1.03-чтение


 * **жирным выделены слова с неправильным ударением** toc
 * **красные буквы - звук произнесён неправильно или отсутствует**
 * ** подчёркивание - не справляетесь со скоростью: очень невнятное чтение или лишние паузы **


 * Encipherment/Decipherment Complexity**

Казаков
Ошибок совсем мало, но очень плохая слитность: даже артикли прочитаны отдельно. This cipher involves no additional work __beyond arranging the letters__ and reading them off again. Therefore, the algorithm requires a constant amount of work per ch aracter, and the time needed to apply the algorithm is proportional to the length of the message. However, we must also consider the amount of space needed to record or store the ciphertext. So far, th e o ther ciphers we have seen require only a constant amount of space (admittedly up to 262 locations). But the columnar transposition algorithm requires storage for all ch aracters of the message, so the space required is not **constant**; it depends directly on the length of the message. Furthermore, we cannot produce output characters until all the message's characters have been read.
 * Cryptanalysis by Digram Analysis**

Цыкарев
Suppose we want to decrypt a message that has used a columnar transposition for its encryption algorithm. The basic attack on columnar transposition s is not as precise as the attack on substitution ciphers. Even though transpositions look less secure than substitutions, they can in fact be more secure. Transposition s leave the plaintext letters intact, so the work for the cryptanalyst is more ex hau sting; more relies on a human's judgment of what "looks right." The first step in analyzing the transposition is computing the letter frequencies. If we find that in fact all letters appear with their normal frequencies, we can **infer** that a transposition has been performed. Given a string of text, the trick then is to break it into columns. Two different strings of letters from a transposition ciphertext can represent pairs

Нефёдов
The first step in analyzing the transposition is computing the letter fr e quencies. If we find that in fact all letters appear with their normal fr e quencies, we can **infer** that a transposition has been performed. Given a string of text, the trick then is to break it into columns. Two different strings of letters from a transposition ciphertext can **represent** pairs of adj a cent letters from the plaintext. The problem is to find where in the ciphertext a pair of adj a cent columns lies and where the ends of the columns are. We must do an exhaustive comp ariso n of strings of ciphertext. The process compares a block of ciphertext characters against characters successively farther away in the ciphertext.

Болтунов
We must do an exhaustive __comparison__ of strings of ciphertext. The process compares a block of ciphertext characters against characters successively farther away in the ciphertext. To see how this works, imagine a m** o **ving window that locates a block of characters for checking. Assume the block being compared is seven characters. The first comparison is c 1 to c 8, c 2 to c 9, …, c 7 to c 14. Then, we try a distance of eight characters, and so the window of comparison shifts and c 1 is compared to c 9, c 2 to c 10, and continuing. For a block of nine characters, the window shifts again to c 1 against c 10, and so forth. This process is shown in Figure 2-5. неправильноечтениенижнихиндексов Substitution and transposition can be considered as building blocks for encryption. Other techniques can be based on each of them, both of them, or a combination with yet another approach. For instance, Sidebar 2-5
 * Combinations of Approaches**

Берёзкина

 * 2.4. Making "Good" Encryption Algorithms**

So far, the encryption algorithm** s ** we have seen have been trivial, intended prim ari ly to demonstrate the concepts of substitution and permutation. At the same time, we have exam in ed several approaches cryptanalysts use to attack encryption algorithms. Now we exam ine algorithms that are widely used in the commerci al world. Unlike the previous sections, this section does not delve deeply into the details either of the inner workings of an algorithm o r its cryptanalysis. (We save that investigation for Chapter 12.)


 * What Makes a "Secure" Encryption Algorithm?**

There are many kinds of encryption, including many **techniques** beyond th o se we discuss in this book. Suppose you have text to encrypt. How do you choose an encryption algorithm for a particular application?

Макарихина
2. The set of k ey s and __the enciphering__ algorithm should be free from complexity.

This principle implies that we should restrict n ei ther the choice of k ey s nor the types of plaintext on which the algorithm can work. For **instance**, an algorithm that works only on plaintext having an e q ua l number of A's and E's is u s eless. Similarly, it would be difficult to **select** keys such that the sum __of the values of the letters of the k ey __ is a pr i me number. Restrictions such as these __make the use of the encipherment pro hibit ively__ complex. If the process is too complex, it will not be used. Furthermore, the key must be transmitted, stored, and remembered, so it must be short. 3. The implementation of the process should be as simple as possible.

Смирнов
4. E rrors in c i phering should not prop agate and c au se corr u ption of further information in the message.

Principle 4 a c knowledges that humans make e rrors in their use of en ciphe ring algorithms. One error early in the process should not thr__ow__ off the entire remaining ciphertext. For example, dropping one letter in a columnar transposition thro__ws off__ the entire remaining encipherment. Unless the receiver can guess where the letter was dropp e d, the remainder of the message will be **unintelli gible **. By contrast, reading the wrong row or column for a **pol yalphabet ic** substitution affects only one character remaining ch aracters are unaffe cted.

Ярушин
These principles were developed before the ready ava i lability of digital computers, even though Shannon was aware of computers and the computatio__nal power they represented__. Thus, some of the **concerns** he expresse d about hand implementation are not really limitations o n computer-based implementation. For example, a cipher's implementation on a computer need not be simple, __as long as__ the time complexity of the implementation is **toler a ble**.


 * Properties of "Trustworthy" Encryption Systems**

Commercial users have several requ ire ments that must be satisfied when they select an encryption algorithm. Thus, __when we say that__ encryption is "commercial grade," or "trustworthy," we mean that it meets these constraints:

 Берёзкина

There are many kinds of encryption, including many **techniques** beyond th o se we discuss in this book. Suppose you have text to encrypt. How do you choose an encryption algorithm for a particular application?