Re: ties problem

[Werner LEMBERG]

> > % . The algorithm to choose the direction of the ties doesn't work
> > %   properly.  I suggest the following for applying ties
> > %   sequentially from top to bottom:
> > %
> > %     + The topmost tie is always `up'.
> > %
> > %     + If there is a vertical gap to the last note above larger
> > %       than or equal to a fifth (or sixth?), the tie is `up',
> > %       otherwise it is `down'.
> > %
> > %     + The bottommost tie is always `down'.
>
> I added this,

Not quite.  In the attached image, the ties should be (left to right,
top to bottom):

  1. up down down down
  2. up down down down
  3. up down down down
  4. up down up down
  5. up down up down

assuming that a sixth is used as the threshold for flipping the tie
direction.

> but it turns out the old algorithm wasn't executed due to <too
> complex>.  Can try both of them (see tie-column.cc) and see which
> one is best?

Will try soon.


    Werner