This is a rather long-term goal probably, but the basic idea is:
An optional mode where, if the user types in a character with a diacritic unavailable in the current encoding, the application tries to match it using X/26 enhancements.
For example, for a character "ĉ", used in Esperanto, an ASCII "c" is combined with the "G0 character diacritical 3", and so on.