I am programming my game, a basic 3D RPG, and have made (in my opinion) significant headway on the engine/renderer itself (I'm a programmer, not an artist, if that isn't apparent). I want the game's environment to have huge castles everywhere, but I am not sure how I should do this, in interest of time. Should I just let the castles be made in Max/Maya/Blender or another third party content program, or do you think it would be worth my time making a VERY BASIC castle editor that uses voxels to make a block castle, and exports them as geometry? Which would you use?

I would use (Blender/Maya/Max/whatever you can afford and are comfortable with) and (Gimp/Photoshop/Paint/whatever you can afford and are comfortable with) to build modular castle pieces that can be arranged to construct any number of castles. This includes seamlessly tiling textures for bricks, blocks, rubble, etc... as well as modular pieces of mesh geometry that align with each other to represent walls, floors, arches, buttresses, statues, doors, and so forth.

I'm not sure making a voxel editor to build your castles is a good idea, unless of course the intention of your game is to use voxel castles. If you are building the voxel editor simply as an intermediate step toward making "standard" 3D geometry for your game, then it feels to me like you are setting yourself to a project that would easily be as large as a game,and probably would only serve as a distraction. Just look at the voxel editor being made/discussed at Procedural World; the dude has been working on that thing for a long time, and while it's getting pretty cool I don't think it's "there" yet. You'd be signing yourself up for the same thing, I think. And those things tend to grow, even if you try to keep them VERY BASIC[sup]TM[/sup].

If your intention is to play with voxels, then play with voxels. But if your intention is just to build a 3D game, focus your efforts there and save the voxels for another day.

The easier it is to create content, the more content you can create, and the easier you can train people.
You could always export to geometry and then modify in Blender for extra features when you needed.

Whether it is worth the time investment to make the editor, you'll have to decide yourself.

I agree, I think you're better off importing models from software a lot of people are already familiar with. I've seen Blender tutorials that make structures and tiling look pretty straightforward (haven't tried it myself yet), so it can be done without an inordinate investment of time. Allowing for import also gives you a huge base of models available free online to at the very least test things with.

Programming wise, you can install Blender, make some castle pieces, and then use the scripting engine to generate castles with your pieces and export those. Or you could just use it as a proof of concept before you code it for real, since the script and models can be edited on the fly in Blender.

- Make some castle tiles (corner tower, walls, front gate, whatever, etc... move them to a hidden layer and turn that layer off.
- Write a script that generates some data, possible like how a roguelike makes levels (lots of articles on that), and then copy instances of your pre-made tiles and line them up.
- Export finished level to your game OR code the same algorythm into your game