If a few citations of author specific characters are enough are sufficient for encoding I have a few more characters to propose
Note : I don't know which I really prefer (encode this kind of rare characters or not).
I prefer to see proposals for such characters. Only by having more proposals can we come to reasonable treatment of the class of such characters.
For characters in scholarly use, if there is a demonstrated need to cite them on an ongoing basis, that would be an argument for adding them. It would establish evidence that they are not *private*.
A mere citation of the form: "look what weird things they used to have in their notation" might weigh less than, say lengthy excerpts in that notation.
The same argument applies to lists of alphabet-variations and text reproductions using that alphabet.
A./
