Re: Spacing oddities

[Carl Sorensen]

On 11/11/11 9:19 PM, "Keith OHara" <address@hidden> wrote:

>Carl Sorensen <c_sorensen <at> byu.edu> writes:
>
>> I have a problem with understanding how [vertical] spacing works
>> in the attached file.
>
>> Page 1 has 32.76 space left, according to annotate-spacing.  The first
>> system on page 2 has an extent-estimate of (-23.98 . 0).  So why won't
>>it
>> fit on the first page?
>> 
>The page-breaker does not know that the 32.76 space will be there, because
>at that stage LilyPond has not yet discovered that the upper staves can
>take less space than the sum of their extent-estimates.

Ahh, this is the point that I had missed.  Although at this point, I can't
see any extent-estimates overlapping (but I'm missing one system's extent
estimate.

>
>I have been drifting toward the opinion that combining the line-breaking
>and page-breaking into one optimization step is too clever by half.
> 
>
>> Why are the extent-estimates for the first and last systems on page 2 so
>> much larger than the visual extent?
>> 
>
>I can only guess that it has something to do with the up-bow and fingering
>and such things on top of beamed notes.

OK, I'll do some investigation in this area.

>
>(I notice that your spacing overrides, trying to pull the staves together,
> are in a tug-of-war with ragged-bottom=##f, trying to stretch the staves
> to fill the page.  I guess this is just an experiment to try to convince
> LilyPond to put more systems on a page.)

Yes.  And if I use ragged-bottom = ##t, I don't get more systems on a
page, I just get systems closer together, and lots of white on the bottom.

Thanks for the explanations -- I think that in order to get things
tighter, I'll need to figure out how the extent-estimate gets so big.

Thanks,

Carl