosystem; the attacker should be presumed to know the cryptosystem but not the key (a.k.a. Kerchoffs' Principle);
(3) it should be easy to change the key and to communicate and remember the key:
(4) the ciphertext should be transmissible by the preferred mode of communication for the traffic in question (typically telegraph or radio, sometimes computer networks), with minimal errors.
(5) if the sender or recipient will need to move around, the apparatus for encryption and decryption should be portable and it should be possible for a single person to handle and operate it. (Portability is important for some correspondents (spies, mobile military units) but not others (embassies, fixed military headquarters).)
(6) originally stated as "the system should be easy, requiring neither the knowledge of a long list of rules nor mental strain." Of course, apparatus may take some of the load off the operator, so we could say that the key is easily memorized (or securely stored), and methods of encipherment and decipherment are simple enough (from the point of view of the operator, given the apparatus provided and the expected conditions of use) that the error rate is tolerable.
(Konheim 7; ACA).
The second maxim, Kirchoff's principle, deserves special emphasis. One of the basic assumptions of modern cryptanalysis is that the encryption must be resistant to cryptanalysis even if the cryptosystem is known, i.e., only the key is secret. While the underlying reasoning is rarely articulated, one must bear in mind that the greater the encrypted traffic volume, the more likely it is that the cryptosystem, and even some keys and plaintexts, will be exposed by burglary, bribery, or blackmail, or outright capture of spies, cipher clerks, and encryption equipment.
In 1949, Shannon suggested that a cryptosystem should create both confusion (depending on the key, there should be many different possible outputs for each input) and diffusion (a change of a single bit in the input should change about half the bits in the ciphertext).
Note however that while diffusion renders the message more resistant to cryptanalysis, high diffusion means that a garble in the ciphertext will render a large part of it undecipherable. That, in turn, may prompt a call for the message to be resent in whole or in part and if the message is resent with slightly different wording but the same key then you have created a cryptanalytic vulnerability.
Cryptographic Literature in Grantville
There are brief essays about various aspects of cryptography in the various encyclopedias available in Grantville.
Insofar as more specialized works, there is surprisingly little to be found if one searches the Marion County library catalogue with limitation to North Marion High School, Mannington Middle School, Blackshere Elementary School, and Mannington Public Library. Most of the relevant holdings are post-RoF. (It is of course possible that the pre-Ring of Fire holdings were withdrawn as new titles were acquired.) The high school has Russell, The Secret War (Time-Life, 1981). There may be relevant passages in more general works on history (especially military history), or books on recreational mathematics. There is also fiction in which cryptography is discussed, most notably Stephenson's Cryptonomicon.
One can only guess what works are in someone's personal library. Dover has published several inexpensive recreational works over the years, including Laurence Smith's Cryptography: The Science of Secret Writing (1955); Martin Gardner's Codes, Ciphers, and Secret Writing (1972), Norma Gleason's Fun With Codes and Ciphers Workbook (1987), Bud Johnson's Break the Code: Cryptography for Beginners (1997) , and even Helen Gaines' Cryptanalysis (1956). Kahn's The Codebreakers (1967) is also a good bet although it is considerably more expensive. Other possibilities include Pratt's Secret and Urgent (1947) , Zim's Codes and Secret Writing (1971) (juvenile text), and Foster, Cryptanalysis for Microcomputers (1982).
Canon says that David Bartley had a lifelong interest in cryptography, and he not only might have one or more of these books, he might have been a member of the American Cryptogram Association, and thus have copies of its publications.
There are a half-dozen up-timers with degrees in computer science, and at least one of them likely to have Knuth's SemiNumerical Algorithms, which addresses pseudorandom number generation. Also possible, albeit less likely, are books on data security that address the creation, and not merely use, of cryptosystems.
There are seventeen up-timers with degrees in statistics, mathematics, or mathematics education, and these may have books on probability, statistics, and number theory. For that matter, up-timers with degrees in the sciences, engineering, or even economics may have books on statistics.
Down-Time Cryptographers Alive as of the Ring of Fire
Antoine Rossignol (1600-1682) began his cryptographic career in 1626. He was a protégé of Richelieu. He is an impressively successful cryptanalyst as well as an innovator in codemaking.
John Wallis (1616-1703) was a mathematician as well as a cryptographer. Strictly speaking, he was not a cryptographer at the time of the Ring of Fire, he came into prominence during the English Civil War. The same is true of the even younger Samuel Morland (1625-1695). He also became an inventor of mechanical devices.