Why and how the text could be Bavarian

[JoJo_Jost]

I believe that the VMS spaces are "honest", in that they were meant to be boundaries between syllables; although they are quite "noisy", since many spaces are spurious or missing. 

Oh, one more thing:

The problem is that the VMS spacing seems to follow several fairly strict rules. We have seven or eight Rules. That can, of course, be made compatible with syllable boundaries - and in that sense it can also roughly fit the Chinese-syllable idea - no question.

But then the ending distribution has to be explained as well. The four most common final blocks cover 81.13% of all token endings, but almost 20% are still formed by other endings. So the system is not simply "a small closed set of syllable endings". It only works if the model also explains why this particular distribution appears.

This is where Occam's Razor becomes relevant for me. I do not mean it as proof! I have no proof. I mean it as a preference between competing structural explanations.

The structure proposed here explains the same spacing rules with one very simple mechanism:

Split the vowel bigrams and place the visible space between the two parts.

That single rule naturally creates strong restrictions at token endings and token beginnings. It also explains why the most common endings are so concentrated, while still allowing a residue of less common endings.

That has a certain charm.