Re: voiced-position malfunctions?

Hi David,

On 12/08/2026 15:45, David Kastrup wrote:
> I think it does what it is intended to do.

I’d need to read more into the commits behind #3902 to find out what 
Keith was after with his compromise. After quickly reading the 
discussion, I don’t understand at all what it is intended to do.

>    The description is of course
> misworded since the rest direction has no way to prescribe the specified
> sign.  What is meant here obviously is
>
>      The staff position of a voiced Rest, negated if the rest has
>      direction DOWN.
>
> The idea behind that negation obviously is that you can just use a
> Staff-wide \override in order to tell LilyPond how far out of center to
> push rests that aren't in \oneVoice.  This also means that it would
> usually a bad idea to specify a negative value for voiced-position: you
> generally want to push the voice in the direction of, well, its
> direction.
>
> Can you propose a better wording here?

For the current behaviour, seems like it might read

“The staff position of a voiced Rest, rounded to the nearest staff line 
(no ledgers) and inverted if the rest has direction DOWN”.

I don’t see how the “no ledgers” part is anything but a shortcut in 
Keith’s implementation, though.

Best, Simon