Segmentation is a significant component in the evolution of many species. The current suggestions about starting a new limb being expensive would make it nearly impossible for the game to evolve millipedes and other many-legged organisms. (Ironically in real life, millipedes were one of the first land-dwelling species)
In later stages we really need to add discounts for things like repeating patterns and symmetry
Thinking from both the Player and Developer side, how would that work? You can create a segment, and for a cost, everything on that segment is duplicated a little further down the line (my current preferred way)? Prefabs? A bunch of different ways.
I think that would even be relevant right now. A discount in mutation points make sense because DNA needs less nucleotide to encode this (related to the idea of Minimum Description Length although I’m not versed in its application in genetics). That would help incentize apparition of roughly regular patterns.
Now, it would often turn out to be advantageous to actually change one of the regular repetitions, triggering differentiation. We’d thus need to allow for progressive evolution of one instance, and its deviation from the original template.
Maybe there are also functional advantages, but I’m a bit clueless about this part. Overall, any input from a theorist would be appreciated!
Maybe up the cost for developing an organelle the first time, and afterward it’s significantly easier (even if you delete it, it goes dormant and doesn’t cost extra again).
I planned to tackle this in light of my germ layer concept, in which I argue there should probably be three “types” of tools within the editor available depending on the player’s evolutionary path: soft-bodied tools for diploblastic and mollusk/worm-esque organisms, exoskeleton tools for animals with an chitinous exoskeleton akin to arthropods, and internal structure tools editing bones for animals akin to vertebrates.
Segmentation would have to be a cornerstone of the middle, exoskeleton based organisms; it is less explicit in vertebrates, only seen in things like assigning individual spine columns.
So perhaps the discussion shouldn’t necessarily be “how do we implement segmentation” but more how do we accurately represent each “path” of life in the editor and what benefits and detriments will be assigned to soft-bodied, endoskeleton, and exoskeleton organisms. That way, we can accurately structure our thoughts and not be overwhelmed with very specific structural phenomena we see in metazoans.
And before I forget, here is a semi-related community post I think would be valuable to understand: Questions about the realistic implementation of repeating parts/segments (and Williston's law) - Multicellular Stage - Thrive Community Forum
This might be a bit tangentially related… But since Deus already extended the conversation to include soft-bodied organisms this might still fit in.
In games, all moving things are in some way segmented as they usually have a segmented rig which they need to be animated.
For fish-like motion, this kind of animation would be very work and ressource heavy. They way ABZU animated their fish and fishlike animals saved them a ton of work and would probably work well with procedually generated organisms as well.
It really starts at 4:00. They used a method called static mesh animation.
I’m tossing this in here because it might be a very valuable tool to animate basically all organisms without an exoskeleton. But I’m sure we wouldn’t want to use completely different modes of animation depending on if an organism has a segmented exoskeleton or not. So it might make sense to discuss this option now rather than later.
2 Likes