One thing to add to this, in this example I used an interface to make the connection abstract from both sides in code, but that's not the only way to do it. For example, in my course, which this example is from, I ended up using a Unity Event. The end result is similar, except that the fire display doesn't need to actively do anything.
I'm surprised you didn't make any mention to the methods explained on your 'Connecting scripts on the same object in Unity' video where you had a class that stored the actions an object could execute and other components subscribe to them. I think that way of doing things is perfectly explained in that video, but I would have liked you to talk more in depth about when to implement that style of method over the ones described on this video. Either way, brilliant information and lovely explaination. Thank you so much and keep them coming!!
Been doing gamedev for around 2.5 years and this is still something that I sometimes struggle with. It has improved the longer I’ve been coding, but sometimes still difficult to decide. Thanks for the video!
Yeah same here. I've learned to take advantage of events as much as possible wherever reasonable. Sometimes I would have a hierarchy of specific scripts being controlled by things a bit more general which are controlled by a manager script. The manager script would use the events thrown out by the scripts it manages to.. well.. manage them. lol
This is exactly what I'm dealing with right now for my school project. I'm either abstracting everything to eternity or can't seperate a class properly fearing of communication between parts will be pain in the ass in future. I guess practice will be the answer to this problem. I will pick a way and it will cause problems when I decide to change something in future and I will learn from it and improve my decision making in future projects. Not doing anything because of fear of failure is always worse than doing something and fail.
the text display in your video is aesthetically pleasing. And the you do a great job in giving options, not doctrines. That's rare in the coding scene :)
Your channel is insanely underrated! I also want to note that the blog's on the website are really useful too! When I am actually working in Unity I prefer reading than watching a video, so it's really helpful!
hey i'd be interested in a video about attributes in unity. ive seen some but dont completely know how to use them, create them, and most built in ones.
Hey man, this is a really great video that touches a topic that we struggle at work too! Just a quick question, at 4:04 you mentioned that Ground Check could be implemented via Inheritance, would it be better to make Ground Check as an Interface instead? Just figured that an interface is more modular. Thanks in advance ! :D
Thank you! Personally, I would want to use an interface when I want to create compatibility (i.e. two different classes can be treated as if they're the same, even though they're not) and inheritance to share functionality (avoid writing the same code twice in both scripts) so for me, it's inheritance. But, this only works if the hierarchy of the scripts is a good fit, and it may still be possible to do something like this with interfaces, now that default implementation is a thing. Just my opinion of course.
This is how I usually do Unity scripts: Usually if something that has movement its all in a movement script for that game Object and then a separate animations script that handles the animations. For everything else if its a global variable I use containers via scriptable objects. UI script for UI systems. Interactions are usually done on the Game Object's movement script unless its a global then I have a dedicated script that handles like a bullet hitting a game Object so it adds force. I like to have my scripts used in a global setting not just to 1 game Object specifically. A door script can have just a door script to every door or if its a portal then a portal script to every portal and have the scene loaded in as a string array and u pick what scene it will go to from that array etc. This is helpful thinking and I enjoyed the video.
I submit that anyone who thinks an entire game's worth of code should go into one script, doesn't yet know enough about what they're doing. This is why we have principles and guidelines and architectural patterns.
I separate feature and architecture, any code will have some specific cuisine, so brstvto contain what's unique in single places. For example, move is a feature, it takes input modifier and position and output back new position, it shouldn't know about anything else, the manager is the one responsible for linking input type, position and controller results to the move element, that keep bug seeking into two type, feature implementation bug, or architecture communication bug, it help refactoring because you either rewrite features or rewrite architecture, even when both need rewrite it's a clear separation of concern. In practice it makes your code more modular, but modularity is only as good as how many utils class your code depends on 😂 the math class sipping everywhere is inevitable.
Imo, single responsibility is best used when dealing with AI behavior. For most other scripts it's just about a free for all. But doing things like splitting up the jump behavior then allows you to adjust exactly how that jump behavior will work as well. Take it to the next step you can store the behavior as a SO which can be swapped as needed.
@@GameDevBeginner More than an idea, it's my usual approach, honestly I'd love to share my state machine architecture with anyone I come across, if you are interested. I have 0 talent for doing videos but I feel the SM would benefit a lot of people.
I've personally been exploring a composition architecture to go along with how Unity already does game objects and components. Unity Events help keep components ignorant, but I fear I'm going to end up over-using them...
In my experience Unity Events can be great for local connections, like on a prefab, and especially if you want to set things up in the Inspector. Personally, I think they work best when the connection is more likely to be unique. Otherwise they can be a lot of work to keep setting them up in the same way on different objects.
@@GameDevBeginner Ohhh I heard that, what’s the next alternative? Only one I have found that runs good on Mac on par is Rider, paid unfortunately ): But great with Unity
I'll probably try VS Code next, or whatever is offered up with Unity by default, if only for the sake of using what beginners will ultimately use too. I have heard Rider is good though.
So in your example, does the fireDisplay still subscribe to OnThrust action defined in IMoveable? And all IMoveable objects can invoke that action when they move?
The Action? it's a ready made delegate, with a void return type and that takes, in this case, a float. I only put system in front of it because I hadn't added the System namespace to the top of the class. See this video for more: ruclips.net/video/J01z1F-du-E/видео.htmlsi=RWUbPiRSqMOxD8UG&t=398
@@chizuru1999 It is and it's a shame because, even though VS Code has gruvbox themes, none of them look exactly like my visual studio one, which I love!
@@GameDevBeginner Yes bro. They feel totally different lol. And I guess its something to do with apple not allowing it. And its a headache to maintain the visual studio for linux/mac. Last time I checked it works on mono only. Thankfully I mostly use win for work 😁
Inheritance… in modern oop… that’s gonna lead to disasters untold. I’m not much of a Unity dev but can’t you just make a global function that can be referenced that checks for the entity state?
to be fer i use all kind off method's. for example the player movement for my game have movement,jump,ground check all on the same code and it's interesting for all the other codes. however some code in my game is like chin because i haven't though for better system for example on the game im working on is multiple game and many off my weapons i will have the same type off ammo, that means i need to have 2 codes one for the ammo type and one for whatever weapon the use the same ammunition. in general i prefer to make evry code impeding with each other in few words to work on it's own but unfortunately for that is not possible for my project the reason is because my game is complex and some system's have to refresh other codes one good example for this is the PlayerHelth for my game it have 3 or more code in the order to work's correctly 1. one code that it's just the player health. 2. that is hurt box. 3. whatever object dealing damages to the player my point no matter what wii do wean i make health for my player i have to use at least 2 code one for the health and the other in the order to do the damage or heal the player. and i don't even mention all the other part that effects by the for health like kill death ect, that it's crucial in multiplayer PvP fps Game.
From what I've heard here, I don't think I'd ever want to work with you. Instead of using the simplest principles to write good code, it seems like you write bad code and then put a lot of thought into justifying your misdeeds by misinterpreting those principles.
One thing to add to this, in this example I used an interface to make the connection abstract from both sides in code, but that's not the only way to do it. For example, in my course, which this example is from, I ended up using a Unity Event. The end result is similar, except that the fire display doesn't need to actively do anything.
I'm surprised you didn't make any mention to the methods explained on your 'Connecting scripts on the same object in Unity' video where you had a class that stored the actions an object could execute and other components subscribe to them. I think that way of doing things is perfectly explained in that video, but I would have liked you to talk more in depth about when to implement that style of method over the ones described on this video.
Either way, brilliant information and lovely explaination. Thank you so much and keep them coming!!
Been doing gamedev for around 2.5 years and this is still something that I sometimes struggle with. It has improved the longer I’ve been coding, but sometimes still difficult to decide. Thanks for the video!
Yeah same here. I've learned to take advantage of events as much as possible wherever reasonable. Sometimes I would have a hierarchy of specific scripts being controlled by things a bit more general which are controlled by a manager script. The manager script would use the events thrown out by the scripts it manages to.. well.. manage them. lol
This is exactly what I'm dealing with right now for my school project. I'm either abstracting everything to eternity or can't seperate a class properly fearing of communication between parts will be pain in the ass in future. I guess practice will be the answer to this problem. I will pick a way and it will cause problems when I decide to change something in future and I will learn from it and improve my decision making in future projects. Not doing anything because of fear of failure is always worse than doing something and fail.
the text display in your video is aesthetically pleasing. And the you do a great job in giving options, not doctrines. That's rare in the coding scene :)
I always think, if the scripts are small are modular I can use them for multiple projects. And then I start from scratch every time 😂
Yeah, me too. I start to feel like I am gonna forget the script if I just use the previous one
My favorite unity tutorial guy.
Your channel is insanely underrated! I also want to note that the blog's on the website are really useful too! When I am actually working in Unity I prefer reading than watching a video, so it's really helpful!
just started learning unity and your vids are a breath of fresh air, easy to follow and very helpful!
A unity scripting course? Sign me up!
hey i'd be interested in a video about attributes in unity. ive seen some but dont completely know how to use them, create them, and most built in ones.
Hey man, this is a really great video that touches a topic that we struggle at work too!
Just a quick question, at 4:04 you mentioned that Ground Check could be implemented via Inheritance, would it be better to make Ground Check as an Interface instead? Just figured that an interface is more modular. Thanks in advance ! :D
Thank you! Personally, I would want to use an interface when I want to create compatibility (i.e. two different classes can be treated as if they're the same, even though they're not) and inheritance to share functionality (avoid writing the same code twice in both scripts) so for me, it's inheritance. But, this only works if the hierarchy of the scripts is a good fit, and it may still be possible to do something like this with interfaces, now that default implementation is a thing. Just my opinion of course.
This is how I usually do Unity scripts:
Usually if something that has movement its all in a movement script for that game Object and then a separate animations script that handles the animations.
For everything else if its a global variable I use containers via scriptable objects.
UI script for UI systems.
Interactions are usually done on the Game Object's movement script unless its a global then I have a dedicated script that handles like a bullet hitting a game Object so it adds force.
I like to have my scripts used in a global setting not just to 1 game Object specifically.
A door script can have just a door script to every door or if its a portal then a portal script to every portal and have the scene loaded in as a string array and u pick what scene it will go to from that array etc.
This is helpful thinking and I enjoyed the video.
I submit that anyone who thinks an entire game's worth of code should go into one script, doesn't yet know enough about what they're doing. This is why we have principles and guidelines and architectural patterns.
I certainly wouldn't do it that way but if it works, it works: ruclips.net/video/cFRT9E0C3XM/видео.html
Very useful, Mr British Man, thank you!
Thanks for the tutorial!
He should be over 100k subscribers.... algorithm do you WORK
I separate feature and architecture, any code will have some specific cuisine, so brstvto contain what's unique in single places. For example, move is a feature, it takes input modifier and position and output back new position, it shouldn't know about anything else, the manager is the one responsible for linking input type, position and controller results to the move element, that keep bug seeking into two type, feature implementation bug, or architecture communication bug, it help refactoring because you either rewrite features or rewrite architecture, even when both need rewrite it's a clear separation of concern. In practice it makes your code more modular, but modularity is only as good as how many utils class your code depends on 😂 the math class sipping everywhere is inevitable.
Imo, single responsibility is best used when dealing with AI behavior. For most other scripts it's just about a free for all. But doing things like splitting up the jump behavior then allows you to adjust exactly how that jump behavior will work as well. Take it to the next step you can store the behavior as a SO which can be swapped as needed.
Neat idea with the scriptable object, would be ideal for that.
@@GameDevBeginner More than an idea, it's my usual approach, honestly I'd love to share my state machine architecture with anyone I come across, if you are interested.
I have 0 talent for doing videos but I feel the SM would benefit a lot of people.
I just did a Live Stream helping someone get into C#, that's my best way to give back.
I've personally been exploring a composition architecture to go along with how Unity already does game objects and components. Unity Events help keep components ignorant, but I fear I'm going to end up over-using them...
In my experience Unity Events can be great for local connections, like on a prefab, and especially if you want to set things up in the Inspector. Personally, I think they work best when the connection is more likely to be unique. Otherwise they can be a lot of work to keep setting them up in the same way on different objects.
Love out! Another great video, on a random side note, what IDE do you use? Looks very clean for Mac which is nice.
Visual Studio, which is unfortunately, now discontinued on mac 😭
@@GameDevBeginner Ohhh I heard that, what’s the next alternative? Only one I have found that runs good on Mac on par is Rider, paid unfortunately ): But great with Unity
I'll probably try VS Code next, or whatever is offered up with Unity by default, if only for the sake of using what beginners will ultimately use too. I have heard Rider is good though.
So in your example, does the fireDisplay still subscribe to OnThrust action defined in IMoveable? And all IMoveable objects can invoke that action when they move?
Yes, and yes. Anything that implements IMoveable would need to include it.
Celeste was coded as one big class afaik.
Really? cool!
Great ❤😊
What was that system. thing?
The Action? it's a ready made delegate, with a void return type and that takes, in this case, a float. I only put system in front of it because I hadn't added the System namespace to the top of the class. See this video for more: ruclips.net/video/J01z1F-du-E/видео.htmlsi=RWUbPiRSqMOxD8UG&t=398
Can you please share what font and theme you use in jetbrains? Awesome video btw.
So it was visual studio, gruvbox theme, and I think the font is menlo 👍🏻
@@GameDevBeginner ye i thought its gruvbox, i use in vscode but isnt the mac version for vs discontinued? Thanks
@@chizuru1999 It is and it's a shame because, even though VS Code has gruvbox themes, none of them look exactly like my visual studio one, which I love!
@@GameDevBeginner Yes bro. They feel totally different lol. And I guess its something to do with apple not allowing it. And its a headache to maintain the visual studio for linux/mac. Last time I checked it works on mono only. Thankfully I mostly use win for work 😁
Inheritance… in modern oop… that’s gonna lead to disasters untold. I’m not much of a Unity dev but can’t you just make a global function that can be referenced that checks for the entity state?
Reupload less than a day in?
no? are you asking because of the thumbnail? RUclips lets you test them.
@@GameDevBeginner if the thumbnail changed it might be why
That'll be it then. There was a clear winner so I ended the thumbnail test early.
to be fer i use all kind off method's.
for example the player movement for my game have movement,jump,ground check all on the same code and it's interesting for all the other codes.
however some code in my game is like chin because i haven't though for better system for example on the game im working on is multiple game and many off my weapons i will have the same type off ammo,
that means i need to have 2 codes one for the ammo type and one for whatever weapon the use the same ammunition.
in general i prefer to make evry code impeding with each other in few words to work on it's own but unfortunately for that is not possible for my project
the reason is because my game is complex and some system's have to refresh other codes
one good example for this is the PlayerHelth for my game it have 3 or more code in the order to work's correctly
1. one code that it's just the player health.
2. that is hurt box.
3. whatever object dealing damages to the player
my point no matter what wii do wean i make health for my player i have to use at least 2 code one for the health and the other in the order to do the damage or heal the player.
and i don't even mention all the other part that effects by the for health like kill death ect,
that it's crucial in multiplayer PvP fps Game.
From what I've heard here, I don't think I'd ever want to work with you. Instead of using the simplest principles to write good code, it seems like you write bad code and then put a lot of thought into justifying your misdeeds by misinterpreting those principles.