Re: User comments on R shorthand

[Olivier Biot]

On Wed, Mar 20, 2013 at 3:54 PM, James Harkins <address@hidden> wrote:

My opinion (as a somewhat-more-than-casual Lilypond user, and as a
contributor to another music software package [SuperCollider]): Any
change in syntax that will break prior usage should be considered
very, very carefully to be sure the gains are worth it.

The proposal is:

- Old: R2  ==  a full measure rest in 2/4 time
- New: R2  ==  *two* full measure rests in any time signature

That breaks backward compatibility.

Apart from breaking backward functionality, it does so dramatically.
 

I do agree with Kieren that it's annoying to have to enter durations
for full measure rests, when the duration of the measure is known
elsewhere. So I think some change, if possible, would be nice to have.

I agree, but it is difficult since the measure size can change, and as I understand it the full-measure rest is treated the same way as any note or rest or space event.

In my humble opinion a full-measure rest should be interpreted as a full-measure rest, otherwise I could simply write it as a "metered rest" and compress the rests with some magic. today this is not the case, it's a special case of a "metered rest" where 'only' the rendering of the rest will differ.

I think Joram's suggestion, R*n, makes a lot more sense. It's related
to the current syntax, just more convenient, and it doesn't break
existing uses since the parser can distinguish among all of the
following unambiguously:

R2 (a full measure rest in 2/4 time)
R2*2 (two full measure rests in 2/4 time)
R*2 (two full measure rests in any meter)

Since a full-measure rest is a different thing, why not make it look differently?

I'm not in favor of using "R" for a full-measure rest since its behavior differs from notes and rests.

Why not write \Rest{2} instead of R*2 or R2*2 for two 2/4 time full measure rests?

The fact that Kieren's original proposal would change the meaning of
the number immediately after R raises a red flag for me -- breaking
compatibility, confusing current users once they are forced to adapt
to the new syntax -- and the only gain over the second proposal is to
lose a * after R. That falls far short of the threshold to justify
breaking existing syntax, IMO.

I'm strongly against Kieren's original idea. I'm cautiously in favor
of Joram's alternative.

Time signatures are set differently than music expressions.

I think full-measure rests should be treated differently as well, to avoid the confusion that will definitely appear when combining music expressions and time signature changes.

Best regards,

Olivier