Conway's Game of Life for curved surfaces (2012) | Hacker News Reader

Conway's Game of Life for curved surfaces (2012)

0fps.net

110 pointsbabelfish16 comments

The end of the post links to two jsfiddles:

> If you want to try out SmoothLife yourself in your browser, I made a few jsFiddles which illustrate the basic principle. Here is a Fourier based implementation that follows the discussion in this article pretty closely: http://jsfiddle.net/mikola/aj2vq/

> I also made a second WebGL/GPU based implementation that uses a discretization similar to that proposed in Rafler’s original paper: http://jsfiddle.net/mikola/2jenR/

The jsfiddles seem to still be alive, but the "Run" button doesn't work for me. Anyone have any luck with them?

Same. Looks like the "numeric" JS library is no longer hosted at the URL they link to in their dependency list— I had to replace it with "https://unpkg.com/numeric", which seems to have fixed it for me.

Where can you see this dependency list? Tried to look everywhere in the jsfiddle UI but don't find it

My best attempt was setting numeric="https://unpkg.com/numeric" in the "Framework <script> attribute" dialog but that did nothing.

EDIT: fixed it by copypasting the numeric.rep code from https://gist.github.com/ftrotter/5605853#file-numeric-1-2-4-... with 'var numeric = {}' in front of it near the top of the JS file

> Where can you see this dependency list? Tried to look everywhere in the jsfiddle UI but don't find it

"Resources" in the left side menu

https://art.muth.org/smoothlife.html#

webgl implementation of SmoothLife from the comments section of the YouTube video linked in this article

Since this is (at least approximately) invariant under continuous shifts, continuous rotations, and time, should there be some kind of conservation of momentum and energy be visible? At least the amount of white/black shade in cells isn't conserved as far as I can tell (unless perhaps the white is spreading out while fading to black over huge distances... but the grid is wraparound here so probably not the case here)

Assuming you're thinking about applying Noether's theorem, you need your system to have something like a Lagrangian (or be symplectic). From a quick glance it doesn't seem to me that you can write down an action S for the system so it's not clear that you could apply the theorem to get conserved quantities corresponding to the local symmetries of S.

That "smooth life" video was amazing! Really has an actual life feel to it.

Calculus plus Conway's Game of Life = infinitesimal approximation of actual life while never reaching its full complexity! But the artifice is cool, its similar in many ways to Gaudi's work, except that Gaudi probably understood, in some way, that what he was doing was artificial, which is what makes it both so fascinating and disturbing at the same time: it plays with, and then expands life. We view this as a subjugation of life to architecture--now, of course, computer architecture. What if the infinite overwhelmed the finite instead? Then we could have a power unparalleled in history. But it would not be an infinity from the perspective of the finite; symbolized, but not represented. We could say, beyond infinity, like Buzz Lightyear!

Discussed (just a bit) at the time:

Conway’s Game of Life for Curved Surfaces - Part 1 - https://news.ycombinator.com/item?id=4808071 - Nov 2012 (1 comment)

That same paper was also cited by Lenia [0,1], another family of "continuous cellular automata"!

[0] https://chakazul.github.io/Lenia/JavaScript/Lenia.html

[1] https://chakazul.github.io/lenia.html

I did a Swift+Metal implementation of SmoothLife a while back: https://github.com/tscheepers/SmoothLife-Swift

Love it when more of the Internet discovers Mikola's work. This guy made so many nifty packages

What's important is there's a whole Conway wiki with an alien language that a LLM could learn to generate.

(2012)

Added. Thanks!

Reply on news.ycombinator.com