Hi! I'm Peter, a 15 year old hobbyist game developer from Scotland. I'm currently using Unity to build my pixel art survival sandbox, inspired by games such as Terraria and Stardew Valley. I make devlogs showcasing its development as well as giving some useful tips where I can.
Ehh, probably not (at least I'm not working on it at the moment). I'm kind of focusing on my main project at the moment and it would probably take up too much time.
Trust me, it does! A lot of the stuff I wrote (especially from back when I started the project) is just so unmanageable, which is why I'm planning to rewrite the whole codebase once the demo is finished.
Very good video and the game looks so solid! I'm trying to get something similar done, altough I'm only uploading shorts for the devlogs and my game is an earlier stage. I do have very little time to work on it unfortunatelly. =( Best of luck on the process!
The colors look so good; you obviously know a thing or two about color theory. Have you considered doing some simulation calculations with DOTS or the Jobs API?
Thanks! Nope, haven't been all that bothered with the whole ECS and DOTS stuff. Since the game is 2D and the textures are low resolution the performance tends to be 'good enough', but I'll definitely look into it.
A* does not rely on a grid or a finite set of nodes. It is a specific implementation of the best first graph search algorithm applied to pathfinding, but the underlying algorithm is used in symbolic AI, such as most chess solvers. (its also used in google maps and unity's pathfinding system) A priority queue might sound daunting - but in reality its just a sorted list, that stays sorted when you add nodes. (C# already has an implementation of this) Heres the simple explanation of how to implement it: Beginning with only having explored the starting node, you repeatedly explore the best node you've found so far. To explore a node, you calculate the score of each unexplored neighbour and add them to the priority queue. The score of a node is its distance from the starting node plus the estimated distance to the goal node.
If you're wanting some good information on doing on-the-fly pathfinding, you should read some of the dev vlogs for Captain of Industry. They talk about it extensively, especially because that game has the capability of allowing a player to have hundreds of vehicles that are all trying to figure out where to go and what to do at any given time (even figuring out if it should give up on a destination without getting there immediately). Another thing you could think about is using signals for animals to communicate, in a sense. But mainly just position information to any animals within X radius of them. That would dramatically increase your A*/pathfinding performance.
Does anyone else feel that some of these devs seem to have more experience with memes and video editing than they do with game design? Not saying you aren't making progress, but it is mildly annoying to hear people say "I have no idea what this coding/gaming concept is, but I know all of these memes and I have fancy transitions and probably paid for my editing software".
Taming would be cool to see. I thought since the animals have memory, you could tame an animal based on objects that they love carrying? So, players would have to observe an animal for a bit before being able to tame it.
Hey dude heres a suggestion you should add some insects and grass so that something like a grasshopper jumped out of the grass when u stepped on the tall grass
The game looks very polished, Im amazed! I immediately subbed! The forest looks very alive, cant wait to see more. The only complaint I have is the Slime enemy, it is so generic and overused in other games, I think you have to make some really unique enemies (also no bats please!) For example, you can make fome forest creatures, like Orcs, Trolls, dark Fairies, Trents, etc. This game has a potential 🎉❤
Just FYI you do not need a statically computed grid for A*. It's sufficient to have a function that takes a position and returns the list of adjacent positions. So you don't have to page anything in or out as you move around. The main things to worry about with a homegrown pathfinder are escaping a deep bucket-like shape where the unit has to move away from the target to make progress, avoiding runtime explosion due to retracing in mazes, and identifying that a location is unreachable without searching the entire world.
It's also worth noting that A* is great for vision based checks, almost like Ray casting. And the algo only needs to be run if the creature/thing is trying to get to a specific location. But just for basic, "this is what I do daily" behavior, you don't need A*.
+1 for the A*. I've used it many times even on infinite procedurally-generated grids, and it works perfectly. To prevent searching infinitely long for a path to an unreachable destination, you could do bidirectional A*, which expands vertices simultaneously from the start and the end. If the destination is unreachable, e.g. is a closed "rom", then the "backwards" expansion will visit all the nodes in a room and realize there's no other vertices to move to. Bidirectional A* is also faster in normal situations, because it tends to expand less nodes in total.
This looks so freaking cool dude! Maybe I’m thinking to much into this … But will only the animals dissolve? Or will all mobs- Would the mobs drop bones and other things?
Yep, definitely overthinking it! Absolutely not realistic for anything to just dissolve, but I figured it would look cool and it saves me from drawing out a proper death animation for everything.
4:56 I think if you're making a believable ecosytem, the player shouldn't be able to see what the animals are thinking of doing. They are wild and should be somewhat unpredictable.
Yeah, I get what you mean. I might make it so that it is only visible when equipping a certain accessory or something whenever I get around to adding equipment.
Genuine question - can I hire you? I want to develop a game with it's own creatures and if they can have their own lives and interact with the player and NPCs that would step it up to another level.
I believe that highlighted sentence just means that on every step of the algorithm, we take out the node with the lowest f(x) value [the formula is on the same page above] and then update the neighbor nodes (since a priority queue keeps the smallest value om the front, we are sure to get the smalled one every turn)
6:10 I see that u use the exact same health bar on world enemies as Terraria does, try to give it some of your style to get it more unique or in your game style. It is not wrong to get some inspiration, but here it looks like 1 to 1 copy, so be aware of that. Besides that great work, and good luck on future content!
