S TANDARDS Nos. 7 and 8 in the list of May 28 are based on the assumption that—other things being equal—the relative value of a cipher depends upon the degree of security it affords; and that, consequently, a cipher offering absolute protection is the most desirable.
This idea is expressed, in one way or another, by a number of authors on cryptography. Bacon— 1605 —for example, says that ciphers should "bee impossible to decypher." Kerckhoffs—1883—referring to military systems, says that they should be "materially if not mathematically undecipherable."
Bazeries—1901—states that military ciphers should be absolutely undecipherable. And Myskowski—1902—says that ciphers should be practically and also mathematically undecipherable. Delastelle and Myskowski go into some detail, however, on this point, and their lead was followed in compiling the above mentioned standards.
Delastelle says that the comparison of a cryptogram and its translation should not permit the discovery of the key, nor the decipherment of an untranslated part of a dispatch of which the rest is translated.
Myskowski goes even further, contending that, even with knowledge of the system, the comparison of the clear and enciphered texts should not lead to a discovery of the key. And he appends two interesting illustrations where the use of a cipher not affording this guarantee might result seriously.
In the first case he explains how a newspaper, in publishing the translation of material telegraphed to it in cipher by a representative, would afford telegraph officials the opportunity to discover the key by comparing the text in its enciphered and deciphered forms, and consequently to decipher other communications in the same key.
In the other instance he shows how the official publication by a government of the clear text of cipher communications received by telegraph from its ambassador may, in the same way, be equivalent to surrendering the secret of its diplomatic cipher to the foreign government.
Discovery of the key may be rendered less liable in such cases by paraphrasing the clear text. And this should generally be resorted to when there is any danger of comparison.
United States War Department regulations require the paraphrasing of any message which must be filed away with its encoded or enciphered copy, recommending for this purpose change of diction without change of meaning; altering the positions of subject, predicate, and modifying phrases and clauses in the sentence; altering the order of sentences in the message; and deletion rather than expansion of the wording, since the original message could be more readily approximated from the latter by an expert.
Ciphers not requiring such precautions would obviously possess quite an advantage. Hence, standard No. 8. Decipherment can sometimes be facilitated by having the same message enciphered by different keys or ciphers. Cipher No. 42, herewith, is one instance of this kind.
Now for the answers to last week's ciphers. No. 39 used a twenty-four-letter simple substitution alphabet combining I-J and U-V, in which the letters of the key word HELP were used as substitutes for the first letters in each group of six, the remaining letters being in reverse alphabetic al order. The message: "Hit him again, he has no friends." The alphabet:
ABCDEF GHIKLM NOPQRS TUWXYZ HFDCBA ENMKIG LTSRQO PZYXWU
No. 40 used the key word FITZGERALD in conveying the message:
"The Moving Finger writes; and having writ, Moves on; nor all your Piety nor Wit Shall lure it back to cancel half a Line, Nor all your Tears wash out a Word of it."
—Rubayiat, LXXl
Probably the most direct method of determining the period in the above example would be by trying commonly used words with the short cipher groups, according to the method described in FLYNN'S WEEKLY for February 20, 1926.
Thus, the assumption that OZL stands for THE would provide the key letters FIT, which, in turn, would indicate a twelve-letter key by deciphering probable message sequences at that interval throughout the cryptogram. Expanding these by context would similarly provide additional key letters; and so on until decipherment had been completed.
The solution to the key-phrase cipher, No. 41, will be published in two weeks. As explained last week, this will give the fans a better chance to decipher it.
Taking up this week's ciphers, No. 42 should afford an interesting e.xperiment with key phrases. In a key-phrase cipher, a twenty-six-letter phrase or sentence is chosen as key. Using the key phrase, "Better silence than ill speech," for example, A becomes B in cipher, B=E, C=T, D=T, and so on; any letter repeated in the key phrase acting as substitute for a corresponding number of different letters in the message alphabet.
A short cryptogram of this kind might thus be rather difficult to solve. But in the present case the same message has been enciphered with two different key phrases, with the result that you have two cryptograms (a) and (b) to work on instead of only one. After you have deciphered the message, try to reconstruct the key phrases. They are both in English.
No. 44 can be done; but, as ciphers go in this column, it presents a problem of the first magnitude. Our correspondent gives both the cryptogram and its translation, and yet challenges the reader to discover the cipher system.
To make this problem still more interesting, the complete explanation of the cipher, but without the key used in this case, will be published in FLYNN'S WEEKLY DETECTIVE FICTION for August 6. If you are unable to dig out the construction principle in the meantime, you should try to find the key after studying the method in the abovementioned forthcoming issue.
CIPHER No. 42.
(a) VEYYO EENV TPG OER ENYY
EDE HNEGV DNEE EPO VEEPO
(b) EMSLE SNOE TOR CMI IOSS
SEM HOKRE EOSN MOE ESMOE
CIPHER No. 43 (P. A. Napier, Louisville, Kentucky).
WECAM GTEKC MPHEE YEPOM AUEMR EMSEA TEAEM CPEAA MTSEG TEPHM FEMJS MAEMK CMORE CEMQT ECSMA EMJRE CVERI M5CMC EMAME DEEBN MGAMK RMBEE COMFN MFTEK HMLEM HTESR MGAME IMFLM COECF MMTEE HMEEM SCMLH EBIMK EMCFE ASEMS EROEA NME
CIPHER No. 44 (J. McA. T., New Haven, Connecticut).
VZPOD BWEZL WBTLO ACXPC PRHGE GBVXN QMMMD ZYXTF RCGKS AVIYE OAEXY ICGHD IBHUK KLDLO IUICT BALAE BWEQS NZFET WIDLU QVIFI UOHDT TAFXK SMYDS CHGKP ZBXBD FKVHP FAQTP ZMETN SBVHN IANKV RMWVD ACCEP ZKMDQ YIOFL OVDFI UPNWV FQAUM MZQQP OZDKY KNBLX MSQSW IYHBE KHDKU
Message: I wondered whether the simple and direct method of encipherment here used is readily discoverable. I think not. The system is not of the multiple alphabet type. It does not require elaborate tables, and it does not employ computations of any kind. It is quick and easy.
Some fine original ciphers are being submitted to this department. But to avoid delay in publishing them we would caution readers to accompany their contributions with full solutions.
The solutions and full explanations of ciphers Nos. 42 and 43 will be published next week.
We have just received from "Primrose" the long awaited list of winners in the National Puzzlers' League "crypt" contest, published in FLYNN'S WEEKLY for March 19. Here it is, showing the cryptograms solved, and the prizes won, by each contestant:
Winners of "Real Puzzles."
Winners of "Enigma" subscriptions.
Solutions were also received from the following members of the league, not entitled to prizes:
Many readers who tried these contest "crypts" and failed may, even at this late date, be interested to learn how others tried and succeeded. Some of the best ideas from the above solutions are therefore appended.
H. B. McPherrin assumed that "qaboggflk" was likely material in No. 1 for an ING termination. Substituting in "skmp" gave -G--, probably OGLE or UGLY; the latter supposition giving -Y---IC, evidently PYRRHIC, for the next to last group.
Wm. M. Ecklund solved No. 2 by listing the three-letter terminations of the longer words. After careful analysis he had EAD, IDE, RON, NER, and EUR. Substituting through the short words INAPT and OPEN, he obtained POSTSCRIPT and PORTRAYS.
No. 3 proved the most difficult of the three. Rufus T. Strohm solved it guessing the second word, FRENCH. Mr. Ecklund took the last four letters of SHYLOCK, and substituted through FOLK, GULF, UNCLE, and FRENCH.
We congratulate the winners upon their success. The crypts were difficult, and they deserve a lot of credit. We also take this opportunity to present the list of solvers of the code messages, published in the April 23 issue of FLYNN'S WEEKLY:
Some of these solvers observed that the keys to code messages Nos. 1 and 2 were the equivalents of "By United Press" and "Doubloons," respectfully, in Simplex code. Messrs. Bellamy and Shea also submitted the following short key to No. 3:
A E I O U
B D G J L N R T W Y
0 1 2 3 4 5 6 7 8 9
C F H K M P S V X Z
A E I O U
Solutions to ciphers in other issues have also been arriving at a lively rate. The autokey cipher of May 14, especially, stirred up a lot of interest.
Keep your answers coming!
"Solving Cipher Secrets" has now been a regular feature for some time, and every week it seems to create new cipher fans. A definite policy of mixing the easy ones with the hard cryptograms appears to please everybody.