Food Web Creation

I have created a simple template for food web creation.It is a GIMP .xcf file that allows you to drag and drop in niches, which you group into trophic levels. The sizes of trophic levels in this are only suggestions. The limits on numbers are the limits on populations of trophic levels and not numbers of species.

The .xcf can be downloaded here.

Please put all arrows on the "arrows" layer, preferably with color 5e5e5e.

This is a quick example of a biome's food web that I have created.

Spoiler:

* Working on min/max arrangements for sucession stages. So far, minimum for stage 1 is about 4 plant types, 2 herbivores, and 1 transient predator.

I'll have more data eventually.

Mysterious_Calligrapher wrote:

* Working on min/max arrangements for sucession stages. So far, minimum for stage 1 is about 4 plant types, 2 herbivores, and 1 transient predator.

I'll have more data eventually.

Thank you for getting on the succession stages, and also for bumping this thread. WE NEED TO WORK ON BIOMES AND FOOD WEBS. THEY ARE EASY, AND MUST BE DONE.

Put out the cattle call. I want to get this on a roll before college comes back and eats my free time.

Suggestion:

Why don't we add some basic symbols for size, to make things easier to read? Maybe some to designate Autotrophe/herbivore/carnivore/omnivore/etc.?

More specifically, I thought we weren't going to have pre-determined food webs, but instead lay dwn some dietary rules in the OE and have the critters naturally head towards those? It's simpler to code and allows smoother transition during migration. If there's auto-populating through biomes than food webs should become rather redundant.

Tenebrarum wrote:

Suggestion:

Why don't we add some basic symbols for size, to make things easier to read? Maybe some to designate Autotrophe/herbivore/carnivore/omnivore/etc.?

More specifically, I thought we weren't going to have pre-determined food webs, but instead lay dwn some dietary rules in the OE and have the critters naturally head towards those? It's simpler to code and allows smoother transition during migration. If there's auto-populating through biomes than food webs should become rather redundant.

We need to somehow regulate what eats what, so that we know what reaction mutations are possible. The only way to do it is to have a food web.

~sciocont wrote:

We need to somehow regulate what eats what, so that we know what reaction mutations are possible. The only way to do it is to have a food web.

Lies.

Remember how biomes will have organism slots? The easy solution is to take said slots and record what organisms fit eachothers dietary requirements. Thus the game will know what eats what. The one flaw with this is that it will make it difficult to predict what can overpower what.

What I'm worried about is that pre-made food webs are static, while our evolutionary system dictates that the actual food web should be in constant flux as creatures develope new measures and counter measures.

Tenebrarum wrote:

~sciocont wrote:

We need to somehow regulate what eats what, so that we know what reaction mutations are possible. The only way to do it is to have a food web.

Lies.

Remember how biomes will have organism slots? The easy solution is to take said slots and record what organisms fit eachothers dietary requirements. Thus the game will know what eats what. The one flaw with this is that it will make it difficult to predict what can overpower what.

What I'm worried about is that pre-made food webs are static, while our evolutionary system dictates that the actual food web should be in constant flux as creatures develope new measures and counter measures.

That makes a lot of sense. You're completely right... maybe we need to restructure how organisms are set up in a biome. Perhaps we can achieve a much more fluid concept.

~sciocont wrote:

That makes a lot of sense. You're completely right... maybe we need to restructure how organisms are set up in a biome. Perhaps we can achieve a much more fluid concept.

The less changes we need to make to our current concepts the better. I like how biomes work thus far. All we need to do is figure out how to calculate what can over power what.

Tenebrarum wrote:

~sciocont wrote:

That makes a lot of sense. You're completely right... maybe we need to restructure how organisms are set up in a biome. Perhaps we can achieve a much more fluid concept.

The less changes we need to make to our current concepts the better. I like how biomes work thus far. All we need to do is figure out how to calculate what can over power what.

Which means more attributes that we need to figure out.

~sciocont wrote:

Which means more attributes that we need to figure out.

Not nessicarily. We could run darwinian style simulations off-screen, but it would suck up processing power and time.

Still, if we want to do full scale attributes we're essentially going right back to the "Catalogue all the things!" problem we had with Lamarkian.

Tenebrarum wrote:

~sciocont wrote:

Which means more attributes that we need to figure out.

Not nessicarily. We could run darwinian style simulations off-screen, but it would suck up processing power and time.

Still, if we want to do full scale attributes we're essentially going right back to the "Catalogue all the things!" problem we had with Lamarkian.

Actually, i'm just saying we need to determine what a creature can do as defined by what it can eat, which can be determined by organs, size, behavior etc, which is all covered in the editor.

Currently the webs are just concepts... I have to figure out how many slots we need per biome, (or at least per sucession stage) and how many we can run before it becomes a new sucession stage.

For example, four species of plants is the bare minimum for sucession stage 1, but as soon as 1 species randomly moves into the area, it is technically stage 1 and is randomly populated in (low) percentages set by the properties of the biomes with three other suitable plants. In a desert, this could mean that you have a cluster of cactus here, a couple grasses there... if there are resources, they'll spread eventually. If not, they're SOL.

Mysterious_Calligrapher wrote:

Currently the webs are just concepts... I have to figure out how many slots we need per biome, (or at least per sucession stage) and how many we can run before it becomes a new sucession stage.

For example, four species of plants is the bare minimum for sucession stage 1, but as soon as 1 species randomly moves into the area, it is technically stage 1 and is randomly populated in (low) percentages set by the properties of the biomes with three other suitable plants. In a desert, this could mean that you have a cluster of cactus here, a couple grasses there... if there are resources, they'll spread eventually. If not, they're SOL.

We should have some rules for building the food webs biomes. First, we need to figure out trophic relationships. I suggest we use something along the lines of biomass* relationships, because that's how it happens in the real world.
Each level has 10% of the biomass of the level below it (we could increase this in gameplay, but bare with me) the lowest level in any biome is the autotrophs, which are grouped with any smaller organisms that aren't big enough to fit into the food web. (insects, plankton, detritus). The biome, according to its sucession stage, gets an allotted amount of biomass for this lowest level. The plants have to fit into that amount in their total biomass. Then we have herbivores and anything that eats detritus or smaller organisms in the next level. Their biomass is directly related to the biomass of the lowest level. After these, we have the upper carnivores, whose biomass is dependent on the herbivore/lower carnivore biomass.
That would tell us how many organisms could be in each level. IT SHOULDN'T MATTER (in this rule) how many species are in each level, as long as they are under the biomass limit. That's the first rule for food webs.

*biomass of a level is calculated by taking the masses of every organism in the level, which is easily done because organisms do have a mass property assigned to them by the OE. The mass is multiplied by their population and voila.

Other Ideas are:

Spoiler:

How we define What Eats What

Still working on this a bit...

Basically, the organisms need to have lots of attributes. Autotrophs are easy, they just need to fit the climate, so we need a list of climate attributes.
This is a list of attributes that I came up with yesterday for autotrophs- it's not complete, but it tells you where we're going.

Spoiler:

I think we should add to that a value for water retention. Also, we need different types of anchor (fibrous, single, spread out, deep/shallow)

Now, for animals, we start out with herbivores (I pointed out that detritivores and lower carnivores need only fit the climate, as they aren't really eating anything in the food chain). By the way, we should give the player the ability to create a smaller than tiny creature, but not to play as one, because there isn't enough detail in the game to make their environment look real. Smaller than tiny creatures will be populated in biomes randomly.

Herbivores (and all other animals) can be defined by intake (how much food they need) and preference (what specific kind of food they eat). Therefore, a root-eating animal cannot feed on fibrous-root plants, woody-rooted plants, or unrooted plants. It will eat EVERY plant that it can. the maximum biomass of that particular creature is calculated by taking 10% of the biomass of all the species it eats. However, if there is competition for a species, the game splits up the biomass of that species evenly between its consumers.

I honestly think this can be simplified by using my Resource Conversion idea. Biomes could be created and have niches added to them each preparation phase.

I don't see what's wrong with:

Sugar -> Protein (Herbivore eats plants)
Protein -> Protein 2 (Predator eats herbivore)

Each animal would evolve different abilities, and the second it evolved the ability to eat something, it would create the niche. My code for the niches already has this ability.

roadkillguy wrote:

I honestly think this can be simplified by using my Resource Conversion idea. Biomes could be created and have niches added to them each preparation phase.

I don't see what's wrong with:

Sugar -> Protein (Herbivore eats plants)
Protein -> Protein 2 (Predator eats herbivore)

Each animal would evolve different abilities, and the second it evolved the ability to eat something, it would create the niche. My code for the niches already has this ability.

Ok, so you're saying each organism is made up of these resources, and certain organisms can eat certain resources? That sounds like a plan.

I've been attempting to propose this for a while..

But anyway, yes. Everything in a biome is a resource. A resource is made up of multiple compounds. Depending on the compounds in a resource, a given organism may or may not evolve to eat the other animal/thing. (A rock is a resource too.)

I believe using a resource conversion chart will help with this. Basically, I propose that any given organisms's stomach can change compounds based on the possible conversions in the conversion chart. The organism would evolve certain conversions and would thus be able to build certain parts as well. (Protein is major.) A conversion chart could be in xml, and could then be edited for some interesting (out of this universe) evolutions.

roadkillguy wrote:

I've been attempting to propose this for a while..

But anyway, yes. Everything in a biome is a resource. A resource is made up of multiple compounds. Depending on the compounds in a resource, a given organism may or may not evolve to eat the other animal/thing. (A rock is a resource too.)

I believe using a resource conversion chart will help with this. Basically, I propose that any given organisms's stomach can change compounds based on the possible conversions in the conversion chart. The organism would evolve certain conversions and would thus be able to build certain parts as well. (Protein is major.) A conversion chart could be in xml, and could then be edited for some interesting (out of this universe) evolutions.

You know what, that does seem like quite a good idea when applied to this. I remember discussing breakdown of these proteins, etc.

That's more or less where we were going with tags, Roadkill, except that we have omnivores and the ability of certain species to eat certain types of plant where others can't. (Species one attribute: Accept Food Type Plant Stem, Leaf versus Accept Food Type Plant Fruit or Accept Food Type Plant All) Not being a programmer, I have no idea how this works into your system, and for that I apologize.

A chart could help immensely, and we wouldn't have to draw arrows anymore.

I agree with Roadkill on the need for a rescource system (pre-civ). While the tag system will tell you what an organism can or can't eat (and what it's prey is made of), we need a way of figuring that out, and that would be the rescource system. For example, if you have a mouth (or hands/other appendage) capable of gathering fruit, teeth capable of processing it, and a stomach capable of digesting it, then you can eat fruit [Accept: General food type: Plant: Fruit], if your missing any of those items then you can't. Similairly to eat meat you need to be able to take bites out of it, chew it, swallow it, and digest it. Depending on your stomach you can process these into different products too (fruit -> sugar/energy, meat -> protein -> used for growth).

While this may sounds like a lot of extra work, once you have a simple rescource system it should all work itself out fairly nicely.

I also like the idea of resource conversions, an organism would need to gather different food types in order to survive, grow, etc. This also means that organism would be dependant on multiple 'prey' types (I use the term 'prey' loosely, this could be another organism, a plant, or a rescource such as a salt lick, watering hole etc.), which helps diversify each species niche.

I'd also suggest that we think of niches in terms of parameters other than diet too. While we don't need to keep track of everything which would realistically constitute a niche parameter, there are atleast a few I'd suggest:

• Nesting habitat (burrow, vegetation, open ground? - also, materials required for nest building)
  
• Temperature/altitude range - this isn't entirely necessary, as biomes will be divided according to these parameters anyway
  
• Waking hours (nocturnal and diurnal organisms won't generally compete with eachother)
  

Many of these are parameters we'd have to keep track of for other reasons anyway, but they serve as additional parameters for defining niches, allowing for greater diversity. I'll try and come up with a more complete list of both realistic niche parameters, and ones we could reasonably use.

It's also worth remembering that species don't need to be in the same exact niche to compete with (and potentially exclude eachother), if multiple species share just one aspect of a niche (diet, habitat etc.) then they will compete for that aspect, hurting both species population (e.g.: rabbits and cows have very little in common, besides eating grass, but if you stick both in the same field then there will be less food for both). This isn't a problem, it simply means that species compete in more specific ways, allowing them to make more specific reactionary evolutions.

I need to think this through when I'm less tired, I'm not sure the above makes complete sense to me, so sorry if I confused anyone, or went over any old ground you've already covered.

I've been working on niche definitions for a bit, and I've come up with factors that define diet. These are:
What an animal:
can eat
can find
can catch
needs to eat

Also, using the OE as a template, I've come up with the different materials that an animal could eat off of another organism.

wood
bark
leaves
stems
fruits
seeds/nuts
roots
tubers/bulbs
fruiting bodies
sap

muscle
organs
fat
bone
skin
wormacle muscle
arthropod muscle
artthropod exoskeleton
body fluids

These materials are already defined by the OE, so they don't require a separate resource system. An organism's body makeup tells what can eat it, and its organs and structures tell what it can eat.
In a biome, there are also "level zero" materials that animals may need/eat. These are:

water
detritus
soil
bacteria
planktonic organisms

If I've forgotten anything, please tell me.

Now, matching up these materials with structures is the first step to determining what can and cannot be eaten. Another thing I've finished off is the use of toxins. All organisms will have a two digit resistance tag to determine what toxins they can resist, and to what extent. The tag looks like this:

[42]

Each digit is from 0 to 9 so there are 100 different resistances. Every animal or plant that creates a toxin or uses a toxin will have a toxin tag with the same formula. The difference between the resistance and toxin tags is the percent chance that a toxin will kill an organism.

So far, this looks like exactly what we need. I don't have much to contribute, but looking at this, and back to Spore's pre-release demos, I got an idea.

Couldn't we conceivably make a prototype for this system to test it out without the need of our game engine? If someone here is familiar with something like flash, or java, or any simple programing platform, we could make the basis of this in there. Organisms would get text descriptions, and we'd limit evolution to non-body based forms (Can you catch it wouldn't be the question, as that depends on values inherently connected to the rendering software.). Once done, we could release it to our fans and let them run the simulations, and report back to us with results. It would let us tweak and moniter the viability of our system while roadkill and the roadkillettes continue their tour of microbe stage.

Thoughts?

Tenebrarum wrote:

So far, this looks like exactly what we need. I don't have much to contribute, but looking at this, and back to Spore's pre-release demos, I got an idea.

Couldn't we conceivably make a prototype for this system to test it out without the need of our game engine? If someone here is familiar with something like flash, or java, or any simple programing platform, we could make the basis of this in there. Organisms would get text descriptions, and we'd limit evolution to non-body based forms (Can you catch it wouldn't be the question, as that depends on values inherently connected to the rendering software.). Once done, we could release it to our fans and let them run the simulations, and report back to us with results. It would let us tweak and moniter the viability of our system while roadkill and the roadkillettes continue their tour of microbe stage.

Thoughts?

That's not a half-bad idea. "Can you catch it" is really more connected to mobility and hunting methods.

~sciocont wrote:

That's not a half-bad idea. "Can you catch it" is really more connected to mobility and hunting methods.

Well yes, but things like weapons, armor, weak points, speed, etc. are all connected to the OE, which is yet to be made.

In this version, all players will be able to do is make template orgs by the looks of it, but that should still give us great information on how auto-evo treats different templates.

Tenebrarum wrote:

~sciocont wrote:

That's not a half-bad idea. "Can you catch it" is really more connected to mobility and hunting methods.

Well yes, but things like weapons, armor, weak points, speed, etc. are all connected to the OE, which is yet to be made.

In this version, all players will be able to do is make template orgs by the looks of it, but that should still give us great information on how auto-evo treats different templates.

Indeed.