How would XV century ciphers encode numbers?
Numbers in texts can be a good starting point when cracking a cipher, but they can be a challenge when using statistical methods. If numbers are expressed as the Roman numerals using normal letter codes (MDCLXVI), then even 5-10 numbers per page will introduce many unusual bigrams and trigrams, not seen in normal text (XXII, III, CCXX), complicating the deciphering. If numbers are expressed in the modern Hindu-Arabic form with separate codes for 1234567890, the plaintext character set will expand and there will be a separate set of bigrams and trigrams. As far as I understand, sometimes numbers were also encoded in a Hindu-Arabic base 10 scheme, but using plaintext letters instead of digits (say, if we use ABCDEFGHIK, then 1420 could be be ADBK, or BECA). In which case the character set stays the same, but adding combinations of letters not normal for a text, an effect similar to the Roman numerals, but with less repetition (unless there is a sequence of numbers with some increment).
Are there any known examples of ciphers from roughly the Voynich MS era with some numbers present in the plaintext (of some provision for encoding numbers)?
Numbers in texts can be a good starting point when cracking a cipher, but they can be a challenge when using statistical methods. If numbers are expressed as the Roman numerals using normal letter codes (MDCLXVI), then even 5-10 numbers per page will introduce many unusual bigrams and trigrams, not seen in normal text (XXII, III, CCXX), complicating the deciphering. If numbers are expressed in the modern Hindu-Arabic form with separate codes for 1234567890, the plaintext character set will expand and there will be a separate set of bigrams and trigrams. As far as I understand, sometimes numbers were also encoded in a Hindu-Arabic base 10 scheme, but using plaintext letters instead of digits (say, if we use ABCDEFGHIK, then 1420 could be be ADBK, or BECA). In which case the character set stays the same, but adding combinations of letters not normal for a text, an effect similar to the Roman numerals, but with less repetition (unless there is a sequence of numbers with some increment).
Are there any known examples of ciphers from roughly the Voynich MS era with some numbers present in the plaintext (of some provision for encoding numbers)?