Re: Horizontal spacing depends on stem 'direction

[David Kastrup]

"Dmytro O. Redchuk" <address@hidden> writes:

> On Thu 08 Sep 2011, 23:57 Trevor Daniels wrote:
>> I'd class this as an engraving nitpick, but we don't have that
>> classification anymore.
>> 
>> It seems the horizontal spacing is slightly dependent on the stem
>> direction.  This can cause a score to mysteriously spread over an
>> extra system when a single note with stem down is changed to one
>> with stem up, if the original system exactly fitted the line width.
>> 
>> The following example demonstrates this by showing the slight
>> difference in the two line lengths.
> Thank you, I've added this as 1912:
> http://code.google.com/p/lilypond/issues/detail?id=1912

In my opinion, this report is invalid.  Since Lilypond does optical
justification of noteheads (cf node "Optical Spacing" in the Lilypond
Essay), of _course_ its spacing depends on the stem direction.  And of
_course_ this implies that there will be a small borderline where page
break decisions _also_ differ depending on the stem direction.

A human engraver will make a random decision, and stick with this
decision even when one stem is flipped around for the next iteration,
changing the spacing slightly.

Lilypond does not remember what it did last time around with that score,
and it does not throw dice.

-- 
David Kastrup