Re: better beam direction algorithm (again)

[Werner LEMBERG]

> > Lilypond still uses a non-optimal beam direction algorithm.  Again
> > I suggest to simply check for the note under the beam which has
> > the greatest distance from the staff's middle line, and to use its
> > natural direction for the whole beam.
> 
> But what happens if there are a lot of those notes with the same
> maximum?

Example, please.

> We'd get a beam with a lot of (too) long stems. Wouldn't it be
> better to take the direction which minimizes the total deviation
> from the standard stemlength?

IMHO it is better to make the beam always stay near the staff lines.


    Werner