Though I can't be much help as I don't get on the forums very much anymore I would like to offer some of my thoughts, I have only skimmed your posts so forgive me if I missed something.
First, projects like this are great, I didn't really understand C++ until I finally took up a project and created a procedural world generator this spring. That said, always try to start small on these projects and work from goal to goal. Before my world generator I had countless projects that failed because I had hugely ambitious plans for a game and did not pace myself in creating it, expecting all of the features to be done in an instant.
Last thing, you mentioned generating a small galaxy with planets orbiting stars and such. I haven't yet worked on any generation code for that, but, I did play around with planetary orbits. As you might know, most planetary orbits are not perfect circles, but ellipses. They have a value called eccentricity. Eccentricity can be valued from 0 to 1. An eccentricity of 0 would mean the orbit is a perfect circle. However, as an orbit's eccentricty moves closer to 1, the orbit becomes more elliptical. What this amounts to is that an orbit has a lowest point, it's aphelion, and a highest point, it's perihelion.
That said, the math involved in simulating elliptical orbits confused the hell out of me, and I found simulating circular orbits much more easy to do. It all probably just boils down to what you mentioned earlier, whether you care more about realism or if circular orbits are fine. If you are interesting in elliptical orbits, I would take a look at "Kepler's Laws of Planetary Motion". Circular orbits are fairly simple, you simply have a circle with the radius of the planet's orbit, and an angle which is the planets angle in it's orbit. You increment the angle at each timestep and calculate x and y like so:
x = OrbitRadius * cosine(OrbitAngle);
y = OrbitRadius * sine(OrbitAngle);
Let me know if you would like more explanation and please someone point it out if I am wrong somewhere.