A little tip. If you're gonna make a game, first make some smaller, simpler games in the same genre. By simple I mean especially the graphics. Good graphics are pretty hard to archieve, most game studio's have whole teams of artists for character and level creation, and in my experience, good graphics are one of the hardest things to archieve (together with a general fun game experience: making the whole game not too hard or too easy, not confusing the player, and keep rewarding the player without giving him more of the same every time).

The first small game will probably be a mess, maybe you won't even finish it. But, as a result, you will have a better idea of the structure the program needs to have. Sometimes, when you're able to implement a new feature in your old engine, you will think 'OMG, if I redesigned the whole thing it would be so much easier!'. Keep your old projects, and comment everything well, so you have usable source to consult or even use directly if you're having trouble. Also, think about splitting things up in classes, if this doesn't hurt the readability or performance of your game. Also save useful functions in a seperate file.
For me, this way of working results in much cleaner and faster code. Please note I didn't ever code a big game, so I don't know how well this way of working scales with the projects, possibly this is too much work for big projects.

Last thing. You've probably heard this before or thought of it yourself, but seperate your engine from the levels, graphics, and gameplay as much as possible. It will be easier to make special effects (graphical effects or gameplay-wise, for exception, inversing the gravity will be easier when you have some force vectors instead of a hardcoded "player.y += 5;" each frame (as a figure of speech)). This can cost a little bit of performance, but things like this can usually be done pretty efficient. Else you just have to find a balance, or try both ways and then decide what's the way to go.

Good luck

Trying to avoid the flame war going on here: Flixel and Actionscript.

I'm a very hardcore C++ programmer (I'm working in the industry for about a month with 8 years of indie game development) and in all that time the only moment where I went "what the fuck" was when I was using Flixel and Actionscript. It was a marvellous moment of joy to see in a week of work a game I could show everyone. In C++, Java, etc you are going to find yourself fighting the game mechanics AND the language, libraries, compilers, etc. With Flixel and Actionscript you are going to get going like a champ in no time. Learn the basics, classes, accessors everything and at the same time make a game fun and fast. For me the motivation is key and specially at the beginning so I'm recommending it to all new people to programming. Also try to make a pong or something along these lines to get going. If you think it's aiming too low, then it shouldn't take you more than a week shouldn't it? ;) And after that you can try new things.

That said, XNA and C# is fast too, Unity 3D is also kind of good for experimentation although I wouldn't recommend it because it has no real practical use in a professional career.

After your first game, review the technology and try different things.

First things first congrats on listening to your gut. Too many people stick with a job/field/career just because they have time invested in it. If you get deep into something and realize that you're not going to be happy doing it then you should find something else. Life is way too short to be stuck doing things you don't like. If you like the idea of coding and making games then dive on in.

There are a million different approaches to learning to code and a million different technologies. The key is to keep learning new things and keep iterating on what you're doing. No one ever gets to the point where they have learned everything about software engineering. If game development is what you think is going to be your passion then try something like Unity where you'll pick up the basics of C#. If you still have that passion after learning the ropes then I would suggest diving into C/C++ next since that will give you the strongest base for jumping to any other language since it's nice and close to the bare metal. Of course that's just what would work for me. Everyone is different.

Best of luck in the new endeavor and enjoy the ride.

LOL tell me a bad line that'll blow up my windows. User code not kernel code.

One bad line of code from a novice, can cause permanent damage to the OS in C++ .

Every program runs in its own address space and one program cannot access any other programs address space, this is called paging. If you try to access an address in memory that your program cannot read it will cause a segmentation or page fault and your program will crash. No damage will occur.