Note value of full measure rests

[rpatters1]

The current note specification states, "If [measure] is set to true, then this event may not specify "value" and this event must either be empty or contain exactly one rest. A small wording issue is that it might be better to refer to "duration" rather than "value", since there is no "value" field. (I'm only guessing that the documentation means "duration" here.)

The larger semantic issue is this. Elaine Gould allows empty measure rests to be represented three different ways:

1. always whole rests
2. whole rests in time signatures less than 4/2 and breve rests in time signatures 4/2 and higher
3. a sequence of smaller rests showing the meter (in mixed meter)

Perhaps we can cover 3 with "real" rests, though that loses some semantics. But there is currently no way to specify if the full measure rest should be whole or breve.

My personal view is that Gould's 2nd option should be extended so that 4/1 time sigs and up to 8/1 use longa rests and 8/1 and higher use maxima rests. I do not think the MNX specification should preclude this, especially since the spec includes Maxima as a note value base type.

[mscuthbert]

I'll comment on the main point before getting sucked into my medieval rabbit hole: I think that as far as MNX being first a semantic format, there is a value in specifying that "this is a rest that lasts a measure" so let's keep something like "measure=true" as a possibility, but I agree that being able to specify a way of display (w/ the default of >= 4/2 is breve; smaller is whole) would be very helpful. The sequence of smaller rests leads to difficulties since it's not a single "duration" (agreed, that not value) here.

There is yet another system that some composers and publishers use where anything shorter than 2/4 is represented by the actual note duration (but longer is whole rests, etc.). Here for instance is the Boosey and Hawkes score for John Adams' "Chamber Symphony" where 3/4 and 4/4 are both represented by whole rests, while 3/8 is represented by a dotted quarter rest.

(Boosey and, in the US, John Adams, are like the definition of "can't say that this is a minor publishing house with a no-name composer")

---

Medieval rabbit hole here: Longa rests are so rare and poorly understood today that I don't think I'd advocate them for appearing "by default" in any system. (Only us true nerds could ever figure out a multimeasure rest written in older symbols if there wasn't a number like "11" above it). Further, I don't think I've ever seen maxima rests in CMN -- they're quite rare even in medieval music (pairing two longa rests was more common, sort of like how the 8-bar + multimeasure rest symbols do it). -- in practice they looked exactly identical to a barline today (=double bar back then), so it's nearly impossible to tell if it's meant to be a measured rest or general pause between sections. There also was the longa rest worth two breves (written as a line through two spaces) and the longa rest worth three breves (written as a line through three spaces), of which only the former is used in modern multimeasure rests even if for a time signature like 6/1 or 12/2 they would be more appropriate.

[rpatters1]

I'm not really advocating the medieval rabbithole, except to say that it is also a notation that a contemporary composer might try in very limited circumstances. (I've seen 11/1 in some piece: Boulez or Pendercki perhaps. I'd have to dig through my library.)

The maxima rest I've seen is the same one that is used under a "8" in old multimeasure rest notation. (Two parallel thick lines extending from 2nd line to 4th line.)

In any case, I was hoping for a spec that doesn't preclude it, rather than any active support for it. For example, one way to specify would be literally just to omit the prohibition on "duration" when "measure" is true. Then any rest duration would be allowed for full measure rests.

EDIT: I mistyped "4" instead of "8" above.