Discussion of "A possible generating algorithm of the Voynich manuscript"

[ReneZ]

One of my main problems with the method is that it is not very well formulated, and this does not allow testing / verifying it. Just an example: If a recipe for a cake gives a list of ingredients (eggs, flour, etc) and then says: "put all ingredients together and apply heat", then the recipe is too unspecific.

What I tried to get from re-reading is how specific the 'prescription' of the auto-copying is. In the most recent paper, there are suggestions about the type of changes applied, but so far I could not get anything on 'how far back' the author/scribe was looking. I have not looked into the code of the application, but there must clearly be some assumptions in there.

What also seems to be missing is the initialisation. How was it started? This may seem a trivial detail, but again there must be something in the code for that.

It is not written down specifically (and it is again something that I wanted to double-check) but it seems to be implied that every word in the MS (after the initialisation) is the result of auto-copying. That is, there are no words that are 'new seeds' or incidental re-initialisations.
(I have asked earlier in this thread about this, but I think that this question was understood in a different way).
In any case, these points clarify that the initialisation procedure is too important just not to mention.

Then, if this assumption (no new seeds) is true, one could verify the auto-copying hypothesis by checking for each word in the MS if there is a recent (how far back?) similar word (which max. edit distance?) from which it could be derived.
This seems to be the most basic test of the method, but I remember no evidence that this was even attempted. Again, something I would still need to check in the earlier papers. They are very long....