(30-06-2025, 07:03 AM)Eiríkur wrote: You are not allowed to view links. Register or Login to view.It's easy to count things, it's not easy reverse-engineer this process-
A possible research topic would be to find out just how much the patterns for selection of sources and rules for evolving words can be restricted and still be able to generate all of the text (or a high percentage, if we ignore some errors, a few unreadable or unidentifiable glyphs where the scribes went crazy with boredom and just had to innovate).
The problem with rather lax patterns and evolution rules is that it is impossible to tell if a possible source-target match (even when there are several words in a pattern) is coincidental or not.
For example, to generate an entire line, a possible restriction would be to allow all sequential transfers (with any number of skips) from a maximum of two lines of the same page with a maximum of two differences per word in one chunk:
103v.3: qokeeor.chedy.qokey.dar.checthy.chor.qoty.shdy.okeedy.qokeey.qokain
A B C D
103v.6: dain.shey.qokeedy.cheol.qoeeor.lshor.qoky.shedy.qokaiin.chedy.qokam
1 2 A 3 4 B C 5 6 7 D
103v.7: daiin.shey.chol.chey.oteey.lkeeor.okaiin.shedy.shedy.qokaiin.ol.chedydy
1 2 3 4 5 5 6 7
It probably won't work with these limitations but we might be able to identify others that do. We need testable hypotheses to progress. If there are no strict rules we might be able to identify the preferences of each scribe.