Skip to content

Latest commit

 

History

History
46 lines (33 loc) · 5.56 KB

lugas7.md

File metadata and controls

46 lines (33 loc) · 5.56 KB

This is the file for Lucas Martin Thomaz with candidate ID (10008) and username lugas7 (also appear as foobar123asdf due to configuring email in git to be my@email.com) the code overview video can be found in the url below https://www.dropbox.com/scl/fi/9bdt7zbvgdec7m6ujtdxr/lugas7-code-overview-video.mkv?rlkey=yv8ncm21t5yxj0oj2ws1jphe3&st=5v5ghnqg&dl=0

Workload:

Lucas: Player: All Minimap: All SwirlEnemy: All CannonEnemy: All Bullet: All Sword: All Gun: All Boss: All World generation (room placement): Some (implemented connecting adjacent rooms, also made game reload after finish) Rooms: some+ (made new rooms that utilize different enemies and layouts, implemented unique boss logic for finishing the game)

following characters made by Kristian, but implemented within the larger game by Lucas e.g. setting the correct layer and masks as well as making the characters work as enemies for the rooms they are used in. aoe_elite_enemy: Some light_dash_enemy: Some

Reflection

Bad code:

sword slash state (slash_state.gd), it should be more modular by having the slash position set in it's usage instead of mouse position being tested within the slash state. This could be remediated by having the slash() method in Sword take in an angle as parameter and set a field parameter equal to the angle parameter, then this angle parameter could be accessed in the slash state. This would allow non players with swords to utilize the swords slash functionality. Though I found no use for the slash functionality outside player so no remediation means were taken.

We should have utilized an enemy class instead of simply setting the signal enemy_died multiple times. Each enemy could have extended our enemy class

The manager (manager.gd) is too large of a monolith and should have been partitioned into smaller modules. Even within the manager.gd script. I am largely happy with the implementation of the sword scene and gun scene as they are reusable in different characters, as they are used in both the player and the enemy. (Though one teammate made his own implementation of a gun/cannon scene for his enemy: aoe_elite_enemy which was unnecessary, another bullet scene was also made)

Another mistake in my code was to connect change in health component to health bar through the character, when this could have been done directly to both save resources and reduce complexity.

I am also unhappy with the placement of our files, they should have been followed using a group agreed structure for the different assets. There are far too many files in the root directory. This is ok since the god handles an architecture through the node/scene hierarchy within the different scenes. So you most of the time don't have to find the file by their location in the file structure, but when you do have to find them you need to search for them, which godot supports. Though this is a weakness the workflow in godot lessens the impact.

Good code:

I am happy with my implementation of sword scene, I made it easily reusable for different characters for both enemy and player (with an exception of the slash state mentioned previously). I am happy with my implementation of it in such a way that when you import the scene as a child of a character who will use the sword you can assign your own sprite and collision shape to it so that you can decide the size, look, damage based on the sprite and collision shape you set in the @export value of the sword.

I am also happy with my implementation of the health bar both in regard to how it looks being configured within Godot editor instead of clunkily with different sprites. I am also happy with the code of the health bar that allows for easy use. Though as mentioned previously it might be better for the health bar to depend on the health component rather than the parent character.

I also think my development of the Player is good and that I made good use of state machines to handle both walking, dashing, shooting vs slashing. I believe handling this functionality with state machines was a good choice. This is also true for the sword scene which also uses state machines to change between doing nothing, slashing or spinning.

Learning outcome

I learned a great deal of a programming workflow I was unfamiliar with through programming in this Godot game engine. I learned asynchronous logic and frame based function logic that was very interesting and new to the traditional sequential programs I am familiar with from earlier courses and projects. One instance of this is when I tried to implement a slash attack for my player that would move a sword 90 degrees in a circle. What I was used to earlier for solving such problems would be something like finding a function that holds the thread for the time until the slash is completed. However in Godot I you must implement logic that is updated each frame for such problems. What I did was find a value for how many radians to rotate per second, then rotate the sword by my rotations per second value multiplied by the amount of seconds passed since last frame, this is done using the delta value. This is better than a function that holds the attention of the sequence with say a time.sleep between each rotation as it is more modular by having a function that activates each frame. This way the game engine can calculate as many frames per second as possible allowing for more dynamic slashes depending on the resources available. I also implemented similar logic for other implementations such as the health bar.

Another portion that I learned was signal handling and activating functionality for different components of the game asynchronously e.g. collisions.