I guess the next step would be to add walls or something and detect collision but I get the feeling that will be much harder haha.

Yeah, just keep in mind that you need to have — review it if you already know — the basics of algebra and trigonometry.

With today's available resources, videos, etc. this is impossible to be a problem.

For each topic in the game development field this web-site has a section available.

Without checking out a guide what comes to my mind is creating some sort of group with every collisionable object on it, and check for collisions on update (With the excepcion of colliding with itself).

I'm really fired up o:

Avoid xna itself as in the framework available from Microsoft and instead go straight to monogame, the open implementation of the same framework which is still going through active development unlike Microsofts implementation which hasn't had a new release since 2010.

Good luck in your game!

Mono is kind of a successor to XNA right? Doesn't that mean I can apply all I learn in XNA to Mono?

You need to decide first if you want to do a game or a game engine. Both things can run in parallel, but trying to reach the first one without taking the second one steps can ruin your learning process.

IF his goal is to become a game engine programmer, you are right. If he wants to be as "hirable" as possible to a wide range of studios, I guess it cannot hurt too even if he applies to other jobs like UI programmer or stuff like that (don't know how "focused" programmers need to be in Game Dev, in Business Dev I see lots of jobs with a small list of mandatory skills, and a HUGE list of nice-to-have skills).

If he just wants to make his own games, or doesn't want to be the engine coder in a studio that does not look for these nice-to-have skills, writing full on 3D engines will not help him much (other than getting a deeper understanding of existing engines, which is great of course). I agree that just diving into working with existing engines without tinkering with writing games from scratch can make understanding the internals of existing engines harder. Still, I wouldn't tell a beginner to start writing his own engines just to get the skills to write a game. A modern 3D engine takes lots of time to create, time that could be spent learning other things or already working on a game.

Of course, writing small engines as a learning milestone I completly agree with. Anything that takes less than maybe 6 months to complete might be a good idea... just don't try to write the next Unreal engine...

Just my 2 cents.

Well, I just got accepted my proposal for course project, so I'll be focusing on that. It's a sidescroll game for android so I'll be working on that o/