Full measure rests

[Carl D. Sorensen]

This following discussion comes from the -user list.  Reinhold appears to
have identified a case where current LilyPond behavior doesn't match Read's
conventions.

This leads to two questions:

1) Should LilyPond be adjusted to match Read's conventions (Reinhold has
identified how it should be changed, at least to eliminate the breve
symbol)?

2) Should there be an option added to allow the user to switch between
current LilyPond behavior and the behavior recommended by Read?

I bring this up because I didn't want it to get lost on the -user list.

Thanks,

Carl

From: Reinhold Kainhofer <reinhold <at> kainhofer.com>
Subject: Re: Again: Whole Bar Rests and Unmetered Music
Newsgroups: gmane.comp.gnu.lilypond.bugs, gmane.comp.gnu.lilypond.general
Date: 2008-12-06 13:36:01 GMT (3 days, 4 hours and 49 minutes ago)

Am Samstag, 6. Dezember 2008 14:02:04 schrieb Dmytro O. Redchuk:
>   Please, can anybody answer me:
>   1. Shouldn't whole bar rest be semibreve (whole) rest for any actual
> time except 4/2?

Gardner Read says about whole-measure rests (p.98):
"Although the whole rest literally signifies only the value of a hole note,
it 
now commonly serves as the symbol for any completely silent measure,
regardless of the meter or time-signature. Formerly, it served for all
measure 
values except 4/2, this rest being indicated by a sign borrowed from the
breve 
symbol of thirteenth- and fourteenth-century music. But today the whole rest
stands for any empty measure -- for all meters from a theoretical 8/2 to a
3/16. For a 2/16 (1/8) or smaller silent bar, an actual eighth (or smaller)
rest would be used."

>   2. How can i tell lilypond to "draw" whole rest instead of breve
> (double)?

You can't, since it's hardcoded to use a breve rest for ALL measure lengths
>=2 (i.e. >= 4/2)...
>From multi-measure-rest-engraver.cc (lines 218-221):

      SCM sml = get_property ("measureLength");
      Rational ml = (unsmob_moment (sml)) ? unsmob_moment (sml)->main_part_
:
                               Rational (1);
      if (ml >= Rational (2))
    last_rest_->set_property ("use-breve-rest", SCM_BOOL_T);

So, apparently, this "violates" the convention Read describes in two ways:
-) breve should only be used for ml == Rational(2) (if at all!)
-) For ml < Rational (3, 16) a rest of appropriate duration should be used
instead of the semibreve

Cheers,
Reinhold