cignox1

Well, they could tell you which is the language (mostly) used to build core engines of games such as Crysis and the like. But as already said, that info will not make you a better coder, not will make you write your first games (which is something far more important).

One of the people assisting me with the game is insisting I use OpenGL, saying that slowly it's going to replace DirectX.  And that it's "The graphics library of the future". I just don't know if I can believe him.

 

Is it possible that he refers to the recent projects of Valve in regards with linux? Valve is currently trying to lead a move of the game industry towards linux, and this means that those developers willing to follow them have to use OGL. Is not yet clear whether or not that shift will happen...

I must say that you have indeed beatiful renders, wich would deserve more complex scenes (i.e. Sponza Atrium).

If you have troubles with kd-trees and the like, try with something easier: BIH (Bounding interval hierarchy) is not as fast as kd-tree with SAH, but estimates suggest it is 60/70% its speed, so it may help you a lot, and is not so hard to implement (I did it as well, but I would rather not give you my implementation as for some reason it is really slow :-(

You can find the paper by googling and the ompf.org forum used to have several threads on that topic, with many implementations discussed (included my own). Now the ompf.org domain no longer points to the forum, but a (temporary?) replacement has been opened. I don't know if those threads have been moved, but if you ask them they will help you.

how should I go about simulating subsurface scattering (which is needed for realistic metal materials)

Honestly this is the first time I hear someone willing to implement SSS for metals (instead than skin, milk or marble)

And you could consider buying the "Physically based Raytracing: from theory to implementation": I have the first edition and it is truly amazing. In the second (IIRC) the show how to implement both BVH and SSS.
If you are interested in Raytracing/GI you should buy that book.

Hope this helps

If you want to target C++, then IMHO you should just decide between Java and C#. Python is nice and easy, but is too different (different syntax, different compilation model ...)

Usually this is not a problem at all, but if you want something that simplifies your path toward c++, then C# or Java.

And among these two, I would suggest Java: C# is a very beatiful language, but too much bound to Visual Studio: you could miss important steps performed by its features under the hood.
I suggest  Java (notepad++ + javac at first) and then try NetBeans or Eclipse.

Mind you, that's just my opinion: if you pick C# instead that will work nonetheless.

Learning C before C++ only means that once you move to C++ you need to forget all the things no longer needed. Why should you learn to use c strings when you can start with std::string? Or why should you ignore references? Or to use malloc/free instead than new/delete?

A better option would be to learn 'C++ without classes', but still, a lot of languages (Java and C# to start with) are highly OO so you can as well learn C++ the OO way...

I don't see reason why you would want to learn C if you target C++. And I don't see reason to learn Pascal if you want to program with Delphi, for that matter...