Voynichese is a numeric cipher?

[oshfdk]

What are the differences? How different is the text with nulls?

I'm talking about the version with nulls. Just looks very odd. As I said, it's my subjective feeling vs yours.

So first we need to define the "similarity" criterion...

Subjective similarity can be tested without defining it specifically. The easiest would be get one piece of Voynichese produced by the script and 3 pieces of the same length from Currier A and see how easy it is for another person to identify the generated piece without the reference text. If it's easy, then the generated text is not similar to the actual Voynichese. However, this is only a bad proxy for proper statistical testing, which requires the creation of a large ciphertext.

My point here is that when someone claims that some method produces a very good replica of Voynichese, this is based on their subjective assessment which may or may not be justified.