This is extremely helpful. Thank you! It's also really reassuring to know that a ton of the practices and fixes you both are making are things I'm already doing either through other tutorials over the past few years or my own trial and error.
Just so you know, transform *is* cached by Unity. There is no reason to create your own cached reference to the transform except for shorthand purposes.
Does the guy on the top have a RUclips channel, for some reason his word choice is extremely understandable when he explains coding concepts. This solidified some of my understanding.
When your code is “tuned” it feels good. Knowing what you need to do is one thing, but actually getting on, and doing a proper tidy up, is the important step. If you never tidy up, you will waste more and more time reading over stuff. Each bit of new progress involves the extra work of reading over the untidy code, The desire to make “progress” often gets in the way of the “tidy up”. Some coders are lazy, and don’t care until they see a performance hit. Thanks for all the great tips. You guys are making a real difference here.......respect!
This was extremely useful for me, even as a beginner. I'm writing what you call monolithic codes all the time, just because I'm not skilled/educated enough to think about good structure early in the process. I just want stuff to run. These refactoring videos are therefore very helpful! Also some smaller tips, like Unity.Random (booom?!)... loved it. Plus: you both come across very "explaining" rather than "telling what to do", which is much appreciated!
Pro tip: watch the 'clean coding' videos of Uncle Bob Martin and make yourself a better programmer in general. Far too few people take clean coding seriously, and if you're a coder, you should.. it will save you quite a few walls to run into. You're welcome 😀😁
@39:00 "what is your first best practice that goes out the window when you're in a hurry" That's such a good question. for me personally single responsibility is what typically goes out the window under tight deadlines. Because often its quicker to bundle a bunch of responsibilities and logic into one class and one object and keep it tightly coupled to itself vs creating modular reusable components.
Noticed you replacing the types with *var* at 25:37. Microsoft recommends against it in that case because the type of the right side of the assignment isn't clear. It was correct as it was.
I think, that invertion of the if block is not refactoring (7:30) and it doesn't make the code better. There's nothing wrong in having a "pyramid" of ifs, it doesn't make reading any harder. But i allows to follow the positive scenario and saves us of from an odd return which is somewhat similar to goto. If I remember right, Code Complete suggests to follow the positive scenario first instead of checking all possible problems and then (finally) doing someting useful. If you have a "pyramid" you see error handling in the very end of a method and it's nice because probably you don't want to read them at all.
Nested indentations is harder to read them a flat structure. Though I agree about preferring positive cases whenever possible. You can also have the best of both worlds by extracting a new method with the contents of each nested if.
Good job, guys! One thing I would like to mention is that, as far as I understand, one should use rigidbody.position and rigidbody.rotation while manipulating Rigidbodies.
Wonderful initiative that you guys have taken. Most freelancing programmers terribly need it. Please keep doing this. Wish you best of luck in all your efforts
An awesome way to practice is actually code reviewing your own code/legacy code. There's code I wrote 6 months ago that I review now and just sit there asking, "But why?". And this has stayed true for the 3 years I've worked full-time in unity. BUT, please don't go overboard on this because you can fall into the trap of never progressing forward as you're just refactoring constantly 😂 Fun excersize though!
As far as i know, transform IS cashed in Unity now. And has been for a long time. I think they did it somewhere around Unity 5 or something... You can test it if you're curious - there's zero difference between caching transform in awake and calling transform every update.
Thank you for this video. At my company I'm sort of isolated from other developers. There's nobody to ask "how would you do this code?" or "is this a code smell?". Not even 10 minutes into the video I'm learning about guard clauses and cyclomatic complexity. Guard clauses are something I've been doing for a while now to avoid "arrow code". I always felt a little worried that it was a code smell for the exact reasons you described here and now I know I can safely use it by extracting everything that follows the guard clause into a separate method. I'm sure there's many more things for me to learn in the next 50 minutes. If you Infallible Code or anybody else who reads this have good yt video resources on best practices, please recommend them to me. Thank you!
Hey guys, I'd love to know why you use "var" instead of saying what type the variable is. I was always told abusing var is a bad idea and if working with a team it could be "what is the var? Is it a string? An int?" - you get me. So why use it?
Well lets think about what var does, it removes the name of the return type and replaces it with a shorter cleaner name. could that be confusing? definitely but a common rule with software is the smaller the scope the longer the variable names, it gets to the point where variables have SomeVeryLongNamesInCode to describe their jobs. the thing is though, variables are a 2-stage process. there is the "declaration" and the "assignment/initialization" usually this means you are effectively writing the name/type of the variable twice. Dog myDog = new Dog(), etc. so in one of the above examples: var force = RandomForceVector(); we are writing the word vector... just doing it once, not twice. Var can be confusing if you have no reference visually to what type is. but is usually fine if you do. Lastly personally I don't worry about var because we live in a world of very powerful editors, and you can always hover over a var to see its "real" type, but in most cases you don't need an extra 3-4 words per variable at all times. var massively shortens your code and with vs or rider, you can literally press one keyboard shortcut on it to revert to a fully qualified name. TL;DR the goal of 'clean' code is readability. in my opinion I try to remove as many words from my code that do not add useful information to the person reading it, and if I named my variables well enough they should not need to know what the type is to read or use my code. even if they do.. its not gone, its one editor click away
The only thing left about that GenerateRandomDirection part at around 35:30 that is left to do is to normalize it and multiply by constant of 10f (it seems) Then throw it all in a garbage can and just use Random.onUnitSphere * 10f. *I know probability distribution would be different but it clearly makes sense there.
I don't know why, but no matter how many times I watch these videos, and no matter how many times I think I understand it, I just can't seem to apply it to my own code.
2:38 Ah, the infamous, nagging, overly intrusive alert sound of windows. _Virtually always_ heard multiple times in a row. Hearing it echo in the background never fails to raise my blood pressure.
I have to say I hate underscore. It takes two hands to write code if I use underscore and it doesn't look so clean. I use them only for properties to tell myself I shouldn't use underscored var. Also I never write private because its very easy to identify myself what is private and what not if modifier is missing.
Learning a lot from this, thank you. Re. the guard clause, I avoid them generally in C# and C/C++ for the reason you mention - they make code underneath harder to maintain because you de-localise flow control and have to mentally track it as a dependency. I think wrapping the code below up in a new function is messier and less efficient than just using the super-clear if slightly indented if tree.
Another Question! You guys are awesome! hahaha Thanks for that content! An Empty Update() Function on a Monobehavior is "running"? Like "Update() { };" Is a good practice to remove empty functions like that?
I'm not sure I also want this answer. I would never do this but as far as I know, there are no benefits for performance because the update is a method from the monobehavior and its directly called from the parent :) Hope someone will answer so I get the notification :)
@@mehmedcavas3069 Back then I worked on something that involves on instantiating grid tiles on a level, like about 50x30 grids, with the grid prefab having a MonoBehaviour script in it. At first after everything's done I got pretty low FPS but it turns out that leaving the Update() method even though I'm not using it just because I don't know better is the cause, removing it makes the game buttery smooth again. In summary, remove Update() method indeed when not using it, hope this answers it!
@@Aditya-us1qn unity creates a list of all objects that has update, lateUpdate or fixedUpdate...and then runs through it every frame...meaning even if its empty, it's still added to the list... So yea...if u don't put it in the script, it won't be added to the list...
Oh wow. I've been coding in Unity for about 4-3 years. I just learned about "foo = foo ?? bar;" Is it just "foo = foo ? bar : foo2;"? But without requiring an "else" after the colon?
As a person that writes a lot of code. I would of told this person, extensions, the transformation and rigidbody will be used a lot like that. An extension makes sense.
Though I understand how crazy computing power is, they are unknowingly making a case for their insistence on following good practices. They keep finding minor optimization after minor optimization. Many small inefficiencies can start to become large.
Using ?? with components is problematic (probably ok with the camera though) since it is not aware of the null check override Unity is using. Null coalescing operators cannot be overloaded, so Unity cannot change their behavior. They only do a pure null check, they do not check if the object has been destroyed. That is still an ongoing discussion on the forums and a current weakness in the Unity API.
Certainly lots of good practices. Just two questions: a) Why make a simple call to Random a whole new method? Is it because the method could be expanded upon later? I didn't quite understand it if you said it in the video. b) Though I too am a user of WaitForSeconds, I find it has one major flaw; it creates an instance of a class that needs to be garbage collected. Furthermore, an input of 0 would still yield, which would mean a loss of at least one frame. Wouldn't an ultimately better solution be something like this (not checking for negative time for simplicitiy) public IEnumerator WaitForSeconds(float seconds){ float time = 0; while (time 0) WaitForSeconds(seconds); if you don't want a "missing frame". EDIT: Unless a yield return null also continuously allocates memory.. Wow, I really need to profile this one day... Anyway, I always found coroutines very intriguing and a fun way to build time-based systems in unity. I've even built an animation system based on them, since the primary function of waiting is a really powerful tool.
Hey, About using Debug.Logs When you leave a lot of Debug.Logs in the code, and then build and run the program outside the Unity Editor, we'll have memory usage problems because of that logs? Or it is heavy only on the editor?
This is heavy in the editor only 😊 When doing development builds however, you can see error logs and hook up the console to the built application. But that's a separate thing.
The fact that you are so uncomfortable seeing prefixes on functions is a great reason to use them for things like the coroutine here. That discomfort means that if you inadvertently attempt to use it as a regular function you will definitely see the prefix and realize the mistake you've made before compile time.I find that doing stuff like this that I don't like is a great way to catch my own attention later on when I return to a piece of code and don't remember everything that it should be doing.
You mentioned to refrain from too many if statements in Update. What else are you supposed to do? If you have a jump input for a player, I could put the jump function in Update, then the function still has an if for the jump input. How would I check for a a player's input if not checking for it every frame?
If you cache it yourself, yes it is slightly faster but you could do that for other unity objects too. You don't have to and that's why unity allows transform. now
One practice I've taken up for differentiating coroutines is to just put an underscore prefix. So the method discussed in 30:00 would be _EnablePhysics(...) I found that it's pretty easy to understand from seeing it, as functions that start with underscore are very rare otherwise. But the main benefit to me is that it just looks weird if you call it on its own. But you get used to reading something like StartCouroutine(_EnablePhysics(...))
Apparently GetComponent() has a non-generic overload which is slightly faster, it is mentioned here: chaoscultgames.com/2014/03/unity3d-mythbusting-performance/
Here's the link to the talk mentioned during the video from the folks at Playdead: ruclips.net/video/mQ2KTRn4BMI/видео.html Still a great talk years later.
`this.transform` isn't cached- it's hard-coded to directly return the Transform. This is possible for Transform (but not most other things) because every GameObject strictly has one Transform (or derived RectTransform) from initialization. Most components are optional; Transform is not. The full story is that a while back, Unity got rid of all the component accessor properties that were expensive, leaving only the ones that are cached and as efficient as any code you could write (which is just `transform` and `gameObject` in the latest version). So if Unity provides an accessor, use it, otherwise make your own cached accessor.
Some things being commented and some not, regardless of if its self evident, is a pet peeve. Not because it doesnt make logical sense, it's entirely a personal for things to match. lol
Personally, i think "this.GetComponent" etc. is more readable than single "GetComponent ", just personall preferention to put whichone MonoBehaviour is calling the method. But good video to practice.
I learned a couple of things, like the log context. But a guard clause in Update is an absolute red flag, 99% of the time Update is abused and this is one of them. The _shouldAssemble boolean should have been removed with the update function altogether, MoveToStartingPosition that was extracted from Update should have been made a coroutine called from StartAssembly. Also the StopAssembly function that you added don't make much sense , I often see code with classes exposed too much, for exemple putting a public setter in every properties by default.
late to this vid, mostly loved right up to the summary/comment argument xD my stance is you are NOT a good judge of if your code is readable or not. You wrote it, you're automatically biased, you have context you don't even realise you have. "i've written this clean, its obvious what it does" is not helpful to the confused person 2 years later. I'm not saying comments should be everywhere, im just saying "code is where your code goes, it should be obvious what it does" is a mcguffin. Also, the most useful thing summaries do in c# is power intellisense - i often use them just for paramater decoration if nothing else!
![Unity async / await: Coroutine's Hot Sister [C# & Unity]](http://i.ytimg.com/vi/WY-mk-ZGAq8/mqdefault.jpg)
👉 Submit your code for a chance to be reviewed on my weekly Game Dev Livestream at infalliblecode.com/code-review
Can we get a link to James' channel/@'s?
@@mysticvalley4566 0
00
@@mysticvalley4566 0
This link doesn't seem to be working for me
This is extremely helpful. Thank you! It's also really reassuring to know that a ton of the practices and fixes you both are making are things I'm already doing either through other tutorials over the past few years or my own trial and error.
Just so you know, transform *is* cached by Unity. There is no reason to create your own cached reference to the transform except for shorthand purposes.
Does the guy on the top have a RUclips channel, for some reason his word choice is extremely understandable when he explains coding concepts. This solidified some of my understanding.
One of the best 1 hour spended in relation to coding practices. Thanks
When your code is “tuned” it feels good. Knowing what you need to do is one thing, but actually getting on, and doing a proper tidy up, is the important step. If you never tidy up, you will waste more and more time reading over stuff. Each bit of new progress involves the extra work of reading over the untidy code, The desire to make “progress” often gets in the way of the “tidy up”. Some coders are lazy, and don’t care until they see a performance hit. Thanks for all the great tips. You guys are making a real difference here.......respect!
Good Job guys, this is a great video for solo developers that never go through the process of Code Review by a fellow Dev!
Really helpful, reasonable and practical. Love how you explained THE REASON WHY the best practices are actually better!
This was extremely useful for me, even as a beginner. I'm writing what you call monolithic codes all the time, just because I'm not skilled/educated enough to think about good structure early in the process. I just want stuff to run. These refactoring videos are therefore very helpful! Also some smaller tips, like Unity.Random (booom?!)... loved it. Plus: you both come across very "explaining" rather than "telling what to do", which is much appreciated!
Pro tip: watch the 'clean coding' videos of Uncle Bob Martin and make yourself a better programmer in general. Far too few people take clean coding seriously, and if you're a coder, you should.. it will save you quite a few walls to run into. You're welcome 😀😁
@39:00 "what is your first best practice that goes out the window when you're in a hurry" That's such a good question. for me personally single responsibility is what typically goes out the window under tight deadlines. Because often its quicker to bundle a bunch of responsibilities and logic into one class and one object and keep it tightly coupled to itself vs creating modular reusable components.
Noticed you replacing the types with *var* at 25:37. Microsoft recommends against it in that case because the type of the right side of the assignment isn't clear. It was correct as it was.
Absolutely fantastic video. A lot of useful insight. I Especially like that there are two of you bring different ideas and perspectives on things :)
I think, that invertion of the if block is not refactoring (7:30) and it doesn't make the code better. There's nothing wrong in having a "pyramid" of ifs, it doesn't make reading any harder. But i allows to follow the positive scenario and saves us of from an odd return which is somewhat similar to goto. If I remember right, Code Complete suggests to follow the positive scenario first instead of checking all possible problems and then (finally) doing someting useful. If you have a "pyramid" you see error handling in the very end of a method and it's nice because probably you don't want to read them at all.
Nested indentations is harder to read them a flat structure. Though I agree about preferring positive cases whenever possible. You can also have the best of both worlds by extracting a new method with the contents of each nested if.
your collaboration made me subscribe, for the first time ever! i´d love you two to work together more often. this was really nice!
Keep it up. Best Unity Coding Channel on RUclips
Good job, guys! One thing I would like to mention is that, as far as I understand, one should use rigidbody.position and rigidbody.rotation while manipulating Rigidbodies.
Wonderful initiative that you guys have taken. Most freelancing programmers terribly need it. Please keep doing this. Wish you best of luck in all your efforts
Great video, I learned alot. I'm new to Unity and not as experienced with C# as I am with Java.
Please keep up the amazing work! This is extremely helpful for learning coding in general 👍!
Good stuff. I have heard before that using GetComponent constantly is not good but didn’t have a good understanding of why.
14:20 that comment about the Start function is auto-generated by Visual Studio
You can change what is auto generated, they talk about that
amazing watch. watching this video gave me so much insights in how to write better code and additional information i never knew. Keep it up !
An awesome way to practice is actually code reviewing your own code/legacy code.
There's code I wrote 6 months ago that I review now and just sit there asking, "But why?". And this has stayed true for the 3 years I've worked full-time in unity. BUT, please don't go overboard on this because you can fall into the trap of never progressing forward as you're just refactoring constantly 😂 Fun excersize though!
Have you tried the Presentation Mode in Rider? Menus would become visible. Thank you!
I'll give it a shot in the next one!
Omg this is one of the best videos I've seen! Learnt so much
As far as i know, transform IS cashed in Unity now. And has been for a long time. I think they did it somewhere around Unity 5 or something... You can test it if you're curious - there's zero difference between caching transform in awake and calling transform every update.
Thank you for this video. At my company I'm sort of isolated from other developers. There's nobody to ask "how would you do this code?" or "is this a code smell?". Not even 10 minutes into the video I'm learning about guard clauses and cyclomatic complexity. Guard clauses are something I've been doing for a while now to avoid "arrow code". I always felt a little worried that it was a code smell for the exact reasons you described here and now I know I can safely use it by extracting everything that follows the guard clause into a separate method. I'm sure there's many more things for me to learn in the next 50 minutes. If you Infallible Code or anybody else who reads this have good yt video resources on best practices, please recommend them to me. Thank you!
Hey guys, I'd love to know why you use "var" instead of saying what type the variable is. I was always told abusing var is a bad idea and if working with a team it could be "what is the var? Is it a string? An int?" - you get me. So why use it?
Well lets think about what var does, it removes the name of the return type and replaces it with a shorter cleaner name. could that be confusing? definitely but a common rule with software is the smaller the scope the longer the variable names, it gets to the point where variables have SomeVeryLongNamesInCode to describe their jobs. the thing is though, variables are a 2-stage process. there is the "declaration" and the "assignment/initialization" usually this means you are effectively writing the name/type of the variable twice. Dog myDog = new Dog(), etc. so in one of the above examples: var force = RandomForceVector(); we are writing the word vector... just doing it once, not twice. Var can be confusing if you have no reference visually to what type is. but is usually fine if you do. Lastly personally I don't worry about var because we live in a world of very powerful editors, and you can always hover over a var to see its "real" type, but in most cases you don't need an extra 3-4 words per variable at all times. var massively shortens your code and with vs or rider, you can literally press one keyboard shortcut on it to revert to a fully qualified name. TL;DR the goal of 'clean' code is readability. in my opinion I try to remove as many words from my code that do not add useful information to the person reading it, and if I named my variables well enough they should not need to know what the type is to read or use my code. even if they do.. its not gone, its one editor click away
I really wish you guys took the 5 minutes to implement the tower manager, would have been cool to quickly show the difference in the assembly etc.
This is a great code review!
I'm new to this channel. I love this live review. I learn a lot. I'm subscribing now with a bell HEHEHE
can you please help?
what is better: a lot of scripts having an Update, or just a few "initiators", that loop and update these scripts?
As a newbie this was so good and really eloquently and clearly explained. Thank you both!!
I wasn't aware of ContextMenu, thanks for the tip!
I almost pressed the like button once again to like it once more
25:50 what hotkey author use for capitalize 1st letter?
The only thing left about that GenerateRandomDirection part at around 35:30 that is left to do
is to normalize it and multiply by constant of 10f (it seems)
Then throw it all in a garbage can and just use Random.onUnitSphere * 10f.
*I know probability distribution would be different but it clearly makes sense there.
I don't know why, but no matter how many times I watch these videos, and no matter how many times I think I understand it, I just can't seem to apply it to my own code.
I would like to see beginners tutorials from you guys. We need pro education.
This is the beginner stuff..
This was unbelievably helpful 😊
What a fantastic stream! Thank you.
2:38 Ah, the infamous, nagging, overly intrusive alert sound of windows.
_Virtually always_ heard multiple times in a row.
Hearing it echo in the background never fails to raise my blood pressure.
I have to say I hate underscore. It takes two hands to write code if I use underscore and it doesn't look so clean. I use them only for properties to tell myself I shouldn't use underscored var.
Also I never write private because its very easy to identify myself what is private and what not if modifier is missing.
Really awesome tips guys! Thank you 👍
Thanks for watching!
You guys are really good at explaining things, wow.
This is gold dudes cheers!
Learning a lot from this, thank you. Re. the guard clause, I avoid them generally in C# and C/C++ for the reason you mention - they make code underneath harder to maintain because you de-localise flow control and have to mentally track it as a dependency. I think wrapping the code below up in a new function is messier and less efficient than just using the super-clear if slightly indented if tree.
Another Question!
You guys are awesome! hahaha Thanks for that content!
An Empty Update() Function on a Monobehavior is "running"?
Like "Update() { };"
Is a good practice to remove empty functions like that?
Yes, always remove empty Update functions if you do not use them.
I'm not sure I also want this answer. I would never do this but as far as I know, there are no benefits for performance because the update is a method from the monobehavior and its directly called from the parent :) Hope someone will answer so I get the notification :)
@@mehmedcavas3069 Back then I worked on something that involves on instantiating grid tiles on a level, like about 50x30 grids, with the grid prefab having a MonoBehaviour script in it. At first after everything's done I got pretty low FPS but it turns out that leaving the Update() method even though I'm not using it just because I don't know better is the cause, removing it makes the game buttery smooth again.
In summary, remove Update() method indeed when not using it, hope this answers it!
Are you sure?
Every class which ingerits monobehaviour goes through the unity life cycle right?
@@Aditya-us1qn unity creates a list of all objects that has update, lateUpdate or fixedUpdate...and then runs through it every frame...meaning even if its empty, it's still added to the list...
So yea...if u don't put it in the script, it won't be added to the list...
More of this please :-) Who is the other guy by the way?
Oh wow. I've been coding in Unity for about 4-3 years. I just learned about "foo = foo ?? bar;" Is it just "foo = foo ? bar : foo2;"? But without requiring an "else" after the colon?
as of unity 2020.2 camera.main is super efficient, and doesnt use findobjectwith tag actually😁
This was beautiful, thank you!
Great video👍
Thank you for watching!
As a person that writes a lot of code. I would of told this person, extensions, the transformation and rigidbody will be used a lot like that. An extension makes sense.
This is so valuable, thanks!
25:20 what is interpolating?
The transform's position and rotation
glad am watching those , thanks a lot !
Though I understand how crazy computing power is, they are unknowingly making a case for their insistence on following good practices. They keep finding minor optimization after minor optimization. Many small inefficiencies can start to become large.
this was pretty cool, lots of practical insights
It was really helpful. Thank you!
Using ?? with components is problematic (probably ok with the camera though) since it is not aware of the null check override Unity is using. Null coalescing operators cannot be overloaded, so Unity cannot change their behavior. They only do a pure null check, they do not check if the object has been destroyed. That is still an ongoing discussion on the forums and a current weakness in the Unity API.
Certainly lots of good practices. Just two questions:
a) Why make a simple call to Random a whole new method? Is it because the method could be expanded upon later? I didn't quite understand it if you said it in the video.
b)
Though I too am a user of WaitForSeconds, I find it has one major flaw; it creates an instance of a class that needs to be garbage collected. Furthermore, an input of 0 would still yield, which would mean a loss of at least one frame. Wouldn't an ultimately better solution be something like this (not checking for negative time for simplicitiy)
public IEnumerator WaitForSeconds(float seconds){
float time = 0;
while (time 0) WaitForSeconds(seconds); if you don't want a "missing frame".
EDIT: Unless a yield return null also continuously allocates memory.. Wow, I really need to profile this one day...
Anyway, I always found coroutines very intriguing and a fun way to build time-based systems in unity. I've even built an animation system based on them, since the primary function of waiting is a really powerful tool.
Hey, About using Debug.Logs
When you leave a lot of Debug.Logs in the code, and then build and run the program outside the Unity Editor, we'll have memory usage problems because of that logs? Or it is heavy only on the editor?
This is heavy in the editor only 😊
When doing development builds however, you can see error logs and hook up the console to the built application. But that's a separate thing.
do more videos like this. i m a beginner so i really need to learn a lot from you guys.
Can we get a link to James' channel/@'s?
The fact that you are so uncomfortable seeing prefixes on functions is a great reason to use them for things like the coroutine here. That discomfort means that if you inadvertently attempt to use it as a regular function you will definitely see the prefix and realize the mistake you've made before compile time.I find that doing stuff like this that I don't like is a great way to catch my own attention later on when I return to a piece of code and don't remember everything that it should be doing.
Do you guys still think Camera.main is awful after last changes?
not only Groovy, for example Python not only public by default, but it's the only possible way. there is no protected or private in Python
You mentioned to refrain from too many if statements in Update. What else are you supposed to do? If you have a jump input for a player, I could put the jump function in Update, then the function still has an if for the jump input. How would I check for a a player's input if not checking for it every frame?
Thanks for this nice knowledges. İt's was helpful
Do u really need to cache the transform? I thought it is already cached after unity 5 came out
Yes, it is still worth doing.
blog.collectivemass.com/2019/06/unity-myth-buster-gameobject-transform-vs-cached-transform/
If you cache it yourself, yes it is slightly faster but you could do that for other unity objects too. You don't have to and that's why unity allows transform. now
One practice I've taken up for differentiating coroutines is to just put an underscore prefix. So the method discussed in 30:00 would be _EnablePhysics(...)
I found that it's pretty easy to understand from seeing it, as functions that start with underscore are very rare otherwise. But the main benefit to me is that it just looks weird if you call it on its own. But you get used to reading something like StartCouroutine(_EnablePhysics(...))
The this keyword isnt required?
Nope! docs.microsoft.com/en-us/dotnet/csharp/language-reference/keywords/this
Nope, it's implied.
I'm a scrub, but I'd learnt that starting a coroutine name with 'i' to indicate an IEnumerator was good practice to distinguish it as a coroutine.
These guys are to smart for me I’m still new
Apparently GetComponent() has a non-generic overload which is slightly faster, it is mentioned here: chaoscultgames.com/2014/03/unity3d-mythbusting-performance/
Here's the link to the talk mentioned during the video from the folks at Playdead: ruclips.net/video/mQ2KTRn4BMI/видео.html
Still a great talk years later.
Helpful advice :)
`this.transform` isn't cached- it's hard-coded to directly return the Transform. This is possible for Transform (but not most other things) because every GameObject strictly has one Transform (or derived RectTransform) from initialization. Most components are optional; Transform is not.
The full story is that a while back, Unity got rid of all the component accessor properties that were expensive, leaving only the ones that are cached and as efficient as any code you could write (which is just `transform` and `gameObject` in the latest version). So if Unity provides an accessor, use it, otherwise make your own cached accessor.
You can write your own test code, that catched transform is a little bit faster than internal.
Happy to know i had some of those things right lmao
Also F# and Kotlin, public by default. IMO its better to specify private
Some things being commented and some not, regardless of if its self evident, is a pet peeve. Not because it doesnt make logical sense, it's entirely a personal for things to match. lol
Personally, i think "this.GetComponent" etc. is more readable than single "GetComponent ", just personall preferention to put whichone MonoBehaviour is calling the method. But good video to practice.
I wanna kiss the dude with the glasses, muah.
🤣🤣🤣 I'm very confused How fast he speaks this man
ew, that guys reverb is off the charts. i don't wanna keep watching this if that's what i need to subject myself to.
This has Brackeys all over it
I miss that guy
Code review... Using Coroutines...
I learned a couple of things, like the log context. But a guard clause in Update is an absolute red flag, 99% of the time Update is abused and this is one of them. The _shouldAssemble boolean should have been removed with the update function altogether, MoveToStartingPosition that was extracted from Update should have been made a coroutine called from StartAssembly.
Also the StopAssembly function that you added don't make much sense , I often see code with classes exposed too much, for exemple putting a public setter in every properties by default.
lol this is so dumb, this is all just micro-optimizations. Such a waste of time. Make games, identify bottlenecks, then optimized.
late to this vid, mostly loved right up to the summary/comment argument xD my stance is you are NOT a good judge of if your code is readable or not. You wrote it, you're automatically biased, you have context you don't even realise you have. "i've written this clean, its obvious what it does" is not helpful to the confused person 2 years later. I'm not saying comments should be everywhere, im just saying "code is where your code goes, it should be obvious what it does" is a mcguffin. Also, the most useful thing summaries do in c# is power intellisense - i often use them just for paramater decoration if nothing else!