A possible way to break down Voynich text

[ReneZ]

I think what Thomas is doing is highly interesting.

I don't think that the method should be required to achieve:
- 100% coverage of  all possible character combinations in the MS
- a set of base combinations restricted to the size of the Latin alphabet (say: less than 30).

Apart from the possibility of errors in the Voynich MS text, we have also to keep in mind that:

- *If* it is some kind of en encoding of a meaningful text
- And *if*  this meaningful text was written down as a source for the Voynich MS text

then this source text was not printed (with very high probability) but was handwritten.
It would have had variations in writing, possibly with capitals, ligatures and abbreviations, and not necessarily using the Latin alphabet.

An example:
cases where q is not followed by o are relatively rare, but probably not negligible. Since the apparent rule that q has to be followed by o seems so strict and is used so frequently, the exceptions (e.g. qy and qk) are hard to understand, and likely to have some origin.