Did the scribes copy the letters correctly?

[Jorge_Stolfi]

"1 scribe"? It is more likely that there were several

Well, experts disagree on that point.  And anyway I believe there is an important fact that invalidated their analyses.

most VMS writings look very fluent, nearly elegant.

Well again, I beg to differ.  Character shapes, size, spacing are all extremely irregular.  Not from page to page, but from word to word on the same line.  Not at all like the handwriting of those professional scribes who worked at the "book factories" of the time.

If somebody tries to copy unknown or barely known characters, he may "read" them as "the strange 9", "the tailed m" etc. -- but is he enabled then to write them down as accurately and unshakened exactly as we see them? ... Each sign will be a struggle

I don't believe in the theory of a copyist who worked from a "cold start".  But, in its defense, consider what we did over the last 40 years.  We looked at this completely original and unknown script for a while, decided what was the alphabet, and then proceeded to copy transcribe it (digitally) assuming that alphabet.  For each glyph we had to decide whether it was an r or an s, an a or a y, two words or one, a glyph or a dirt smudge... Now and then we ran into a "weirdo" -- a glyph that was too unlike any of our canonical shapes; then we choose a special code for it, and forged on.   It was indeed slow and tiring work -- but we did copy the whole book.  Multiple times.

I think these scribes ...were "at home" within the source language and well educated/trained in this character set.

Agree with the second part, but it does not imply the first part.  People who transcribe Etruscan or Linear A have a good knowledge of the alphabet, but no idea of the language.  At one point I had memorized the Phaistos Disk alphabet myself...

I believe that the Scribe (or "each" Scribe, if you will) was recruited by the Author himself and trained by him on the alphabet.  The Scribe had to practice copying it until the Author was satisfied with the result.  But that probably happened in Europe, and the Scribe was some local guy who was not a professional book copyist but had some writing experience -- a student, a small-town notary, lawyer, or accountant, a letter-writer, ...  There was no need to teach him the language (or the encryption, or the gibberish-generation algorithm)...

When VMS turns out (again) to be unexplainable, this is the last refugee for many theorists: to assume that the VMS authors made massive, bizarre and numerous mistakes...

I believe that indeed the transcriptions that we use contain a large number of errors, due to various causes (including  some causes which most Voynicheers refuse to recognize).  Maybe as many as one error every 10 words or more.  And most of the apparent errors are not bizarre, but can be explained by similarity of shapes and other likely excuses.

Yet, a high rate of errors would not make the solution impossible. It would only frustrate some simplistic approaches, and would require a lot more work to find (and prove correct) than a typical cryptographic puzzle would. 

All the best, --stolfi