And the blue is a little jarring against the red/orange, it might help to tone that down a bit.

Literally like a 10 second edit, just purpled and desaturated-

I'm in the middle of remaking the crystal cruiser from FTL (see this thread), using Facet's ever useful post as a very close reference, and I feel like I'm really starting to lose some definition in favor of lots of detail, but even the details don't seem to come across right. I have no idea what I'm actually doing, just putting down whatever I feel like and seeing if it works, is there some kind of logic I could be following to get the forms to read better at this size and have more natural reflections on the inside?

This year I'm in a really open ended class over the fundamentals of art, and considering this is the only static visual art I can do well I knew I'd work it in somehow. I did some pixels physically with squares of paper a while ago but that was a little boring, so I figured I'd take advantage of this and be a little more ambitious.

So here's what I'm thinking of doing- I'll (somehow) get 256 wooden cubes (which fills 16x16) and paint each side a certain color. These will be out in a public exhibition space, and over the year, maybe every week, I'll doodle things until I'm happy with one, probably focusing on the really important parts of pixel theory, go out and make that for display, and then work on one for the next week. Maybe I'll even encourage people to work with it themselves since it'll be made to be modular.

But since I'm nowhere near an expert there are still a few things I need to figure out, including (and some of these are partly/mostly rhetorical)-

-How do I choose the most versatile six colors possible?
-What would best represent the basics of pixel theory?
(as an aside, I just realized that I can use negative space between cubes as a seventh color)
-Where do I find 256 wooden cubes?
-How do the relationships between pixels change when viewed so close and slightly separated? Would antialiasing still work if I can't imply shapes as easily?
-Are certain kinds of art (abstract vs. realistic vs. design oriented vs. object oriented, etc.) better suited to these limitations?
-Are there ways I can take advantage of how I won't be restrained (rotating the grid 45 degrees, varying spacing between cubes)?

I'm rambling a bit but there are a lot of possibilities here, it might be interesting to see how everything we talk about holds up when you stretch the medium.

EDIT: I didn't really get into this much but the "point" of this would be how the cubes are always the same cubes with the same colors all the way through, but they can be used by anyone to make an infinite amount of anything. It's a little pretentious but it works as message of community/universality and all that, and it really hits the nail of simplicity in pixel art right on the head. I'd probably keep a blog or whatever and have people email pictures of what they do, that could make for a pretty decent project by the end of the year.

I also am very curious how I can make choices such as these.
Not only with purple but those lovely ramps I've seen Helm (and others) do with stuff like dark shades of muddy green mixed in.
It reminds me of this - http://wayofthepixel.net/index.php?topic=13186.msg133574#msg133574
Is there any logic to getting such different colours to blend together so well?
I've been messing around with that off and on for a while but can't seem to grasp it 
How to blend without colour separation??

I think this information would also benefit the OP...

This is something I've always wanted to know too, sometimes you even see it in (stylized) portraits and things and I have no idea what the theory is behind it.

Bit of a long bump, but I saw this today.