On plain texts and ciphers (a thought experiment)

[-JKP-]

...

Considering examples like Hildegard von Bingen and Giovanni Fontana, I would say that the alphabet was created by a single person in the XV century. The "small community" speaking an "isolated language" still fit the story of course. 
If I remember correctly, both Hildergard's and Fontana's newly created alphabets are documented in more than one manuscript. Voynichese could be another example of alphabet infant mortality.

One thing I noticed about von Bingen's code (not the constructed part, but the alphabet substitution code) is that it's designed for ease of reading.

• b and d are normal except they have an extra flourish (which is easy to ignore) and the d is mirrored
  
• p and q are like b and d in the sense that one is mirrored and the other has an extra flourish
  
• the l (ell), like the b and d, has an extra flourish (which can be ignored)
  
• u is simply slanted
  
• r simply has a long ascender
  
• x has the crossbars offset but still resembles an x
  
• the y is mirrored in the vertical direction
  
• i has two extra ticks that can be ignored, and o similarly has two extra flourishes that can be ignored
  
• f and t are similarly easy to see/remember—the f simply lacks a crossbar and the t is normal except for an extra wiggle in the crossbar
  
• the s is substituted with a z, which is easy to remember because of sound similarity
  

There are only a few letters that are distinctly different. You could learn to read it in 10 minutes and could probably write it from memory in 20.


What makes it a tough code is that the words themselves have been substituted with von Bingen's own constructions. In other words, the shapes are not especially different from Latin glyphs, just as the VMS glyphs are not particularly different from Latin letters/numbers/abbreviations, but the underlying words are not German or Latin, they are Bingenese, and a glossary is needed to keep track of them and discern their meaning.

Even with that, the code was probably not intended to be used for entire manuscripts, since the "secret" (encoded) words in Bingenese, that were found in a fragment, are in the minority, mixed in with Latin. It's very much like diplomatic and personal letters of the time that were mostly readable but had the sensitive parts (gossip, people's names, strategic info, etc.), encoded while not obscuring the rest.