Those are really interesting ideas and they make total sense to me. We could try and implement them since it's now open-source, if only I weren't so busy with college...

This is a neat suggestion. It would however fundamentally change the simulation, and require a pretty significant rewrite to the codebase. There are also a few points that need to be addressed, like how does an organism decide to use energy for movement or reproduction or survival? It's certainly doable and I'd be interested to see results, but I'll leave it to the open source community to figure it out. Thanks for the input 😁

​@@EmergentGarden I didn't think it would require that big of a change, to be honest. For reference in case anyone wants to give it a try, this is what I can come up with: Since food_collected is already a thing that grows with food consumption, it can be used as an energy storage. Adding a max would be just like doing a (if variable > max) variable = max; with the max maybe being the number of cells, double that or something along those lines. For movement cost it should be possible to just add this.food_collected--; in attemptMove(). Or subtract a movement cost, which could be given by a function similar to the cost for reproduction. It could even be a division of the total number of cells by the number of mover cells. Stopping lifespan from calling the die() function should be enough to disable the age limit, unless I'm missing something in other scripts. And for the when to reproduce point, there are many ways to do it, and more complex ones are very annoying, but the way I usually do is let them reproduce if they meet the energy requirements. That would change nothing from how it currently works, and would just pull food_collected back to near zero, allowing for the possibility to take starvation damage due to other expenses. However, to make things more dynamic, we could use the lifetime variable differently, appropriating it for a reproduction counter. It could ++; every time the organism consumes food (regardless of whether or not that energy is used) and when that reaches the foodNeeded(), it reproduces, possibly throwing food_collected in the negatives. And as for starvation damage, the simplest way to do it is (if food_collected < 0) harm(); Of course you could also add an int to harm(), and do damage += that_int, instead of damage++; which would allow a better control of what is damaging the organism and by how much. This could allow us to control the ratio between starvation damage and damage from other organisms. Any int that goes into harm(int) that is greater than 1 would multiply the damage by that, and so you could times the damage dealt by other organisms by 10, as well as multiply the maxHealth() by 10, and have starvation damage be 10 times milder than combat damage. Increasing maxHealth() would allow things like doing food_collected--; every frame, or even food_collected -= this.anatomy.cells.length; This is the passive food consumption required to just exist, which combined with reproduction should help accumulate damage and provide the substitute for a lifespan. Of course the values could be tweaked to provide better combinations, such as the ratio between max health, combat damage, starvation damage, passive energy consumption and movement energy consumption. The same goes for the reproduction cost, balancing it with max food_collected, how much food_collected is incremented by each food eaten, etc.

@@EmergentGarden Hum, could be fun to give this a try. I've also wanted to learn to program in Rust, and Rust can run both on hard metal and in a browser through web assembly.

@@EmergentGarden this video might be of interest to you if you ever do decide to try implementing some of ops suggestions ukposts.info/have/v-deo/hmOlhnyYsKeH1Z8.html

Same

I've been missing out it seems lol

it’s fascinating watching these little organisms compete with each other you can see organisms evolve and die out completely in only like 5 minutes

Oh yes please that would be wonderful.

octogonal

how

@@handledav idk it was a year ago lol

Would pheromones just be randomized colors? How would you represent the different pheromones?

For three pheromones, like the bibites, I would use the bibites way. Pheromone 1 is red, 2 is green, 3 is blue. Each pheromone blends into each other using the RGB color system. For more pheromones, just show the most plentiful one.

And you need a nose or similar means, to notice any pheromones.

Maybe a "tooth"-cell, that takes one unit (of food) from an adjacent organism and transfers it to its own organism. As organisms accumulate food for reproduction, they usually should have some food in store(stomach).

I'd think more like a muscle cell, it won't be extending, it will just expand and contract. The hard part being that the engine isn't designed to handle cells getting bigger and smaller, however if he could get expanding, and compressing cells working then it would help a lot.

Convergent evolution. There are many more people out there working on things like this.

Every cell you add increases the lifetime

Back from looking at the code. It's exactly as simple as OP says, everything complex happening in the simulation is an emergent property of very basic constituent components. At a glance it seems easy enough to to add to or modify. Can't wait to play with it. Thanks for making it open source!

Now i know how to make filter feeders.

That sentence was unintentionally hilarious

You can set the value to 0.1 or something like that, but if you do that you get only like 2 fps

You can reduce the grid-size i.e. the size of individual cells. By reducing the size from 4x4 to 1x1, you increase the world by factor 16.

I am a little sad also that the scenario instantly resets when the last organism dies. I feel like it would be cool to get a summary of everything at the end. You have pop saved already, but like, a 'speciation' event could also be tracked, where it tracks identical species, and if there's a proportionate growth of a new species. If you log at the moment of birth the configuration, and whenever a species dies it also logs the configuration, you can track the major species. Not totally sure how the math could work on that, I'm sure there are a lot of solutions. Proportionate representation might be a good metric, anything with a stable number of like 10 or 15 species of the same type at any given time could be considered 'stable' and makes it on a list. The neural nets would all be flattened in this, but I think that's fine.