3 ms·

I've been doing a lot of programming in Unreal lately, and I agree: It is hard. The docs are shockingly terrible, and 99% of the time I just end up reading the

I've been doing a lot of programming in Unreal lately, and I agree: It is hard. The docs are shockingly terrible, and 99% of the time I just end up reading the source code of the engine for documentation. It is also re-igniting my passion for software development and learning. So far I have: - Learned a LOT of linear algebra and vector math through "learning by doing" with 3d games - Learned C++ - Made a bunch of dumb little games during "game jams" - Built a particle system using Niagara, made it into a component, and attached it to my player character for a movement indicator - Made a valiant attempt at developing a procedural animated suspension system for a little car. If you have a path for learning how to do this procedural animation stuff, please @ me!! - Learned how to configure and implement the pathfinding and AI systems for Unreal - Realized I need to do a bit of a deep dive into shader development, and picked up a book as an introduction to shaders and have been watching tutorials. Fun stuff!! - Had to think a LOT about architecture for my game. - Learned basic blender modeling and rigging, rigged a model I have spent so much time and... I still do not have a functional game. But that is OK. If I was trying to do something simpler, I would have something. But I'm not, i want to learn and build something cool. Overall I see my gamedev as a "punk rock" kind of thing: I don't do it for the money, I do it because I want to and it fills the time with intellectually stimulating work.

valyagolev 5y ago
Not sure if that's what you're looking for, but check out Pixar's collection on Khan Academy. It helped me a lot even though it's not extremely deep https://www.khanacademy.org/computing/pixar https://www.khanacademy.org/computing/pixar
The_rationalist 5y ago
spywaregorilla 5y ago
Try blueprints. Not because they're easier, but because they expose the engine API and make it so much more discoverable. IMO its a much better way to learn the engine. When I write c++ for unreal or c# for unity, I find I'm much more likely to end up time trying to implement a concept that the engine already handles, in a far superior approach than what I'm doing. Procedural animation isn't quite there yet, but control rig is slowly improving.
- Agentlien 5y ago
  On the other hand if you make blueprints without really knowing the engine you'll quickly end up with a mess of cobbled together poor solutions which sort of work.
  - kwertyoowiyop 5y ago
    You can do that in C++ very easily too.
    - Agentlien 5y ago
      Absolutely, but I've seen it happen a lot more in visual scripting languages than in C++. Mainly because it lowers the entry requirements to people who don't understand what they're doing.
      - kwertyoowiyop 5y ago
        Agreed 100%. Then you have to bring in a Sr. Engineer to reorganize and rearchitect all the “not code.”
  - spywaregorilla 5y ago
    That's fine. By the end you'll know the engine and be able to make it better.
    - Agentlien 5y ago
      That sounds fine for hobby projects but I used to work in a AAA studio where much of the game logic was made this way and a lot of engineering time went into identifying and fixing issues (both functional and performance) caused by designers or TAs working this way without asking help when they needed it.
- honkycat 5y ago
  My workflow tends to be: - Build it in blueprints for rapid iteration - Port it to C++ for refactoring and long-term code quality I was bullish on the node based programming, but over time I've liked it less and less. I've been a software engineer for 10+ years so I am very comfortable in the IDE and coding workflow. But yeah, discoverability in blueprints is out of this world. And the C++ classes use the same functions as the blueprint classes. Once you figure out how to find the C++ the blueprint is calling into, it becomes very productive.
d23 5y ago
> - Made a valiant attempt at developing a procedural animated suspension system for a little car. If you have a path for learning how to do this procedural animation stuff, please @ me!! > - Realized I need to do a bit of a deep dive into shader development, and picked up a book as an introduction to shaders and have been watching tutorials. Fun stuff!! Can you point me in the direction of resources for both of these? Or give me your contact information? Would love to chat!
- honkycat 5y ago
  For procedural animation: I have yet to find the path to learning this skill. It is very opaque, and I think you just need to attack it with your brain and learn it. For shader and graphics programming: I liked this tutorial series as a "first steps" by Freya Holmér: https://www.youtube.com/watch?v=kfM-yu0iQBk https://www.youtube.com/watch?v=kfM-yu0iQBk Here is another list: https://www.alanzucconi.com/2018/01/03/learning-shaders/ https://www.alanzucconi.com/2018/01/03/learning-shaders/ The Book of Shaders is a popular suggestion, but it appears to be abandoned: https://thebookofshaders.com/ https://thebookofshaders.com/ For general Unreal? Tom Looman's $300 Unreal C++ class, available on his website. It is pricey but really excellent. I also discovered this AWESOME list of learning paths for all gamedev concepts: https://github.com/miloyip/game-programmer https://github.com/miloyip/game-programmer
smrtinsert 5y ago
I spent a good number of years doing similar. Back in day with Blender, Microsoft XNA. I did a bunch of work on shaders, modeling concept art, linear algebra, C# and then had reached demo stage where my main vehicles were interacting in a physics engine I had essentially coding using 3d models with basic art I put together. At that point I realized I needed marketing, business/legal framework, and more importantly distribution. I didn't even own an XBox, why am I targeting it? I decided to start over with LibGDX with primary platform being Android which was exploding - and the rest is history. Off and on 3-4 years down the drain I suppose. My advice to anyone considering this is do a minimal indie game compo. Set the bar as low as possible so that you get a feel for all of these important pieces of actually producing a game from start to finish. I still think my concept is sound, and I have 1 other idea I'd love to work on, but there's no time now that I'm old and with a family. My interests have pivoted back to music which is more quickly gratifying and easier to share. It doesn't help that the gaming community is a toxic user base in my experience. Finally, what does it even mean to make a game these days. It seems all people care about are in game items, markets to trade them etc. Nothing fun about that for me at all.

valyagolev 5y ago

Not sure if that's what you're looking for, but check out Pixar's collection on Khan Academy. It helped me a lot even though it's not extremely deep https://www.khanacademy.org/computing/pixar https://www.khanacademy.org/computing/pixar

The_rationalist 5y ago

spywaregorilla 5y ago

Try blueprints. Not because they're easier, but because they expose the engine API and make it so much more discoverable. IMO its a much better way to learn the engine. When I write c++ for unreal or c# for unity, I find I'm much more likely to end up time trying to implement a concept that the engine already handles, in a far superior approach than what I'm doing. Procedural animation isn't quite there yet, but control rig is slowly improving.

Agentlien 5y ago
On the other hand if you make blueprints without really knowing the engine you'll quickly end up with a mess of cobbled together poor solutions which sort of work.
- kwertyoowiyop 5y ago
  You can do that in C++ very easily too.
  - Agentlien 5y ago
    Absolutely, but I've seen it happen a lot more in visual scripting languages than in C++. Mainly because it lowers the entry requirements to people who don't understand what they're doing.
    - kwertyoowiyop 5y ago
      Agreed 100%. Then you have to bring in a Sr. Engineer to reorganize and rearchitect all the “not code.”
- spywaregorilla 5y ago
  That's fine. By the end you'll know the engine and be able to make it better.
  - Agentlien 5y ago
    That sounds fine for hobby projects but I used to work in a AAA studio where much of the game logic was made this way and a lot of engineering time went into identifying and fixing issues (both functional and performance) caused by designers or TAs working this way without asking help when they needed it.
honkycat 5y ago
My workflow tends to be: - Build it in blueprints for rapid iteration - Port it to C++ for refactoring and long-term code quality I was bullish on the node based programming, but over time I've liked it less and less. I've been a software engineer for 10+ years so I am very comfortable in the IDE and coding workflow. But yeah, discoverability in blueprints is out of this world. And the C++ classes use the same functions as the blueprint classes. Once you figure out how to find the C++ the blueprint is calling into, it becomes very productive.

d23 5y ago

> - Made a valiant attempt at developing a procedural animated suspension system for a little car. If you have a path for learning how to do this procedural animation stuff, please @ me!! > - Realized I need to do a bit of a deep dive into shader development, and picked up a book as an introduction to shaders and have been watching tutorials. Fun stuff!! Can you point me in the direction of resources for both of these? Or give me your contact information? Would love to chat!

honkycat 5y ago
For procedural animation: I have yet to find the path to learning this skill. It is very opaque, and I think you just need to attack it with your brain and learn it. For shader and graphics programming: I liked this tutorial series as a "first steps" by Freya Holmér: https://www.youtube.com/watch?v=kfM-yu0iQBk https://www.youtube.com/watch?v=kfM-yu0iQBk Here is another list: https://www.alanzucconi.com/2018/01/03/learning-shaders/ https://www.alanzucconi.com/2018/01/03/learning-shaders/ The Book of Shaders is a popular suggestion, but it appears to be abandoned: https://thebookofshaders.com/ https://thebookofshaders.com/ For general Unreal? Tom Looman's $300 Unreal C++ class, available on his website. It is pricey but really excellent. I also discovered this AWESOME list of learning paths for all gamedev concepts: https://github.com/miloyip/game-programmer https://github.com/miloyip/game-programmer

smrtinsert 5y ago

I spent a good number of years doing similar. Back in day with Blender, Microsoft XNA. I did a bunch of work on shaders, modeling concept art, linear algebra, C# and then had reached demo stage where my main vehicles were interacting in a physics engine I had essentially coding using 3d models with basic art I put together. At that point I realized I needed marketing, business/legal framework, and more importantly distribution. I didn't even own an XBox, why am I targeting it? I decided to start over with LibGDX with primary platform being Android which was exploding - and the rest is history. Off and on 3-4 years down the drain I suppose. My advice to anyone considering this is do a minimal indie game compo. Set the bar as low as possible so that you get a feel for all of these important pieces of actually producing a game from start to finish. I still think my concept is sound, and I have 1 other idea I'd love to work on, but there's no time now that I'm old and with a family. My interests have pivoted back to music which is more quickly gratifying and easier to share. It doesn't help that the gaming community is a toxic user base in my experience. Finally, what does it even mean to make a game these days. It seems all people care about are in game items, markets to trade them etc. Nothing fun about that for me at all.