Re: cluster building with ties

[Kieren Richard MacMillan]

Hello, David:

> Your tie is no-win.  You seem to be trying to indicate which time
> value on the shared notehead is originating the tie by the tie
> direction, but tie direction indicates nothing at all.

Actually, I'm not (consciously) using direction to indicate anything. 
However, of all the possible notation options I tried -- multiple 
voices with rests, all eighth notes with multiple ties, and every 
imaginable combinations thereof -- the one I settled on is the only one 
(IMO) that properly shows the specific voice-leading that I intend. 
(n.b. More accurate/specific yet would be to individually slur the top 
and bottom voices, but that ended up looking too "grobby"...)

As another possibility, is it possible to automatically "mask" the part 
of the tie that passes "through" (behind) the notehead, like so:

tieMask.pdf
Description: Adobe PDF document

[detail from original image, manually adjusted in a graphics program]

Although rare, I have seen examples of this in scores from almost every 
engraving house you can name. This is what should happen when a tie or 
slur passes "through" a time signature (does this happen in Lilypond?), 
so why not for a note head?

It would be best if one could apply a setting which indicated which 
objects would force a mask (e.g., time signatures, note heads) and 
which objects wouldn't (e.g., stems, staff lines), and of course the 
mask width.

Is this something I can add to the slur-engraver code? Happy to dive in 
myself, if necessary.

Best wishes,
Kieren.