multiple timesignatures

[Ruud van Silfhout]

I saw that in January of this year a new feature called compoundTime signatures 
is introduced in Lilypond. A good effort. compoundTime specified the timing of 
each measure. However, for church music (esp. hymns) this does not suffice. 
What 
you see there is that multiple meters are specified a the start of the music. 
This means that the of a song each measure is in one of the initial meters. 
This 
can be seen in the image below (This is BTW a part from Bernstein’s America, 
West Side Story, so this does not happen in church music only ;-)).
 

 
So two measures, each with  a different meter. In church music you also see 
that 
the number of measures that are written in one meter differs from the number of 
times a second meter occurs. So typically at the end of a chorus line or near 
the end of a song the second meter is used. What I would like is that multiple 
meters can be specified at the beginning of the piece and that the \time 
command 
is used to switch between the different meters. So something like (in this 
case):
 
<image should be here, but this is posted through gmane, which doesn't accept 
images>

\multiMeter { 6/8 3/4 } g8 g g c c c \time 3/4 a4 f c.
 
The time signatures for these multiple meters is printed at the start of the 
music.
 
The first fraction in the \multiMeter list is the initial \time to start. 
Subsequently the user specifies \time 3/4  which selects another meter from the 
 
initial meters. The effect is that, because the \time specified as one of the 
initial meters, it is not printed. When e.g. a \time is specified that is not 
in 
the set on initial meters (e.g. 5/8 in this example) , the time signature is 
printed, When the \time changes back to an item from the multi meter list, that 
time signature is printed (or perhaps all time signatures from the \multiMeter 
list: that is not clear to me, but that could be an additional option). 
Normally 
the \multiMeter list is empty, which ensures that all time signatures are 
printed as it was before.
 
I’m willing to help with the implementation, but I’m not experienced in the 
lily 
code and I have currently no environment where I can compile and build a 
version 
of lilypond. I saw that much of the code now is written in scheme and I'm not a 
schemer although I developed some software of Lisp albeit some 20 odd years 
ago, 
so is rusty.