The o's in VMS

[BessAgritianin]

This is very good observation, and I think it connects to a larger question of how the language was designed.
There are many other symbols which are ambiguous glyphs. Like "P", or "r" or "ch" (EVA) etc..

        I believe these variations are not random, but deliberate. They encode additional information which leads to multilayer understanding. 

Now, however there is another point in this direction- manuscript may had been copied by different scribes, some of whom were not initiated to understand the subtle rules, but copied mechanically what they see and understand.
         
         This may explain why the text has not been deciphered- we are not just dealing with a code, but a code which had been corrupted by the copyists.
 This could also explain why EVA transcription system which tries to map to each symbol fixed set of symbol(s) cannot be sufficient and as a result all the statistics is distorted.