Syntax Proposal: Ranged State Application (RSA)

[andjsrk]

• "single bracket command" refers to a notation in the form of (...) that affects following input.
• "state" refers to what "single bracket command"s change.

Motivation

It is good for user experience that not needing to remember what the previous "state" value was.
Probably because of that, many syntaxes exist: (^c), (vc), (scdef), (:ceg) or (|ceg).
What if we extend this to every "state"s?

• Set volume and get back to the previous state later.
• Set instrument and get back to the previous state later.
• Set arbitrary default note length and get back to the previous state later.
• etc...

Semantic and Temporary Syntax

(ⓐ: an input. a sequence of either a note/rest/tie or a "single bracket command")
(rsa ⓐ)

Plays ⓐ. After that, every "state" gets back to right before the RSA bracket is encountered.

Note that the proposed syntax is not something like (rsa[ⓒ] ⓐ)(ⓒ: "single bracket command"s). This makes the syntax more powerful:

(q=8)c(rsa (q=32)C(q=64)d(q=128)D)e~ /= same as (q=8)c(q=32)C(q=64)d(q=128)D(q=8)e~

Examples

(q=8)ceg(rsa (q=32)gec)ceg /= same as (q=8)ceg(q=32)gec(q=8)ceg

(p=0)ceg(rsa (p=1)ceg(p=2)ceg)ceg /= same as (p=0)ceg(p=1)ceg(p=2)ceg(p=0)ceg

(q=4)c(rsa (p=55)(v=40)g)^c~ /= same as (q=4)c(p=55)(v=40)g(p=0)(v=80)^c~

Unresolved Questions

• Should it be allowed that picking which "state"s get back? For example:

  (rsa[q,v] c(p=1)(v=50)(q=4)e)g` /= same as c(p=1)(v=50)(q=4)e(v=80)(q=16)g