change defaults for dot spacing in repeat sign to accommodate tab staves

change defaults for dot spacing in repeat sign to accommodate tab staves

follows up issue 2762

[Keith, 2012-09-08 03:12:30]

#(set-global-staff-size 5)
{ R1 \repeat volta 2 R1 }

[Keith, 2012-09-08 18:54:07]

http://codereview.appspot.com/6488097/diff/3001/scm/bar-line.scm
File scm/bar-line.scm (right):

http://codereview.appspot.com/6488097/diff/3001/scm/bar-line.scm#newcode178
scm/bar-line.scm:178: ;; the default distance between centre of dots is composed
of
Have you considered working entirely in staff-positions?
.. placing repeat dots at the first pair of positions with at least a full
staff-position gap for each dot.

Then style changes (tiny, heavy staves for miniature scores, or widely-spaced
tabStaff that really should have fatter repeat-dots) will not change the
position of the dots.

[dak, 2012-09-08 19:11:46]

On 2012/09/08 18:54:07, Keith wrote:
> http://codereview.appspot.com/6488097/diff/3001/scm/bar-line.scm
> File scm/bar-line.scm (right):
> 
> http://codereview.appspot.com/6488097/diff/3001/scm/bar-line.scm#newcode178
> scm/bar-line.scm:178: ;; the default distance between centre of dots is
composed
> of
> Have you considered working entirely in staff-positions?
> .. placing repeat dots at the first pair of positions with at least a full
> staff-position gap for each dot.
> 
> Then style changes (tiny, heavy staves for miniature scores, or widely-spaced
> tabStaff that really should have fatter repeat-dots) will not change the
> position of the dots.

I agree with Keith that ignoring the various thicknesses altogether would likely
make things more predictable.

[benko.pal, 2012-09-08 19:14:15]

ok, thinking about a new patch.

[Keith, 2012-09-08 20:45:55]

I guess you are thinking we bring the dots inside the staff if there is at least
one staff-postion of space for each dot (as in 2-line percussion staves) and
continue to move them closer to the center if we find locations with at least
two-staff-positions of space for each dot, or more space than we have found so
far.

[benko.pal, 2012-09-08 22:26:20]

sorry, I'm confused.
do we want to support
- NR 2.5.1 style 2-line percussion staves (setting both line-count and
staff-space to 2 instead of setting just line-positions to (-2 2))?
- default TabStaff's (even line-count, 1.5 staff-space)?

if we want to support both of those without changing dot size, calculations
cannot be based on line-positions only, staff-space must be considered.  if we
consider staff-space, issue 2648 must be taken care of.

[Keith, 2012-09-08 22:39:01]

On Sat, 08 Sep 2012 15:26:21 -0700, <benko.pal@gmail.com> wrote:

> do we want to support
> - NR 2.5.1 style 2-line percussion staves (setting both line-count and
> staff-space to 2 instead of setting just line-positions to (-2 2))?
> - default TabStaff's (even line-count, 1.5 staff-space)?
>
> if we want to support both of those without changing dot size,
> calculations cannot be based on line-positions only, staff-space must be
> considered.

I suggest you consider both the gaps between line-positions (prefer gap of 2
staff-positions)
and the full range of line-positions (prefer to have repeat dots within the
staff)

If someone writes tablature and wants a staff-space = 2 for some reason, he
probably want the repeat dots in their usual positions. (His choice on whether
to also increase the dot size.)

If someone writes 2-line percussion somewhat more narrow, he will probably want
the repeat dots inside the staff.

>
> http://codereview.appspot.com/6488097/

[benko.pal, 2012-09-09 07:22:44]

2012/9/9 Keith OHara <k-ohara5a5a@oco.net>:
> On Sat, 08 Sep 2012 15:26:21 -0700, <benko.pal@gmail.com> wrote:
>
>> do we want to support
>> - NR 2.5.1 style 2-line percussion staves (setting both line-count and
>> staff-space to 2 instead of setting just line-positions to (-2 2))?
>> - default TabStaff's (even line-count, 1.5 staff-space)?
>>
>> if we want to support both of those without changing dot size,
>> calculations cannot be based on line-positions only, staff-space must be
>> considered.
>
> I suggest you consider both the gaps between line-positions (prefer gap of 2
> staff-positions)
> and the full range of line-positions (prefer to have repeat dots within the
> staff)

I can see two possibilities:
1. special casing two-line staves.
   simple but will fail for staves with a big space in the middle but
otherwise dense.
2. implement a more complicated algorithm.
   the current one looks for a good place
   - starting from the centre
   - going up and down strictly synchronised
   - stopping at the first convenient point.
   instead we can try
   - starting from the outside
   - going inwards very loosely synchronised
   - scoring several paths
   - choosing the best.
   smells like a huge can of worms.

> If someone writes tablature and wants a staff-space = 2 for some reason, he
> probably want the repeat dots in their usual positions. (His choice on
> whether to also increase the dot size.)

is dot size customizable currently?  if so, we must include some regtests.

> If someone writes 2-line percussion somewhat more narrow, he will probably
> want the repeat dots inside the staff.

I still lean towards special casing 2-line staves.

p

[Keith, 2012-09-09 07:58:16]

On Sun, 09 Sep 2012 00:22:43 -0700, Benkő Pál <benko.pal@gmail.com> wrote:

>> On Sat, 08 Sep 2012 15:26:21 -0700, <benko.pal@gmail.com> wrote:
>>
>>> do we want to support
>>> - NR 2.5.1 style 2-line percussion staves (setting both line-count and
>>> staff-space to 2 instead of setting just line-positions to (-2 2))?
>>> - default TabStaff's (even line-count, 1.5 staff-space)?
>>>

> I can see two possibilities:
> 1. special casing two-line staves.
>    simple but will fail for staves with a big space in the middle but
> otherwise dense.


> I still lean towards special casing 2-line staves.
>

That would be fine.

[dak, 2012-09-09 08:16:44]

Stupid question: anybody have an idea where the dots are to fall in four-line
plainchant?  Is this compatible to where we want to see them in six-line and
four-line tablature?

[benko.pal, 2012-09-09 08:26:09]

plainchant definitely has no bar lines; I guess it has no repeat signs as well,
but if it has, it's (still guessing) another glyph LilyPond doesn't know.

I'm fine with not distinguishing four-line plainchant and four-line tablature.

[dak, 2012-09-09 08:38:33]

Anybody else annoyed just how hard it is to get simple things like that right,
when they probably weren't thought through before anyhow?

In any case, is there a "natural" already existing override for hand-tweaking
dot distance?  With vertical time signature position, the user would always have
the obvious option of using Y-offset himself, and that makes it easier to favor
the simpler rule and put the responsibility for corner cases into the hands of
the user.

I presume that we don't have anything "obvious" like that for dot positions?  I
am not suggesting we actually invent such a thing without being quite sure that
it is actually needed, just curious whether such a fallback would be already in
the toolbox.

[benko.pal, 2012-09-09 16:36:53]

I messed up something, the new patchset is at
http://codereview.appspot.com/6506090