Prescriptive or descriptive: descriptive (this standard attempts to describe existing conventions)
TODO: Create identifiers to distinguish the SiGN definition for cubes / normal puzzles / other puzzles.
Programs implementing a custom parser who wish to keep it as simple as possible should support reading and writing Simple General Notation (SiGN), defined at the end of the first section below as
Programs wishing to provide “full” algorithm should support Larger General Notation (LGN), defined at the end of the second section below as
Definitions below use a variant of Backus-Naur form. Consider this example:
foo = "A" bar = "1" baz = "C" / "-D" thing = foo bar / foo baz
Then the valid values for
These standards aim to satisfy these goals:
This allows programs to interoperate easily by following the standards in a straightforward way.
The exact set of base moves depends on the puzzle. For cubes and “normal” twistypuzzles, we define
base-move to be the following specific definition:
bare-move-family = x / y / z / m / e / s / M / E / S layer-move-family = U / L / F / R / B / D range-move-family = u / l / f / r / b / d / Uw / Lw / Fw / Rw / Bw / Dw positive-int = [1-9][0-9]* dash = "-" base-move = bare-move-family / layer-move-family / positive-int layer-move-family / range-move-family / positive-int range-move-family / positive-int dash positive-int range-move-family
Other puzzles may define variants on this, but they must never contain whitespace or end in a digit.
repeated-move is a
base-move with an optional suffix to indicate repetition.
# TODO: Is R0 allowed? prime = "'" repeated-move = base-move / base-move positive-int / base-move positive-int prime
The prime serves the purpose of a negative sign, indicating repetition of the inverse move.
sign-alg is a sequence of moves written out with spacing between them:
single-space = " " single-spaced-move-sequence = repeated-move / repeated-move single-space single-spaced-move-sequence sign-alg = single-spaced-move-sequence
lgn-alg can be expanded and normalized to a
A repeatable unit is a unit that can be repeated without being wrapped in a repeatable unit itself.
(Definitions of group / commutator / conjugate are below.) repeatable-unit = base-move / group / commutator / conjugate
This is similar to
repeated-move above. In fact, every
repeated-move is a valid
repeated-unit = repeatable-unit / repeatable-unit positive-int / repeatable-unit positive-int prime
The same requirements as documented for
repeated-move apply, except with units instead of moves.
TODO: document when
white-space can contain newlines.
white-space = single-space repeated-white-space = white-space / white-space repeated-white-space sequence = "" / repeated-unit / repeated-unit repeated-white-space sequence
optional-white-space = "" / repeated-white-space embedded-sequence = optional-white-space sequence optional-white-space group = "(" embedded-sequence ")"
The following identities hold:
(A B C... X Y Z)' == Z' Y' X' ... C' B' A'
(A B C... X Y Z)n' == (Z' Y' X' ... C' B' A')n
conjugate = "[" embedded-sequence ":" embedded-sequence "]" commutator = "[" embedded-sequence "," embedded-sequence "]"
The following identities hold:
[A: B] == A B (A)'
[A, B] == A B (A)' (B)'
A general LGN algorithm is anything that is a valid
lgn-alg = sequence
There are 6 axis puzzle rotations:
This is best defined using an example:
x moves “like
R” means that means that
x rotates the entire puzzle using the same rotation transformation that the
R layer moves during an
R move. One can think of this as an
R move “taking the entire puzzle with it”.
Note that this corresponds to the axes defined in standard 1.4.1.