Hi, I've always been fascinated by the idea of writing game engines..and I wanna see myself in the future as an engine programmer. I'm currently 19, in my second year of undegrads(CS), and with 3.5 years of C++ programming experience, on my way on reaching the intermediate level. After going through quite a few threads related go game engines, I've decided to first learn directX, create a small FPS using that for graphics, then bring all my code together and using it, frame my own engine. After that, learn python for scripting, and learn a suitable physics library (Looking at physx and bullet) for physics in-game.

The thing is, I'm not really into making (only)games, I'm more into the low level subsystems that come together to make that game.. so what should I build for my portfolio which would show off my engine programming skills? Are tech-demos okay? Or should I do something else? If yes, what should it be?

Thanks!!

I think a good engine is something that allow someone not that skilled in programming to make a good game/software really easily.

And also that make it easy to a non-programmer (artist) to integrate their content into the software.

I am now also striving to do my own engine, but in HTML5 / WebGL / JS. Why? Because C/C++ is pratically taken by Unity, Unreal and CryEngine. Also you forcefully need to use the best algorithms and patterns to speedup stuff, so it is a good challenge. For what I experienced now, I think what is obligatory to offer a good starting point is:

- Wrap Graphics calls so the user do not need to manage them (he do not need to manage buffer ids, pointers, shader compilation,...)

- So create "Mesh" "Texture2D" "Shader" "Material" and classes that handle this stuff for him.

- The ideal pattern would be the user creating a Camera and the object and it would be already in the screen.

- Offer Initialization, Update, PhysicsUpdate (constant delta time), Destroy callbacks with a set of usefull data at the user hand;

- Handle creation and destruction of stuff so the user can avoid pointer handling.

- Useful data == Time.deltaTime, Input.GetKey(...), Time.elapsed, Time.physicsDeltaTime, ...

- Mesh loading from popular formats (Collada == XML wich is more easy) or FBX wich is more complex but can embed mesh and texture data.

- Animations (Skining / Bones) curve based interpolation of data.

- Physics, you can plug PhysX using C/C++ or C#. I will not even tell you to do one because it is an engine on its own.

- But at least should have rigidbodies, collisions with/without real interaction, raycasting, ...

- Sound (we tend to forget that)

- Post Processing effects (put your render in a buffer and render again in a plane to apply Image Effects like blur)

So never forget that an engine isn't just about rendering. It can cover from low-level to hi-level stuff.

You should cover the Core (initialization, update, render, destroy), Rendering, Sound, Physics, Resouces (Mesh, Texture, ...) sections very well to make it easy to use.

I am too working on a game engine.I think it's great to have one in your portofolio,especially if you want to apply for a programmer job.

I think a good engine is something that allow someone not that skilled in programming to make a good game/software really easily.

And also that make it easy to a non-programmer (artist) to integrate their content into the software.

I am now also striving to do my own engine, but in HTML5 / WebGL / JS. Why? Because C/C++ is pratically taken by Unity, Unreal and CryEngine. Also you forcefully need to use the best algorithms and patterns to speedup stuff, so it is a good challenge. For what I experienced now, I think what is obligatory to offer a good starting point is:

- Wrap Graphics calls so the user do not need to manage them (he do not need to manage buffer ids, pointers, shader compilation,...)

- So create "Mesh" "Texture2D" "Shader" "Material" and classes that handle this stuff for him.

- The ideal pattern would be the user creating a Camera and the object and it would be already in the screen.

- Offer Initialization, Update, PhysicsUpdate (constant delta time), Destroy callbacks with a set of usefull data at the user hand;

- Handle creation and destruction of stuff so the user can avoid pointer handling.

- Useful data == Time.deltaTime, Input.GetKey(...), Time.elapsed, Time.physicsDeltaTime, ...

- Mesh loading from popular formats (Collada == XML wich is more easy) or FBX wich is more complex but can embed mesh and texture data.

- Animations (Skining / Bones) curve based interpolation of data.

- Physics, you can plug PhysX using C/C++ or C#. I will not even tell you to do one because it is an engine on its own.

- But at least should have rigidbodies, collisions with/without real interaction, raycasting, ...

- Sound (we tend to forget that)

- Post Processing effects (put your render in a buffer and render again in a plane to apply Image Effects like blur)

So never forget that an engine isn't just about rendering. It can cover from low-level to hi-level stuff.

You should cover the Core (initialization, update, render, destroy), Rendering, Sound, Physics, Resouces (Mesh, Texture, ...) sections very well to make it easy to use.

Thanks for the suggestions! they'll surely help me when on my way to develop the engine.

But,

Well, the point is, I'm not trying to make the next AAA game engine which would beat Unity or UDK, Since it seems i'll be doing this alone, I won't be able to do that anyways :P
The reason I chose c++ is because that's the language I'm most comfortable with ATM, and I'm ready to learn anything new related to the language as and when needed.
Besides, if I become an engine programmer, I'll get the opportunity to work and improve upon already existing engines in the market, or maybe develop an entirely new engine with a team! So the question is, how should I be able to reach that goal? of becoming an engine programmer?

I think a good engine is something that allow someone not that skilled in programming to make a good game/software really easily.

And also that make it easy to a non-programmer (artist) to integrate their content into the software.

+1

A tech demo is fine. You don't have to create a lot of content or gameplay to showcase engineering skills.

But you should realize that most of the value you'll potentially provide to your creative co-workers is in making it easy for them to create content. So beyond your ability to create run-time features, you'll want to showcase your ability to facilitate content creation.

I think a good engine is something that allow someone not that skilled in programming to make a good game/software really easily.

And also that make it easy to a non-programmer (artist) to integrate their content into the software.

+1

A tech demo is fine. You don't have to create a lot of content or gameplay to showcase engineering skills.

But you should realize that most of the value you'll potentially provide to your creative co-workers is in making it easy for them to create content. So beyond your ability to create run-time features, you'll want to showcase your ability to facilitate content creation.

Thanks for the advice! Any examples of what common formats, if any, should I support in my engine?