"You havent learned how to make games. You've learned how to make games in Unity. And if Unity were to disappear for example, you would no longer know how to make a game at all." This aged perfectly
I mean, you've still learned higher level concepts like level design, fine tuning controls, and other 'game' stuff. Depending on how much of the underlying API the engine exposes, you *might* even learn a bit about programming games. I'm pretty sure the only reason I was able to learn OpenGL as fast as I did was because concepts from it were simplified and streamlined in the engine I used to use, GMS2, so I wasn't nearly as confused as I could have been when I went to learn about vertex attributes and stuff. And in the end, unless you're just doing drag and drop asset go there click click finish build, you're still going to have programming fundamentals, all you'll need to learn is a new API. That being said, I never really gave Unity a shot, so maybe it's different and really mind numbing, but I do watch a lot of Sebastian Lague videos and he seems to be able to program fairly well, despite using Unity for rendering, so I don't know.
"You haven't learned how to make games. You've learned how to make games in Windows." "You haven't learned how to make games. You've learned how to make games in Linux." It's an incredibly unintelligent argument. There are so many layers of abstraction that the engine one doesn't really amount to much compared to the rest of it.
@@Ed17151 Well he's not talking about those, he's talking about game engines. Some algorithms can be mathematically proven to be the best option available, but time will tell.
The other thing that gets forgotten when talking about not reinventing the wheel is that despite the thousands of years ever since the wheel was invented, we still tweak it, changing the materials we used to make it, the proportions and so on. To do those effectively, it's very important to know _how_ the wheel is works and, what do you know, trying to reinvent the wheel can be a very good way to know how and why it works.
Some people here are completely missing the point. Casey never said that you shouldn’t use an engine. He’s simply stating the obvious, which is the lower level you go, the more knowledge you will have, and the more control you have. This is the difference between gameplay scripting and actual graphics engineering.
I've ALWAYS found this behaviour very puzzling. But then, there are a cliche of developers who simply think WAY too highly of themselves (alpha devs), and actively work against others to ensure that they will always stay one step ahead. And then there are those weaker developers (the betas) with an inferiority complex, who encourage others NOT to become better developers. Because they themselves will then fall behind! So just insane human behaviour being played out really. And since a lot of devs aren't known for being emotionally intelligent... Particularly if they work with machines way more than their fellow human beings. Well therein lies the reason why these traits can be more profound in this particular field.
I think it comes from a corporate mindset. In business, time is money so you need to get things done as fast as possible, which means not reinventing the wheel. That makes sense in business, but the best way to learn something, at least in CS, is to implement it from scratch yourself.
The funny thing is that actually, we HAVE been reinventing the wheel for thousands of years. Choose any wheel from any object today and compare it with wheels from millennia in the past - they’re so different. In fact, we don’t just have a single “wheel” today. We have different wheels for cars, different wheels for trucks/lorries, for baby strollers, for airplanes, for bicycles, for motorcycles, and the list goes on and on. Yes, the fundamental physics of wheels involving angular momentum and torque have not changed. But how we employ those physics in order to move different classes of objects have been refined over thousands of years.
I'm writing a game and its engine, from scratch, because I find it fun to learn how everything works in detail. By using engines I learned that one must still learn how that engine works, for each engine, unless you have knowledge how everything works. If you know how all engines work under the hood, then you also will have a lot easier time learning a new engine's interfaces.
+VerySeriousUser Exactly. You become much more efficient using other engines if you know how stuff works, you're no long a newb copying and pasting stuff you see online. If you're writing your own stuff you might be interested in a book called Tricks of the 3D Game Programming gurus by Andre Lamothe. It's one of a kind. You start from scratch and end up with a Quake-like engine.
"Don't reinvent the wheel" is a valid statement from certain business perspective. Valid examples are very easy to come along. When you need to solve problems not related to the core business, and you need a solid solution for it, in terms of performance, security, scalability. You would normally resort to a battle-tested library or framework to do that. The same goes for the sadly normal situation of the business requiring things for yesterday and you don't have time to make a wheel in that case. Just grab the best one you can. And most surely, when talking about security of peoples life or information, don't reinventing the wheel for a serious project is the way to go. But, for everything else, go invent new wheels, reinvent existing ones. Again, and again. Go do it, it IS the way to learn. It also makes you appreciate the complexity of the things we use everyday.
The concept of the wheel is not the same as it was thousands of years ago, and it has indeed been reinvented many times. They are no longer carved from a single mass of solid material. They are no longer comprised of 90-degree 2-plied wooden planks nailed together. They are no longer creaking wooden rings compressed against thick spokes. The modern bicycle wheel operates with fine spokes under tension, no longer relying on its operational support by compression. That stands for a very different way of understanding and designing wheels. Then there's technology involved in rims, inner-tubes and tyres, common to larger vehicles as well. "Reinventing the wheel" is a tired and false overused excuse to deny the value of invention. It is not a contrast against the motive of hand-made games, but rather a strong point to further justify learning to make things from scratch. "Reinventing the wheel" is not only an educational process, but when done right, provides better tools for the purpose at hand and inspires further new innovations.
Pumped hella morale in my brain! I wanna do something about not being able to do anything if Unity disappeared. Perhaps start with that game engine book, hand-made hero, and JB's new programming language.
@@sugrado Jai would be cool - can't wait until it gets out in 2030 or so. But for anyone else interested, I think CS Primer may be a good way to start. I'm certainly very excited about it. It's covering virtually the full stack of software and you can pick and choose the topics you're interested in.
For me is actually depends on yourself, if you are programmer or computer science student better for you to create from the scratch as it will useful for you in the future. But if you is just a hobbyist, indie game developer, or artist just learning the game engine is enough.
Not really. I’m a computer science student and I would rather use something more high level like Monogame or Unity. I like Casey but this type of low level programming is just not enjoyable to me and in most cases it means you won’t be able to make an actual usable product. I’m personally not interested at all in the R&D aspect he talked about, I just want to program GAMES not engines.
This is why he said "depends on yourself". Also, you may get better programmer or have a good grasp of the way engines work, if you tinker on low-level stuff while making games with engines.@@EhKurd
@@EhKurd I'm aware this was 4 years ago, but still wanted to leave a comment. Writing things from scratch doesn't mean it can't turn into a usable product. Sure, it's an investment of time, but the quality of the final product will be much better if you used something that fit the needs of your project. And it's certainly not impossible to do either, even for single developers. And as Casey pointed out, even as an educative project it will be very useful, and make you a much better programmer for the higher level tools.
Even if the wheel existed in game engines, what about bike wheels and car wheels and rollerblade wheels... there isn't a one perfect wheel that fits every single purpose
I just wanna program my own shit without some asshole telling me I did it wrong according to some best practice and framework, or whatever the fuck. I know what I'm doing.
I know it’s a long time later but I wanted to mention something. He is not saying that everyone needs to learn how to do this low level stuff. He is not even saying that people who do know how to do it should be using it over an engine like Unity or even a minimal cross platform library like the STB library. He goes on to show removing the C runtime library which is not something that ever makes sense outside of education. The point is that someone needs to be the one to improve Unity, or make a new one, or code the C runtime library, or anything else. Many times he says that he hasn’t done something for a long time because he uses the same platform layer each time and changes it slightly. Not everyone will want to be one of these low level programmers, but there has to be some. The point of that stream is to act as a way for the next generation of low level programmers to learn from someone who has had experience with it, so that when the next people take over Unity, or Unreal, or whatever, there will be people who know and are interested in doing that and thus don’t start from scratch.
The ideal game engine captures the scope and intent of the game well enough to optimize itself down to the smallest code size, or at least makes such optimization easy.
I would kind of argue that "the wheel" - the concept of how to make games - has been invented. But you're not reinventing the wheel, you're making building a kind of wheel that fits your needs. There are wildly differently wheel for different purposes, steering wheels and wheels for planes are wildly different so it makes sense to make new kinds even though one or many kinds have already been made
Indeed! But alas, in this TikTok Gen Z generation, people would rather not learn a damn thing, and simply get that instant hit of success! Which is of course, completely non-sensical! You can't get something from nothing! And EGGaming has nicely proved my point!
People who complain about reinventing the wheel have no idea what they're talking about. EVERY profession where you create something is learned by copying. Programmers start with Hello World. Painters start with shading spheres. Musicians start by playing very easy songs to get used to the basics. Yet for some reason when you build your own XYZ in the software world suddenly people say it's wasted effort because that already exists. Well okay, I guess we don't need to create any new racing game anymore because we already have racing games, right? Why design carbon wheels for high-end roadbikes when we already have bike wheels for many decades? This is so stupid.
Great analogy. Lately, I feel like the thinking of "this and that already exists" comes from both laziness and people believing these things are complicated enough that it's impossible to learn them efficiently. Reinventing the wheel is actually one of the best ways to get more efficient at building complex systems.
I like scratch 1.4 engine you can learn allot because you can actually look at the code vs unity . Unity is locked . It's still cool to have unity but depends on your goals. Fast vs control
Using a game engine, is almost like having your Maths teacher do your exams for you! You won't learn as much, and you're totally reliant on your teacher to get the best grades you possibly can... In effect, you no longer have direct control!
I disagree. Game engines weren't designed for programmers, they were designed for game designers and artists. You need all of them to make a functioning, decent game.
Your not reinventing the wheel, you can't reinvent something, unless you discovered timetravel and your from the future. You are undertanding wheels at a fundemental level by designing them.
Starting from scratch is not always the right thing, it's good if you want to learn low level stuff. It's not good if your goal is to finish a game because it will take much more time doing it that way
This is true. Different goals require different journeys to achieve them. I feel like there is a lot of animosity between the camp that advocates for and against engine use. Use whatever makes you happy, or helps you achieve what you want to achieve.
@@sagitswag1785 No solution is a one-size fits all. But modern software is really crappy because the knowledge of low-level stuff is so poor. I don't think the problem is that low-level stuff is hard to learn - or that high-level developers are too lazy to even try figuring it out. I think it's that the quality resources on it are getting increasingly hard to find, if you don't already know how it works or have a great network. If just 10% of developers took a one month course on how a computer works, then I bet we'd have incredibly powerful high-level languages, compilers, libraries, engines and broad software in a fairly short amount of time. But not only is modern software built in a way that is incredibly ignorant of the hardware it runs on, it's also become unnecessarily complex and abstract to an enormous degree, with modules and frameworks that go in and out of fashion for arbitrary reasons (because they all suck, but the grass _could_ be greener with the new shiny one, right?). This makes it much harder to work with than it should be, and I think our entire field has most likely been in denial about this - me and my coworkers certainly was for the last 10 years.
If you're making games as a hobby, making them from scratch is going to be just fine. If you're doing it professionally, or you want to release your game to Steam, etc. on a reasonable schedule, you _have_ to use an engine. The asset pipelines and development lifecycles in the industry are going to be oriented around one of the big two (Unreal or Unity), unless you work for a shop that has an in-house engine like Valve or Bethesda. And if you want to get a job in the industry, you will have to demonstrate proficiency with one or both of these middlewares. That's just the way things are in 2018. It's like GIMP vs. Photoshop. GIMP is fine if you're just messing around, but if you're doing digital art or photo manipulation that's going to see print, Photoshop is a must.
Just because the industry cares a lot about Unreal or Unity doesn't mean you need to use them. So what if you release a game to steam thats made from scratch? The industry doesn't control how you make games. You do.
@@bgri9823 Stardew Valley, Minecraft, The Witness. In most cases you'll want to use an engine but saying you have to if you are working professionally is really dumb. Some people and some projects will do fine building from scratch. Its also extra dumb because if you want to work professionally as a developer in the games industry, making your own engine is by far the best way to get relevant experience. No one will hire programmers who've only scripted in engines.
I mean, the concept of a wheel will be the same forever, but so will the concept of a game engine. But modern wheels are better than 5000 year old wheels, and modern game engines are better than the very first ones. It's good to ask why you spent months writing your own game engine instead of just using Unity. Sometimes that is the right thing to do, but you should have a reason.
tbh, I disagree, the wheel has in-fact been to some degree reinvented many many times, the concept might not have changed but the details have so many variations so I don't think reinventing the wheel is actually a bad analogy as we wouldn't have the modern world if we settled for a one size fits all generic wheel
+Iggyinabox I'm not trolling at all, you can't build a race car with horse cart wheels(actually they basically redesign them every race season to gain an edge over their competitors) or land an airplane without pneumatic tires, etc. Yet when someone wants to redesign a piece of garbage software because it sucks everyones all like "don't reinvent the wheel" like it's somehow a bad thing to do.
+DjChronokun Well now that you are saying something completely different than your first statement yes you are right the wheel has been redesigned to handle different things, but it has never been reinvented. His main point here is that there is no wheel in when it comes to game engines, no silver bullet so to speak. I know if I want to move something putting it on wheels allows me to take it across a surface much easier. There is no game engine that can do this for games, nothing that works perfectly across the board. It's not that "Don't reinvent the wheel" is essentially bad advice. It just doesn't apply in case.
+Iggyinabox Right, but even if there was such a game engine it would still need to be redesigned, reimplemented, reoptimized to efficiently handle different requirements. So sure, a game engine isn't a wheel, it's more like an automobile that you might decide to reinvent because it needs tracks and skis to run on snow rather than the wheels it came with, no game engine is simple enough (or could be simple enough) to compare to a wheel in conceptual complexity. So ok, you're right, Casey's right, but what I was trying to get at was that the people who say "don't reinvent the wheel" when you're trying to *redesign* some software would say the same thing if you were trying to redesign the wheel and we'd be stuck in the age of horses and carts because it's a fundamental attitude problem that makes them short sighted and always say "this is good enough" whether it is or not and they're generally demonstrating their ignorance of both software and mechanical wheels when they say it - which is why I didn't think it was a bad analogy because they're demonstrating the same ignorance on both sides.
using a computer with a von Neman architecture isn't doing it your self either. Start by inventing your own basic system architecture to avoid just using someone else's idea of how cpu and memory should work.
The difference is you can starting writing your own engine from scratch right now for free. You can't manufacture your own computer architecture from scratch without a billion dollar factory.
championchap it’s not even a comparison, it’s a wider scope of the basic idea of doing it yourself. Just by using a prebuilt FPGAs can get you some low level parallel stuff a normal CPU can't. Asics aren't that uncommon either even though super powerful ones like a high performance CPU might is. Look at what sony is doing with PS5, as I understand it they are using a custom ssd controller that allows the system to copy directly from the hard drive to the GPU without involving the CPU/PCI bus and they have stream compression integrated in their IO and a RAM cache directly at the storage as well to further increase transfer performance. This is not an extremely complicated thing to do and an an example of what you can do if you go beyond conventional PC system architecture.
The correct question is "why choose to learn ??" This isn´t make any sense, why to learn ????? This Casey is crazy man, let´s give all our brains to google and going back to the job -.- ....
So.., should I just give up on making games if I don't learn from scratch? Because apparently what he's trying to say is that learning and making games from pre-existing engines is a bad thing and shouldn't be done? So should I just give up? Because I really don't think I am able to commit enough time to learn how to "Make From Scratch" but things like pre-existing engines makes it easier for me to learn and keep interested. Someone please tell me what to do.
Hey there, first things first, Casey is an old school programmer. l33t engine programmer for 30+ years. Making things from total scratch is how he did all his things. There was no Unity back then, but times have changed. It totally depends on what your goals are and what motivates you. If you just want to make games and have fun but don't have the time to learn all the stuff that's involved in making an engine, then by all means start using Unity or Unreal or what have you. Make a few games, feel satisfied, get the ball rolling. You will get a much better idea of what you want to do then, whether to keep using an engine or invest time in making your own, also depends on what type of game you're making and what platforms you're shipping on. I personally would be very grumpy and annoyed if I had to deal with Java/Android or Swift/iOS directly so I'm glad that Unity does all the platform stuff for me if I'm releasing on mobile. On the other hand, if your goal/what motivates you is to make 'game tech' or tools or low level stuff that's used to ship a game, then you probably want to learn this low level stuff Casey preaches. If you want a role as an engine programmer in some company, you definitely want to know how things work inside out at the lowest level, doing things from scratch is one of the best ways of doing that. If you get excited and your programmer scenes are tingling and aroused just by thinking of making a game in a super limited space 32kb RAM and a 100 hertz CPU or something, if you enjoy the process of learning how things work, how everything fits together, if you like system engineering/architecture problems, API design, etc all those things are good indicators that you're more of a lower level guy and doing things from scratch is a thing you do for fun. But if you just want to make games in the fastest way possible, you don't have to go through all that. There's absolutely nothing wrong with that. You can do a fantastic job using Unity or Unreal. A lot of the bad rep Unity gets is not because it's a bad engine, but because of the low quality products that get shipped using it by hobbyist and beginners, that will go on to make RUclips tutorials spreading some really bad programming content. Casey just wants people to avoid doing that by providing high quality education so that when you do use Unity, you are a lot more efficient and powerful because now you're armed with the knowledge of how the internals work so you can do a better job. If you're just starting out, I personally would suggest to just start doing it, just do anything using any engine just get your feet wet try a bunch of things get a taste of everything see what you like. Definitely don't give up on making games just cause you don't have time to do things from scratch. Doing things from scratch is just one way of doing it, there are many ways to reach an end goal. There's too many sources and people saying different things, do this, do that, don't do those etc. It can get easily overwhelming. It's like starting a workout program, how many reps, what tempo, what exercises, what times, do you have your shake before/after the workout, etc etc. Million questions at the end of the day you just give up and don't do anything. All you have to do is just write a simple program or find one online, stick to it and just start doing your workouts. Eventually you'll just tweak the program to suit your lifestyle best. I personally started out using engines and frameworks myself. XNA, Processing, Unity, Unreal, etc. But I'm the type of person that wants to understand how everything works and implemented under the hood, and having a black box that does things magically for me bothers me. So I tend to write things from scratch myself as much as possible. I do use tools when doing things offline (e.g. asset preprocessing with Assimp), or only in debug mode (e.g. Dear IMGUI), or super lightweight well written no pain in the ass tools (e.g. Sean Barret's libs) as long as the tool doesn't bite me in the ass or stand in my way. Hope that helps! Cheers Ali
"You havent learned how to make games. You've learned how to make games in Unity. And if Unity were to disappear for example, you would no longer know how to make a game at all."
This aged perfectly
Godot, hehe
I mean, you've still learned higher level concepts like level design, fine tuning controls, and other 'game' stuff. Depending on how much of the underlying API the engine exposes, you *might* even learn a bit about programming games. I'm pretty sure the only reason I was able to learn OpenGL as fast as I did was because concepts from it were simplified and streamlined in the engine I used to use, GMS2, so I wasn't nearly as confused as I could have been when I went to learn about vertex attributes and stuff.
And in the end, unless you're just doing drag and drop asset go there click click finish build, you're still going to have programming fundamentals, all you'll need to learn is a new API.
That being said, I never really gave Unity a shot, so maybe it's different and really mind numbing, but I do watch a lot of Sebastian Lague videos and he seems to be able to program fairly well, despite using Unity for rendering, so I don't know.
"You haven't learned how to make games. You've learned how to make games in Windows."
"You haven't learned how to make games. You've learned how to make games in Linux."
It's an incredibly unintelligent argument. There are so many layers of abstraction that the engine one doesn't really amount to much compared to the rest of it.
"Nothing we have developed is a wheel." Damn, that is an eye-opening statement.
I've been using Unreal to make a game for over a year, and this is a big mood right now TBH.
What about algorithms and data structures. Some of them haven't changed many years.
@@Ed17151 Well he's not talking about those, he's talking about game engines. Some algorithms can be mathematically proven to be the best option available, but time will tell.
The other thing that gets forgotten when talking about not reinventing the wheel is that despite the thousands of years ever since the wheel was invented, we still tweak it, changing the materials we used to make it, the proportions and so on. To do those effectively, it's very important to know _how_ the wheel is works and, what do you know, trying to reinvent the wheel can be a very good way to know how and why it works.
Some people here are completely missing the point. Casey never said that you shouldn’t use an engine. He’s simply stating the obvious, which is the lower level you go, the more knowledge you will have, and the more control you have. This is the difference between gameplay scripting and actual graphics engineering.
The real question is not why we want to reinvent the wheel, but why other programmers are trying to discourage us from doing so?
my best guess is they are so scared of leaving their comfort zone. so when they see someone else doing it, it just confuses them.
I've ALWAYS found this behaviour very puzzling. But then, there are a cliche of developers who simply think WAY too highly of themselves (alpha devs), and actively work against others to ensure that they will always stay one step ahead.
And then there are those weaker developers (the betas) with an inferiority complex, who encourage others NOT to become better developers. Because they themselves will then fall behind!
So just insane human behaviour being played out really. And since a lot of devs aren't known for being emotionally intelligent... Particularly if they work with machines way more than their fellow human beings. Well therein lies the reason why these traits can be more profound in this particular field.
I think it comes from a corporate mindset. In business, time is money so you need to get things done as fast as possible, which means not reinventing the wheel. That makes sense in business, but the best way to learn something, at least in CS, is to implement it from scratch yourself.
The funny thing is that actually, we HAVE been reinventing the wheel for thousands of years. Choose any wheel from any object today and compare it with wheels from millennia in the past - they’re so different. In fact, we don’t just have a single “wheel” today. We have different wheels for cars, different wheels for trucks/lorries, for baby strollers, for airplanes, for bicycles, for motorcycles, and the list goes on and on. Yes, the fundamental physics of wheels involving angular momentum and torque have not changed. But how we employ those physics in order to move different classes of objects have been refined over thousands of years.
I'm writing a game and its engine, from scratch, because I find it fun to learn how everything works in detail. By using engines I learned that one must still learn how that engine works, for each engine, unless you have knowledge how everything works. If you know how all engines work under the hood, then you also will have a lot easier time learning a new engine's interfaces.
+VerySeriousUser Exactly. You become much more efficient using other engines if you know how stuff works, you're no long a newb copying and pasting stuff you see online. If you're writing your own stuff you might be interested in a book called Tricks of the 3D Game Programming gurus by Andre Lamothe. It's one of a kind. You start from scratch and end up with a Quake-like engine.
Wish I had watched this when I started, but oh well! 573 episodes later he's still doing this! I love it.
"Don't reinvent the wheel" is a valid statement from certain business perspective. Valid examples are very easy to come along.
When you need to solve problems not related to the core business, and you need a solid solution for it, in terms of performance, security, scalability. You would normally resort to a battle-tested library or framework to do that. The same goes for the sadly normal situation of the business requiring things for yesterday and you don't have time to make a wheel in that case. Just grab the best one you can.
And most surely, when talking about security of peoples life or information, don't reinventing the wheel for a serious project is the way to go. But, for everything else, go invent new wheels, reinvent existing ones. Again, and again. Go do it, it IS the way to learn. It also makes you appreciate the complexity of the things we use everyday.
The concept of the wheel is not the same as it was thousands of years ago, and it has indeed been reinvented many times. They are no longer carved from a single mass of solid material. They are no longer comprised of 90-degree 2-plied wooden planks nailed together. They are no longer creaking wooden rings compressed against thick spokes. The modern bicycle wheel operates with fine spokes under tension, no longer relying on its operational support by compression. That stands for a very different way of understanding and designing wheels. Then there's technology involved in rims, inner-tubes and tyres, common to larger vehicles as well.
"Reinventing the wheel" is a tired and false overused excuse to deny the value of invention. It is not a contrast against the motive of hand-made games, but rather a strong point to further justify learning to make things from scratch. "Reinventing the wheel" is not only an educational process, but when done right, provides better tools for the purpose at hand and inspires further new innovations.
Pumped hella morale in my brain! I wanna do something about not being able to do anything if Unity disappeared. Perhaps start with that game engine book, hand-made hero, and JB's new programming language.
Major O Why did you think I went off? ;)
So you wouldn't talk to me :P
Jk xD please do keep sharing anything useful regarding that matter on here.
So, how is it going?
@@kcarlsson89 He's probably still waiting for Jai.
@@sugrado Jai would be cool - can't wait until it gets out in 2030 or so.
But for anyone else interested, I think CS Primer may be a good way to start. I'm certainly very excited about it. It's covering virtually the full stack of software and you can pick and choose the topics you're interested in.
0:45 Talk about a prophecy, haha.
I am a professional game developer and I totally agree with Casey.
Thank you. The answer that I want everyone I know to know.
For me is actually depends on yourself, if you are programmer or computer science student better for you to create from the scratch as it will useful for you in the future. But if you is just a hobbyist, indie game developer, or artist just learning the game engine is enough.
Not really. I’m a computer science student and I would rather use something more high level like Monogame or Unity. I like Casey but this type of low level programming is just not enjoyable to me and in most cases it means you won’t be able to make an actual usable product. I’m personally not interested at all in the R&D aspect he talked about, I just want to program GAMES not engines.
This is why he said "depends on yourself". Also, you may get better programmer or have a good grasp of the way engines work, if you tinker on low-level stuff while making games with engines.@@EhKurd
@@EhKurd I'm aware this was 4 years ago, but still wanted to leave a comment. Writing things from scratch doesn't mean it can't turn into a usable product. Sure, it's an investment of time, but the quality of the final product will be much better if you used something that fit the needs of your project. And it's certainly not impossible to do either, even for single developers. And as Casey pointed out, even as an educative project it will be very useful, and make you a much better programmer for the higher level tools.
I love the cursed thumbnail, Casey looks like he in the middle of a nuclear blast
You probably did mean "...static into dynamic...". Great take btw.
Even if the wheel existed in game engines, what about bike wheels and car wheels and rollerblade wheels... there isn't a one perfect wheel that fits every single purpose
I just wanna program my own shit without some asshole telling me I did it wrong according to some best practice and framework, or whatever the fuck. I know what I'm doing.
I know it’s a long time later but I wanted to mention something. He is not saying that everyone needs to learn how to do this low level stuff. He is not even saying that people who do know how to do it should be using it over an engine like Unity or even a minimal cross platform library like the STB library. He goes on to show removing the C runtime library which is not something that ever makes sense outside of education. The point is that someone needs to be the one to improve Unity, or make a new one, or code the C runtime library, or anything else. Many times he says that he hasn’t done something for a long time because he uses the same platform layer each time and changes it slightly. Not everyone will want to be one of these low level programmers, but there has to be some. The point of that stream is to act as a way for the next generation of low level programmers to learn from someone who has had experience with it, so that when the next people take over Unity, or Unreal, or whatever, there will be people who know and are interested in doing that and thus don’t start from scratch.
The ideal game engine captures the scope and intent of the game well enough to optimize itself down to the smallest code size, or at least makes such optimization easy.
I would kind of argue that "the wheel" - the concept of how to make games - has been invented. But you're not reinventing the wheel, you're making building a kind of wheel that fits your needs. There are wildly differently wheel for different purposes, steering wheels and wheels for planes are wildly different so it makes sense to make new kinds even though one or many kinds have already been made
Totally on your side ! people should learn to program, not use engines.
fuck you kiddo. Just makes games who cares about some rendering or fysics bs lol
@@eggaming8292 I hope you're kidding
@@igorswies5913 If you want to make games for most people its better to use engine.
Indeed! But alas, in this TikTok Gen Z generation, people would rather not learn a damn thing, and simply get that instant hit of success! Which is of course, completely non-sensical! You can't get something from nothing!
And EGGaming has nicely proved my point!
@@sdwone Boomer moment
People who complain about reinventing the wheel have no idea what they're talking about.
EVERY profession where you create something is learned by copying. Programmers start with Hello World. Painters start with shading spheres. Musicians start by playing very easy songs to get used to the basics. Yet for some reason when you build your own XYZ in the software world suddenly people say it's wasted effort because that already exists. Well okay, I guess we don't need to create any new racing game anymore because we already have racing games, right? Why design carbon wheels for high-end roadbikes when we already have bike wheels for many decades?
This is so stupid.
Great analogy. Lately, I feel like the thinking of "this and that already exists" comes from both laziness and people believing these things are complicated enough that it's impossible to learn them efficiently. Reinventing the wheel is actually one of the best ways to get more efficient at building complex systems.
I like scratch 1.4 engine you can learn allot because you can actually look at the code vs unity . Unity is locked . It's still cool to have unity but depends on your goals. Fast vs control
Well said. I learned so much more from coding games from scratch
the prophecy is true
Смотрю этот видос и вспоминаю windertona:
Не ожидал тут подписчика вина увидеть)
Using a game engine, is almost like having your Maths teacher do your exams for you! You won't learn as much, and you're totally reliant on your teacher to get the best grades you possibly can... In effect, you no longer have direct control!
I disagree. Game engines weren't designed for programmers, they were designed for game designers and artists. You need all of them to make a functioning, decent game.
As long as you know how to control the teacher all is fine :)
Your not reinventing the wheel, you can't reinvent something, unless you discovered timetravel and your from the future. You are undertanding wheels at a fundemental level by designing them.
Starting from scratch is not always the right thing, it's good if you want to learn low level stuff. It's not good if your goal is to finish a game because it will take much more time doing it that way
This is true. Different goals require different journeys to achieve them. I feel like there is a lot of animosity between the camp that advocates for and against engine use. Use whatever makes you happy, or helps you achieve what you want to achieve.
@@sagitswag1785 No solution is a one-size fits all. But modern software is really crappy because the knowledge of low-level stuff is so poor. I don't think the problem is that low-level stuff is hard to learn - or that high-level developers are too lazy to even try figuring it out. I think it's that the quality resources on it are getting increasingly hard to find, if you don't already know how it works or have a great network. If just 10% of developers took a one month course on how a computer works, then I bet we'd have incredibly powerful high-level languages, compilers, libraries, engines and broad software in a fairly short amount of time.
But not only is modern software built in a way that is incredibly ignorant of the hardware it runs on, it's also become unnecessarily complex and abstract to an enormous degree, with modules and frameworks that go in and out of fashion for arbitrary reasons (because they all suck, but the grass _could_ be greener with the new shiny one, right?). This makes it much harder to work with than it should be, and I think our entire field has most likely been in denial about this - me and my coworkers certainly was for the last 10 years.
If you're making games as a hobby, making them from scratch is going to be just fine. If you're doing it professionally, or you want to release your game to Steam, etc. on a reasonable schedule, you _have_ to use an engine. The asset pipelines and development lifecycles in the industry are going to be oriented around one of the big two (Unreal or Unity), unless you work for a shop that has an in-house engine like Valve or Bethesda. And if you want to get a job in the industry, you will have to demonstrate proficiency with one or both of these middlewares. That's just the way things are in 2018. It's like GIMP vs. Photoshop. GIMP is fine if you're just messing around, but if you're doing digital art or photo manipulation that's going to see print, Photoshop is a must.
Just because the industry cares a lot about Unreal or Unity doesn't mean you need to use them. So what if you release a game to steam thats made from scratch? The industry doesn't control how you make games. You do.
what about, lets say, Path of Exile? fully competitive game on Steam based on a custom engine.
@@bgri9823 Stardew Valley, Minecraft, The Witness. In most cases you'll want to use an engine but saying you have to if you are working professionally is really dumb. Some people and some projects will do fine building from scratch.
Its also extra dumb because if you want to work professionally as a developer in the games industry, making your own engine is by far the best way to get relevant experience. No one will hire programmers who've only scripted in engines.
@@La0bouchere It depends on the position. I want to do game design stuff and code my designs so I don't necessarily need to be John Carmack.
If you're indie, you don't *have* to do anything.
I mean, the concept of a wheel will be the same forever, but so will the concept of a game engine. But modern wheels are better than 5000 year old wheels, and modern game engines are better than the very first ones.
It's good to ask why you spent months writing your own game engine instead of just using Unity. Sometimes that is the right thing to do, but you should have a reason.
tbh, I disagree, the wheel has in-fact been to some degree reinvented many many times, the concept might not have changed but the details have so many variations so I don't think reinventing the wheel is actually a bad analogy as we wouldn't have the modern world if we settled for a one size fits all generic wheel
+DjChronokun This made me LOL hard, but then I realized you were just trolling, or at lest I am hoping you are.
+Iggyinabox I'm not trolling at all, you can't build a race car with horse cart wheels(actually they basically redesign them every race season to gain an edge over their competitors) or land an airplane without pneumatic tires, etc. Yet when someone wants to redesign a piece of garbage software because it sucks everyones all like "don't reinvent the wheel" like it's somehow a bad thing to do.
+DjChronokun Well now that you are saying something completely different than your first statement yes you are right the wheel has been redesigned to handle different things, but it has never been reinvented.
His main point here is that there is no wheel in when it comes to game engines, no silver bullet so to speak. I know if I want to move something putting it on wheels allows me to take it across a surface much easier. There is no game engine that can do this for games, nothing that works perfectly across the board. It's not that "Don't reinvent the wheel" is essentially bad advice. It just doesn't apply in case.
+Iggyinabox Right, but even if there was such a game engine it would still need to be redesigned, reimplemented, reoptimized to efficiently handle different requirements. So sure, a game engine isn't a wheel, it's more like an automobile that you might decide to reinvent because it needs tracks and skis to run on snow rather than the wheels it came with, no game engine is simple enough (or could be simple enough) to compare to a wheel in conceptual complexity. So ok, you're right, Casey's right, but what I was trying to get at was that the people who say "don't reinvent the wheel" when you're trying to *redesign* some software would say the same thing if you were trying to redesign the wheel and we'd be stuck in the age of horses and carts because it's a fundamental attitude problem that makes them short sighted and always say "this is good enough" whether it is or not and they're generally demonstrating their ignorance of both software and mechanical wheels when they say it - which is why I didn't think it was a bad analogy because they're demonstrating the same ignorance on both sides.
Henry Ford, why are you reinventing the wagon! lol
LOL.
using a computer with a von Neman architecture isn't doing it your self either. Start by inventing your own basic system architecture to avoid just using someone else's idea of how cpu and memory should work.
The difference is you can starting writing your own engine from scratch right now for free. You can't manufacture your own computer architecture from scratch without a billion dollar factory.
championchap it’s not even a comparison, it’s a wider scope of the basic idea of doing it yourself. Just by using a prebuilt FPGAs can get you some low level parallel stuff a normal CPU can't. Asics aren't that uncommon either even though super powerful ones like a high performance CPU might is.
Look at what sony is doing with PS5, as I understand it they are using a custom ssd controller that allows the system to copy directly from the hard drive to the GPU without involving the CPU/PCI bus and they have stream compression integrated in their IO and a RAM cache directly at the storage as well to further increase transfer performance. This is not an extremely complicated thing to do and an an example of what you can do if you go beyond conventional PC system architecture.
Ok now we're talking. EUV lithography from scratch let's go!!
The correct question is "why choose to learn ??" This isn´t make any sense, why to learn ????? This Casey is crazy man, let´s give all our brains to google and going back to the job -.- ....
That was a joke.... Right?
Bro ... just unplug your computer. It's obvious why you are asking such a dumb question.
So.., should I just give up on making games if I don't learn from scratch? Because apparently what he's trying to say is that learning and making games from pre-existing engines is a bad thing and shouldn't be done? So should I just give up? Because I really don't think I am able to commit enough time to learn how to "Make From Scratch" but things like pre-existing engines makes it easier for me to learn and keep interested. Someone please tell me what to do.
Hey there, first things first, Casey is an old school programmer. l33t engine programmer for 30+ years. Making things from total scratch is how he did all his things. There was no Unity back then, but times have changed. It totally depends on what your goals are and what motivates you. If you just want to make games and have fun but don't have the time to learn all the stuff that's involved in making an engine, then by all means start using Unity or Unreal or what have you. Make a few games, feel satisfied, get the ball rolling. You will get a much better idea of what you want to do then, whether to keep using an engine or invest time in making your own, also depends on what type of game you're making and what platforms you're shipping on. I personally would be very grumpy and annoyed if I had to deal with Java/Android or Swift/iOS directly so I'm glad that Unity does all the platform stuff for me if I'm releasing on mobile.
On the other hand, if your goal/what motivates you is to make 'game tech' or tools or low level stuff that's used to ship a game, then you probably want to learn this low level stuff Casey preaches. If you want a role as an engine programmer in some company, you definitely want to know how things work inside out at the lowest level, doing things from scratch is one of the best ways of doing that. If you get excited and your programmer scenes are tingling and aroused just by thinking of making a game in a super limited space 32kb RAM and a 100 hertz CPU or something, if you enjoy the process of learning how things work, how everything fits together, if you like system engineering/architecture problems, API design, etc all those things are good indicators that you're more of a lower level guy and doing things from scratch is a thing you do for fun. But if you just want to make games in the fastest way possible, you don't have to go through all that. There's absolutely nothing wrong with that. You can do a fantastic job using Unity or Unreal. A lot of the bad rep Unity gets is not because it's a bad engine, but because of the low quality products that get shipped using it by hobbyist and beginners, that will go on to make RUclips tutorials spreading some really bad programming content. Casey just wants people to avoid doing that by providing high quality education so that when you do use Unity, you are a lot more efficient and powerful because now you're armed with the knowledge of how the internals work so you can do a better job.
If you're just starting out, I personally would suggest to just start doing it, just do anything using any engine just get your feet wet try a bunch of things get a taste of everything see what you like. Definitely don't give up on making games just cause you don't have time to do things from scratch. Doing things from scratch is just one way of doing it, there are many ways to reach an end goal. There's too many sources and people saying different things, do this, do that, don't do those etc. It can get easily overwhelming. It's like starting a workout program, how many reps, what tempo, what exercises, what times, do you have your shake before/after the workout, etc etc. Million questions at the end of the day you just give up and don't do anything. All you have to do is just write a simple program or find one online, stick to it and just start doing your workouts. Eventually you'll just tweak the program to suit your lifestyle best.
I personally started out using engines and frameworks myself. XNA, Processing, Unity, Unreal, etc. But I'm the type of person that wants to understand how everything works and implemented under the hood, and having a black box that does things magically for me bothers me. So I tend to write things from scratch myself as much as possible. I do use tools when doing things offline (e.g. asset preprocessing with Assimp), or only in debug mode (e.g. Dear IMGUI), or super lightweight well written no pain in the ass tools (e.g. Sean Barret's libs) as long as the tool doesn't bite me in the ass or stand in my way.
Hope that helps!
Cheers
Ali
Bru he literally said that its fine to use engines and that you should do that if you just want to make a finished game
The only decent advice you'll ever get is don't listen to anyone's advice. Everyone is full of crap including the gas bag in this video.
audio quality sucks