John Cowan wrote,
> Non-characters aren't encoded, they're reserved either for specific > purposes or for any desired purpose. > If it's a specific purpose, it seems like it should either fall under "character" or mark-up. I can understand reserving code points for any desired purpose, such as control characters or escape sequences. These may well differ from application to application. Once a meaning like "INTERLINEAR ANNOTATION ANCHOR" has been assigned to a code point, any application which chooses to use that code point for any other purpose would be at fault. In other words, if these characters are to be "used internally for Japanese Ruby (furigana), etc.", then they ought to be able to be used externally, as well. I understand that having common internal use code points might be considered handy from an implementer's point of view, but suggest that such conventions should be shared among implementers only, and should not be enshrined in a character encoding standard. > > Is there such a thing as a non-character with a specific semantic > > meaning? > > Why not? > Because it seems to be an oxymoron. If it has a specific semantic meaning, then it should be possible to store and exchange it without any loss of meaning. In other words, it's a character and should be so encoded. (Logos and such notwithstanding.) Best regards, James Kass.