Thanks a lot for your feedback Mekamani! I appreciate it.
TL;DR Start
I've done this once before on a student project, but it was limited. Trying to add robustness for usability in as many different types of games as possible.
Interfaces let you disregard what objects/algorithms one uses and define relationships based solely on the type of interactions they have. This would let me define my own framework, supply my own sample set of classes to actually fit the framework (implement the interfaces), but also leave other developers open to build their own versions should they want to.
I want to minimize code duplication and have the capacity to fully dynamically define the behavior of events in a game.
Ex. pressing "A" does (1). Pressing "B" does (2). (2) changes how (1) works without explicitly requiring me to create a new function/object for (1)'s functionality.
Ex. I press "space" to trigger a skill. I have 3 skills (A), (B), ( C ). I have 1 effect (1). I have 3 ways in which (1) is delivered. (A) uses delivery method (a), (B) uses (b), ( C ) uses ( c ). (A), (B), and ( C ) should each rely on the same function for (1), but should have 3 different objects responsible for the delivery based on (a), (b), and ( c ), minimizing code duplication. I should also be able to easily switch out (1) for some other effect (2) and suddenly get 3 new skills that each cause effect (2) in 3 different ways (a), (b), ( c ). I should also be able to make it so that if a skill ( C ) has multiple stages to it or if a skill (A)/(B) is used, I have the capacity to influence the functionality of the (1) or (2) that is in ( C ), all without changing the underlying implementation of (1) or (2).
TL;DR End
Now the second point and I have to admit that I haven't used unreal engine much, so it might have some underlying classes that one must use in order to accomplish something, but.... Lets say that this is the first time you are building the system, I honestly think that you should try keep things as simple as possible. If on the otherhand you have already made one version and you are trying to improve your previous design, then you can just skip the rest.
I actually already worked on a much simpler version of this skill system for a UE4 project at Baylor (it is used in the Steam Greenlight tactical turn-based game SYNCH). While I was able to do some nice things with it in that project, I often found myself in situations where features that we wanted to have in the game turned out to be very difficult to do with the given framework. As such, this is my current attempt to redo the framework's construction to be more viable for a variety of scenarios.
Anyways the last chapter thoughts came from seeing so many interfaces, that I cannot comprehend how complex of a system the pdf looked like from first glance. Or highly possibly I just do not understand what an interface is.
In the simplest terms, an interface just lets you define the behavior you want without concern for the underlying implementation. That is, I can build a framework for the system that I want without making requirements about how anyone actually implements it, i.e. what objects/algorithms they decide to use to fit into the framework. The ubiquitous interfaces were my attempt to decouple a sample implementation I was planning to make from the design that would be publicly visible to the UE4 system. I wanted to build something that anyone would be able to use and, if desired, make their own by creating their own implementations.
Anyways what I was thinking when I read your post and quickly glanced your classdiagram was, that you haven't given us any information what you are actually trying to do. Basically there is no information of what it is that you are actually trying to do. And at least I do not know what you mean by skill system, the definition could be pretty much anything, but judging from the class diagram it has something to do with spells/player attacks or something?
Ah, I see what you mean. I initially thought that people would understand what I meant by a "skill system," but I see that was a poor assumption on my part.
I tried to have a short summation of the thought process for the skill system in the diagram itself (big Note object in the bottom-right), but I you have to look in the diagram to see it, you may not notice it, and it may not even be good enough at explaining things...
The core usage I had envisioned was for RPGs or tactical turn-based games (like the one I helped produce), but I was also intentionally abstracting the concepts to hopefully make it viable for a large variety of game genres like shooter, action/adventure, RTS, MOBA, etc. Fundamentally, I want a system that allows me to modularly put together "abilities", that is, the culmination of a discrete set of manipulations from one game object upon 1 or more other game objects, and which provides a powerful and robust method of delivering these abilities' manipulations ("effects"). In addition, the system must make it possible for the specifications of "effects" to be variable and influenced by other effects.
For example...
My character has the ability to press A and shoot a bullet in a forward-facing direction. They can also press B to "charge", an action which leaves them unable to move/attack for 2 seconds and then it modifies the effect(s) of the "A" skill. It 1) increases range and 2) adds an area-of-effect element to the bullet. So, if the person presses B, charges, and then presses A, they will have a projectile that hits at farther range and hits several enemies in a given radius rather than just a single enemy at shorter range. This would be an example of where one skill's "effect" is variable and can be influenced by another "effect."
Presumably, ANYTHING would be able to layer on these "modifications" to skill effects though, so power-ups, equipped armor, skill tree bonuses, etc. You could even create a different skill entirely by just having permanent skill modifications to the existing skill. Indeed, if I wanted the second effect to be its own permanent skill, I wouldn't need to actually create a whole different skill that has the exact same effect with a few different stats. I could simply never remove the modifications to the skill (they are essentially added components of the skill that acts as an entity). In this way, I need only have the code that "deals damage to another game object" be in one location: the DealDamage SkillEffect. If I edit the DealDamage object, I can change the way ALL skills deal damage. It is also fully dynamic, so I can add or remove damage dealing from any given skill should I desire to.
Another example...
I have a bullet (deals fixed damage on impact), a fireball (deals fixed damage at regular intervals for a fixed period, i.e. a Damage-Over-Time effect), a lightning bolt (deals damage to one target, then finds another nearby target and deals the same damage to them, triggering up to 3 times, i.e. jumping twice).
The things these each have in common is the singular effect, DealDamage. They each differ in the delivery of that effect. I should be able to make it so that the underlying implementation only differs in the object that is responsible for distributing the effect (as would be possible - ideally - using the SkillDistributor object in the diagram).
In addition, I should be able to fully manipulate the effects and modifications for these skills. For example, I could take the lightning bolt skill and change its effect to be something that slows down enemies (so each hit object has a 50% "Slow" status condition applied to them - 'status conditions' aren't currently a part of the diagram, but it's a viable thing to be added). You could then also make it so that each successive "hit" resulted in an added 10% effectiveness, achieved by having the skill's own effect apply a mod to itself after each hit causing a change to its variable attributes.
I should also be able to switch up the way in which the skill is distributed (a volumetric area, a projectile, just targeting the self, etc.) simply by altering which targeting method is supplied to the delivery object for a skill.
I hope that made it slightly more clear.