It looks like the author has a pretty simple procedure for computing the 'identity' sandpile (which they unfortunately don't describe at all):
1. Fill a grid with all 6s, then topple it.
2. Subtract the result from a fresh grid with all 6s, then topple it.
So effectively it's computing 'all 6s' - 'all 6s' to get an additive identity. But I'm not entirely sure how to show this always leads to a 'recurrent' sandpile.
EDIT: One possible route: The 'all 3s' sandpile is reachable from any sandpile via a sequence of 'add 1' operations, including from its own successors. Thus (a) it is a 'recurrent' sandpile, (b) adding any sandpile to the 'all 3s' sandpile will create another 'recurrent' sandpile, and (c) all 'recurrent' sandpiles must be reachable in this way. Since by construction, our 'identity' sandpile has a value ≥ 3 in each cell before toppling, it will be a 'recurrent' sandpile.
In the case of piling sand exactly in the centre, the intermediate states between the initial state and reaching the final equilibrium seem to get closer to having a circular boundary as the grid size increases, instead of the diamond-shaped boundary you might expect for a symmetrical object in a planar grid. Take a look at the largest resettable grid doing this within a couple seconds of being reset.
> The rules of abelian groups guarantee that these identity sandpiles must exist, but they tell us nothing about how beautiful they are.
This has causality backwards—being a group requires an identity element. You can't show something is a group without knowing that the identity element exists in the first place.
In fact, a good chunk of how this article talks about the math is just... slightly off.
Isn't this single frame state of a classic cellular automata? Note, not "just" because I mean no disrespect. I don't understand how this differs from Conway's life other than nuances of the live or die rule.
It looks like the author has a pretty simple procedure for computing the 'identity' sandpile (which they unfortunately don't describe at all):
1. Fill a grid with all 6s, then topple it.
2. Subtract the result from a fresh grid with all 6s, then topple it.
So effectively it's computing 'all 6s' - 'all 6s' to get an additive identity. But I'm not entirely sure how to show this always leads to a 'recurrent' sandpile.
EDIT: One possible route: The 'all 3s' sandpile is reachable from any sandpile via a sequence of 'add 1' operations, including from its own successors. Thus (a) it is a 'recurrent' sandpile, (b) adding any sandpile to the 'all 3s' sandpile will create another 'recurrent' sandpile, and (c) all 'recurrent' sandpiles must be reachable in this way. Since by construction, our 'identity' sandpile has a value ≥ 3 in each cell before toppling, it will be a 'recurrent' sandpile.
I implemented this in Rust some years back. It is connected to some serious research mathematics (see f.ex https://www.ams.org/notices/201008/rtx100800976p.pdf)
https://github.com/FredrikMeyer/abeliansandpile
In the case of piling sand exactly in the centre, the intermediate states between the initial state and reaching the final equilibrium seem to get closer to having a circular boundary as the grid size increases, instead of the diamond-shaped boundary you might expect for a symmetrical object in a planar grid. Take a look at the largest resettable grid doing this within a couple seconds of being reset.
I found 'xsand.c' (X11) in 1995 by Michael Creutz, that simulated these sandpiles; I had fun with the sand but also learned C from it.
> “…an abelian group is both associative and commutative…”
If something is not associative it is not a group. An abelian group is a group which is commutative.
Very related (yet idiotically titled, as always) veritasium video https://youtu.be/HBluLfX2F_k?si=6lVPLvJNc2YH_4go
It's like reverse clickbait with him
Yeah, I wish he'd do a second channel that is just reposts with normal titles.
> The rules of abelian groups guarantee that these identity sandpiles must exist, but they tell us nothing about how beautiful they are.
This has causality backwards—being a group requires an identity element. You can't show something is a group without knowing that the identity element exists in the first place.
In fact, a good chunk of how this article talks about the math is just... slightly off.
Isn't this single frame state of a classic cellular automata? Note, not "just" because I mean no disrespect. I don't understand how this differs from Conway's life other than nuances of the live or die rule.
I don't believe that Game of Life is Abelian.
I don't think you could even define an associative binary operator on states in the Game of Life because of its computational irreducibility.
CGL doesn't have the scale invariance ("fractality") of ASM. ASM criticality is stable and persistent. "fractal life on edge"?
what that looks like
https://youtu.be/rKD51IUNK3A?t=40s
So that gets to how it differs, but it doesn't say its not a cellular automata. It could say "it's a cellular automata with different rules"
It is a cellular automata distinguished by commutativity. You used CGL as the basis for comparison, that's highly nonAbelian.
According to Wolfram (& I agree :), everything is a cellular automaton, so comparing to CGL made more sense to me.
It seems the sand only spills up and to the left.
Now I want to redo a bathroom. Good job, writer!