Some time ago I was lost same as you. I'll try to do my best to give you few advises on what you could do at the moment. As I understood from your post you have pretty decent knowledge on C++ ( I assume that OOP paradigm [at least the basics like types of constructors and how to use them, inheritance, maybe polymorphism etc.], at least basics of STL (very, very useful in game programming at least for me), strong foundation about raw pointers/smart pointers are not something that is difficult for you ) as well as I hope you spent some time to study IDE you are using - especially the debugging aspect, as it can save you from hours (literally!) of frustration while looking for some dumb bug like forgotten semicolon at the end of class definition
Well, about the tips. Firstly, it's very good that you have a goal you'd like to achieve and in my opinion you should stick to that goal and do everything that's possible to make it happen. Most important in order to "survive" is to find a realistic goal, then to make sure that you possibly could make it happen with the knowledge and assets that are available to you and then just keep going for it and remember to NOT TO abandon what you already have started, because of laziness as it will bring MUCH reluctance to future projects or maybe even programming. That's why you should always keep in mind to keep your goals relatively simple and doable.
Next tip would be ( when you already have started some project ) to try to split bigger tasks into easier ones like:
"Implement basic logging to my engine" could be broke into:
- "Make an decision what single log line should contain"
- "Create an abstract skeleton class containing only what's most necessary"
- "Fill methods with code"
- "Instantiate logging class where it's necessary"
- etc.
Above breakdown could be done far better with some time spent on it, but it's just an example, which I hope is quiet clear and understandable. Bare in mind that even tasks that are child of parent tasks can be broke into even easier tasks if you find them to complex. Just don't go for something like: "Wash my teeth", please
First thing I propose is - don't try to create an engine pointlessly. If you are serious about making a game from scratch with help of OpenGL first thing you'd have to do is to spend some time to gather thoughts about what kind of game would it be, what would it need, would you be able to make it, would you be able to find necessary assets ( I assume your first game wouldn't be a commercial one ) or maybe some people could possibly make them for you ?
Then break it down to smaller ones just like with programming tasks, so for example if you'd like to make a game similar to Minecraft you should ask yourself any question about the gameplay you can think about, like would it be from 3rd person or 1st person or maybe side camera ? Would it just be single player or do I have necessary skills/could I learn necessary skills to make it multiplayer ? Would the world be generated based on seed or would I rather use presets ? etc. etc.
After finding an general outline you can and you should break down those thoughts into even more detailed and so on as long as there are any thoughts in your mind about the game. Oh, and remember to note EVERYTHING. It's really, really essential to do so. After you are done with non-technical description you should try to transform it into a kind-of framework documentation though not really a one. Just clean-up your thoughts into some framework-compatible order, maybe try adding some technical comments about implementation details, maybe try to attach some pseudo-code, umm you could try UML as well.
Basically that's what you should do before you start coding. Project planning as well as project management are HUGE fields to play with and you could find tons of books about it, but as you are programmer it's not your goal to be an expert in that field ( as long as you do not want to be one, but in my opinion as a beginner you shouldn't bring too much on yourself ). Remember to keep doing it in funny way to make the creation process entertaining as well. It will help you work longer. You could try to involve your friends as well, even if they are not programmers or 3D-artists or 2D-artists or audio-engineers they can still give you some ideas and make your game just better.
That's about it for my tips for now. If I'll remind myself anything more I'll edit my post later.
Most probably you won't take what I wrote here seriously ( I wouldn't not so long ago ), but just that you read it makes it count.
Best of luck to you !
Btw. Unity3D is free for commercial use as long as you do not exceed 100,000 $USD of income yearly though free version is a bit stripped you could still find it very useful. Commercial license is not that expensive as well. It's 1,500 $USD if I remember well.
Btw2. UDK by Epic Games and CryENGINE 3 SDK by Crytek are both something that could possibly interest you as they are very cheap ( CE3 SDK is free for indie devs if I remember well ) and very mature and advanced tools built by efforts of dozens of programmers and artists from all over the world through the last decade.
Btw3. It's totally your call, but I personally find Direct3D much easier to understand and much clearer in code then OpenGL. You could also make yourself familiar with the quiet a bit of lecture in here
http://programmers.stackexchange.com/questions/60544/why-do-game-developers-prefer-windows