I'm myself having fantasies about creating a mod (having a job and a family with 4 kids ranging from diaper-age to adult, I got time constraints that forces my spring modding plans to be defined as 'fantasies'
) and for anything gaming related I've taken a key principle of extreme programming at heart:Start with something so ridiculously simply that it takes little effort to create it and to make it "completely working". (That way, one has something that "works" from the start and thus can be verified to "still work" even during the first iterations of improvements that is part of the development cycle).
So, assume you try to create something with the most primitive graphics you can imagine (64x64 textures does not look like a commercial game but they can still define something and to define a unit with some volume, you only need 4 triangles to define a tetrahedron, really) and that you use a minimal set of unit types like this:
- one single type of mobile unit for combat and construction
- one single type of immobile unit as a factory popping out the mobile kind of unit I just mentioned
- one single type unit being a metal extractor
Focus on creating a mod like the one described above, without any kind of cool innovative features and gimmicks we all have percolating in the back of our head as "damn, that I would like to see in a mod" (those features can be developed later) and I think you have accomplished to
- to define a manageable number of hurdles and obstacles to confront and overcome. Yes, even if the above mod specification have an ambition level that is "lower than whale shit", it still contains a number of pitfalls, caveats and lurking dragons that will bring quite some experience points to an adventurous, novice mod creator.
- to define something that takes minimal time before you can say "ok, it looks like whale shit, but man, it is playable! (...involve a co-developer, and you both can play with each other and get tons of encouragement and motivation by playing your own creation!). How long time that is depends on tons of factors, and I'm not even a n00b on mod creation so I can only give you a bad guess... 40 hours? That bad guess includes time you need for getting up to speed on the various tools involved though and selecting some of said tools in the first place (Blender for 3d? Gimp for 2D textures? Choices, choices, choices...) etc.
Ambitious projects that never manage to reach usable status might be educating, but they don't have that "wow" feeling of something that is complete and at least runnable at a basic level. And how often haven't one happened to be over ambitious in various situations...
Then of course, aim for the stars and reach the top of the trees etc, but make sure that the steps you try to take to reach those stars are small enough to be manageable.
Good luck!
) it would've been impossible. There's a varied skillset you need before you can start producing standalone content for any game engine. or Spring, you need to know how to model, map, skin, edit FBI, write Lua, make COB, use various spring-related programs, make decent balance, and advertise, along with many I have probably neglected to mention. Joining a game development team can help you acquire these skills, and earn you some prestige among the community, so that others may volunteer to help you. Your co-developers at a game development project can give you useful advice, teach you some fundamental concepts, and help bounce ideas, all of which are very valuable services. Most importantly, you actually contribute something to the community, rather than sitting around with some dreams in your head. 
