From the research started on this thread:
https://monero.forum/thread/monero-tower-defence-game
The current thread will be focused on Canvas2D, as started by @lIlIIllIIIIllllII on post #10.
the proof of concept looked like this:
It was in ECS Rust, fully custom from scratch without any dependencies.
(unlike the WebGL proof of concepts that were using things like Three.js, Babylon.js and so on as the base for game engine, that one was fully done from scratch without any existing engine)
Also was fully procedurally generated from Blender through python scripts without using any actual rendered assets.
Unfortunately, when trying to move to a spritesheet based approach with actual 3D models rendered into spritesheets, then that's when it became troublesome.
So that's what is being revisted here, instead of starting of with the game, starting of with a browser based "Model Viewer" (inspired by "WoW Model Viewer").
It's been a decent approach regarding the WebGL proof of concepts with the full 3D models being used in either Three.js or Babylon.js from their .glb or .gltf files, making sure they render properly, that their animations are working and so on.
As of right now, what is being worked on is the "Model Viewer" for Canvas2D, so similarly as the earlier proof of concept game, it doesnt use any dependencies and is being written fully from scratch.
And instead of loading the 3D models directly, it's loading the spritesheets, so for example, 17 frames running animation looks like this:
(forum upload is lowering the resolution, quality used is currently higher but same principle)
dang actually lemme try again lol, cause cant see shits rn on the upload downscaling things lol
so like, individual frames like the one below, from different angles, with a script that makes them into spritesheets and it does the animations with 8 different angles for the isometric thing pretty much:
So at 25 frame tick per second, a 75 frames animation (like the 2nd animation playing in the video (AoeCast)), would be a 3 seconds cast time.
With items or specialization tree, cast times could be reduced as well, and for the animation, skipping frames evenly across the animation, still playing at 25 frames per second would also make it play faster.
Eventually with different characters and more work then "Item Layering" could be done as well, but this is kinda complex already so for the proof of concept, there won't be none of that, just characters and their animations from Mixamo:
https://www.mixamo.com/#/?page=1&type=Character
So for now the idea for items is just that they wont show on character model and wont be rendered at all, but just from character inventory similar as the "WoW Armory" or any characters planners like:
https://maxroll.gg/d2/d2planner/rag0mp0u
or:
https://maxroll.gg/d3/d3planner/46449470
and things like that
(which i guess, that part could probably even be done without Javascript, so there would also be a way maybe that users from Tor with NoJS could manage their characters and do the trading from an UI that doesnt require Javascript)
so pretty much where it's at for now, not much progress, started again few days ago, still havent started on the game itself and still on the "Canvas2D-ModelViewer" part of the thing
seems like it gonna be troublesome already, like, having the character scaling normalized across all animations or whatever... idfk... feet positioning or whatever... idk... lots of headaches coming in for even basic things lol
for the game itself, for now the idea is... AutoChess/Roguelike kinda? like, Teamfight Tactics (the rito games League of Legends AutoChess thing) but also Roguelike for dungeons etc, so like, autobattler maybe, so on that it will be troublesome as well for the automation, pathing and pretty much everything
idk, still TBD, but it could probably be a thing
like, that way it could have PvP from players facing other players (similar to AutoChess) and then dungeons, raids and so on from Roguelike type thing with like going to the next room, do the PvM fight, then next room and so on, could be a thing, idk, just an idea for now but could be a thing idk lol
Focus would still be around items and trading.
Better gear meaning faster clear times meaning more drops meaning finding better drops.
But for now, focus for this thread is Canvas2D and just getting a base thing to a working state then the rest will come later, the model viewer working with the spritesheets is already a first step, will see how it goes.