Q: Is there a `smart' multi-voice accidental mode?

[David Petrou]

I need to typeset music with multiple voices per staff.  Think 4-part
counterpoint (chorales, fugues, etc.)  I'd like the following
accidental behavior:

- Lily should consider voices on the same staff separate when adding
  accidentals (or naturals).  That is, if one needs an accidental on a
  particular note, and the later, in the same measure, another voice
  needs the same accidental on the same note, then lily should output
  the accidental for both notes.  You should be able to read each
  voice independently of the other voices.

- Exception to considering the voices separate: if a note in one voice
  is accented (or naturalized...), but then another voice in the same
  measure has an unaccented version of that same note, then the note
  needs to have the correct accident or natural or whatever.

Does this make sense?  This is the practice followed by many editions.
Does this fall into `sounds nice David, implement it yourself' or
`wow, there's a burning need for this, we'll do it now.'?  :)

thanks,
david