Voynich Authorship

[tavie]

- a top cipher developer would know very well that it is a very bad idea to encrypt a very long text with a single method/key

Why would it be a very bad idea?

One reason is that generally in cryptanalysis, the more source material you have to work with, the better your chances are at cracking the cypher.  The more sentences there are, the better your chances of finding repetitions or other patterns that lead to the breakthrough.  The Voynich definitely violates this general principle! 

That said, many successful decipherments have owed more to mistakes in how the cypher system was implemented by its users, and less to weaknesses in the cypher system itself.  If the mind(s) behind the Voynich knew enough about cyphers to design/adapt and use one but lacked practical experience or knowledge of how cyphers are attacked, that is one explanation of why this general principle wasn't followed.  A little knowledge is sometimes more dangerous than no knowledge.  Believing that your cypher is unbreakable can also lead to overconfidence and result in implementation mistakes, such as using it for an extremely long text. 

Alternatively, they may only have wanted to protect the text from cursory snooping and never expected it to be subject to lengthy cryptanalysis.  There are probably a few other reasons we could speculate about.

(Also it is possible that the Voynich uses more than one key, hence the appearance of different languages/dialects)

The differences between the Currier dialects are often centre stage, but what I find most fascinating is how those differences play out against a backdrop of similarities.  This poses interesting implications for any changes to the system, including if there is more than one key.