Patch Map Design Discussion

[tjwhale]

A few points based on the discord discussion:

I think what we’re really trying to do here is to teach a computer how to draw nice patch maps when given planet variables. I think this is an interesting project because it needs all of our skills to be pooled to produce something really good. It’s got to look nice (art team), respond to the changing planetary variables smoothly (programming + theory), a computer needs to have a list of rules from which it can draw the map (programming + theory) and it needs to have good gameplay (gameplay = everyone). I think we’ve got the potential to make something pretty great which is better than anything any one of us could produce so thanks everyone who is contributing, I think this is a cool process.

I like your thoughts goro about working out a grid pattern for how to layout patches, that’s the sort of thing we need to do, have a list of rules from which the map is drawn.

Re geography I think it’s helpful to detach from the idea of geography as that actually gives us a lot of freedom. I haven’t had time to prototype this yet but I think one idea that might look great is if we start in the middle of the map, put down a region (rectangle maybe). Then iteratively we can put down a new region near those already placed and connect it with lines to those that are already there, such that the lines don’t cross. This means it will grow kind of like a snowflake or something which I think could look really good.

On reflection I like the idea of vertical stacks of ocean patches because of the marine snow , as you were saying Naro. The main energy source in mid ocean patches, where it’s too dark for sunlight, there’s no H2S source and no iron will be debris falling from above, so having the patches in a stack makes computing this really easy.

I also wondered about maybe different thicknesses of connection? So some region to region connections are really easy to cross, shown as thick lines, (maybe between two close regions of ocean) and maybe some are hard to cross, shown as thin lines, like into an underwater cave with a narrow entrance. Might add some interest to the flow of species. Maybe when you move some of your pop to the new patch less of them make it through a thin connection. Anyone have thoughts about that?

Here’s an example of what a growing map might look like. I put some 45 degree subway lines in, not sure if people like that. (Looking at it again it’s not such a good example as if red is abyss and green is tide pool then they definitely shouldn’t be connected, but you get the idea.)

[DonGororo]

I also wondered about maybe different thicknesses of connection? So some region to region connections are really easy to cross, shown as thick lines, (maybe between two close regions of ocean) and maybe some are hard to cross, shown as thin lines, like into an underwater cave with a narrow entrance.

Then the entrance should be really narrow wouldn’t it? We are talking about microbes here xD
But seriously, this is an interesting idea, maybe ocean currents could gate the way to new patches?

Also, what would be the incentive for the player to leave their cozy sea vent? If I recall correctly, that kind of patches would have a lot of resources, but the ones above them would be more barren. Maybe once the vent gets overpopulated they will have to venture outside? Or maybe all patches will start with a high level of resources that will decrease over time?

[tjwhale]

Yeah Crodnu brought that up and I thought it was a good point, we should encourage spreading. I like the idea of an overpopulation penalty, where if you have a lot of pop in a patch it becomes harder to get more, which means you are incetivised to spread out, also yeah having starting resources in each patch means getting out early is good. So both nice ideas there.

[tjwhale]

Ok I did some more work on the prototype today. I implemented a new generation algorithm which has two nice properties, none of the connection lines cross (though they come close) and all the patches are certainly connected to the network, which I think are kind of important.

I think the improvements that might be good for this release would be things like:

• Nicer line drawing, maybe more like a subway map
• General visual improvements, maybe rounded corner boxes, nice colours, icons etc
• implementing rules about which patches can be connected to which

things which it’s probably best to leave for a later release

• different strength connections
• connections changing over time, forming, being broken etc

Is everyone ok with this being the design we go forward with? I guess this is probably a good time to check in case anyone wants a major redesign. Personally I am reasonably ok with it, I think it’s not quite visually strong enough yet but we can work on that, also I think mechanically it will be fine. Have a think about it and lets see if we’re all happy enough with it.

[MirrorMonkey2]

@tjwhale Nice work!

implementing rules about which patches can be connected to which

Funny that you mentioned that, because I just formulated a proposal for these rules:

Things like saline seas, underwater caves etc. can wait, but there is one problem I think we should solve in this release: Making the transition from deep waters to shallow waters more gradual.

Right now our virtual shores „look“ like this:

With the addition of a bathypelagic biome (or do we just summarize abysso- and bathypelagic as one biome?) this is the smoothest transition I can think of:

Now this is probably way more complicated than necessary. Somewhere in between these two images lies the optimal input/output ratio. The question is: Where exactly?

[tjwhale]

Yeah interesting.

I like the rules list of which biomes can connect, that’s cool.

I guess in terms of number of steps to move around that’s partly related to the species splitting discussion. If it takes 3 minutes to swim around and get enough compounds to reproduce and you can move to a new patch every 2 editor sessions then to move (vent -> abyss -> meso -> epi -> coastal) might take 24 minutes, which is a reasonable amount of time I guess. I’m not sure we’d want it to take that much longer.

Re ocean floor depths we could do it continuously, so an ocean floor biome can be at any depth from 1m to 4000m if we like, rather than chunking up with titles. It’s an option.

[DonGororo]

Maybe making those new ocean floor biomes is easier that it seems. We already have an ocean floor background so we’d just have to make 3 more versions of it, each increasingly lighter. Then in the patch map we could include one ocean floor patch for each ocean layer.

By the way, I’ve been reading about shallow sea vents and it turns out they are fairly comon in volcanically active areas, like Iceland or the Azores: https://news.nationalgeographic.com/2018/06/hydrothermal-vents-discovered-azores-science-environment/

So maybe they could appear in other layers besides the hadal zone, that would make colonizing the surface easier for phototrophs.

[tjwhale]

Here’s another version which has subway style connections. All feedback welcome, I am not so happy with how it looks so that would be a nice thing to work on. Part of that is the random node colours.

[DonGororo]

Looks really great! Yeah the main issue is the random colors, but that will be solved once we use icons with a coherent palette. I’ve been wanting to make a patch map gui concept for a while, I’ll see if I have some free time by the end of this week to to finally get to it.

[tjwhale]

Here’s some more progress on the map. I moved the connection lines to the sides of the boxes (but it’s not totally successful, I think its more complicated than I thought). I also managed to clean up the line crossings quite a lot so lines don’t go through boxes or other lines so much, though that could still be improved too. It’s suprisingly hard.

[OBPSG]

I definitely think that having a hierarchical approach will help with keeping things manageable for both developers and players. The main questions I have currently are what essential qualities each patch in a main group will have in common e.g. land or sea-based, general temperature range, etc; also whether we will want only one sub-patch in each group to serve as the link to an adjacent main patch, or if we want to allow more than one link to between two main patches.

[tjwhale]

Yeah interesting questions.
I think one of the regions I like the best is the big ladder which is like 6 high and has all the layers of the ocean in it (or a smaller cropper version of it). I think that’s a nice, obvious, thing to group together.

Not sure about other options, maybe having a coastal + estuary + a couple of tide pools in a square would make sense?

I think re moving if only one patch links to the outside that makes everything very linear, it’s nice if there are lots of possible connections I think, but we can experiment with it.

[MirrorMonkey2]

Only having one link between regions would create huge problems. If only the epipelagic regions were connceted to each other species without a resistance against low pressure or high radiation couldn‘t spread from one deep patch to the other. If only the deep regions were connceted photosynthesizers would face similar problems with high pressure and absence of sunlight.

[tjwhale]

I’ve managed to make a little progress on the patch map prototype, not a huge amount though. It now just draws the lines a bit more nicely.

The main things remaining are to make it a bit more efficient in how it generates the maps, I think I know how to do that. And then to put all the actual patches in so that they only connect in sensible ways, I am not sure how hard that will be.

[DonGororo]

Awesome! But I’m a bit skeptical about using diagonal lines. How about using right angles instead? You could also make the bends rounded, that way conections will look similar to boxes.

[tjwhale]

Yeah maybe, there’s some other stuff I want to do first but that’s something worth considering. I can imagine straight lines with rounded corners looking good.

[tjwhale]

Some decent progress today. I started with the graph colouring problem of trying to assign biomes to each patch. I just did 1x1 patches in each region just to make it easier. This is probably the best one produced so far.

The property this has which none of the previous ones have is that each connection makes sense, as in moving to a neighbouring patch is reasonable.

[tjwhale]

Ok so I’ve made some decent progress today. The map now generates actual worlds, so they definitely have all possible patches in them and all the connections obey the connection rules. (It’s got writing all over it so looks terrible but then you don’t need a legend to know what is happening)

Something to discuss. I set it so that whenever there is an ocean stack it’s always Epipelagic down to Abyssopelagic. We can change that, add more etc. I think think though maybe it makes sense that if you are in the ocean you can always go up, so if there is mesopelagic there is always epipelagic above that, does that make sense? (The epipelagic might get frozen into an ice shelf but is still there) We could have stacks of different heights, maybe have an ocean floor at the bottom of each stack, not sure really.

Basically what should the rules on stacks be? Can you always go up? Should they always have a different depth floor on the bottom? Should they always be the same size?

[Narotiza]

Looking good!

I figure you should have an epipelagic patch above every ocean section, mostly for the purposes of simulating marine snow drifting down when we get to that. Different depth stacks would be interesting.

By the way, here are all the (slightly updated) icons I designed for the patches, I don’t think I’ve posted them here yet:

https://drive.google.com/open?id=1ErXyWB87wOTQWyOBqRcpt7K6KF4ybKW2
It seems you’re using the hadopelagic icon for the floor instead of the abyssopelagic floor, perhaps I’d change that

I’ll get to posting the individual icons eventually, I was a bit short on time as of writing this

[tjwhale]

So I tried a totally different approach, as the one I was doing got quite fiendish (it still might be worth pursuing though I just wanted to try this).

How this works is it has a tile library of 25 tiles with different connectors on (Up, Down, Left, Right) and then it picks tiles at random and connects them together.

Compared with the other approach it feels a bit gamey maybe, like it starts to look a bit like a boardgame. However I like how fast it is and also I like that we could design all the tiles by hand. For example I really like the tile at the very top right at the start of the video, it’s a coastal with 3 tide pools coming off it, I think that’s quite interesting. Another one is the underwater cave connector which is two caves on a single tile in a line. I think they could provide interesting opportunities for life to evolve. Also it would be possible to draw nice designs on each tile and have special rare ones for deep ocean trenches or giant river deltas or something.

I think in general it will be able to generate pretty much anything the other prototype can generate, so I don’t think there’s a loss of complexity in general. The other prototype is essentially this with tiles with only 1 patch on. Also having the map change over time is quite doable as we can just swap tiles out for other ones with the same connectors.

I’d be interested in feedback on it. Do you like this idea at all, do you think hand crafting a tile library would be interesting?

Here is the current tile library I just threw together