Rubik's Cube for the 83+

[DarkerLine]

As my first real attempt at assembly, I'd like to present this as-yet-unfinished Rubik's cube program. (You may have seen me announce my plans for it on my blog, in not as many words)

Here's a screenshot of it in action -
[attachment=1997:attachment]

Right now all the actual workings of the cube are finished - it's for the MirageOS and saves your state when you exit. I plan to add more features later, and also optimize the code to a (hopefully, much) better size than it is right now.

For people who want to try it, and I urge you to because I want feedback, the program is attached. The controls are as follows: CLEAR exits, arrow keys rotate the cube, and pressing three number keys such as 1,2,3 in order (which I hope can be done by brushing your finger against the keyboard) rotates a layer of the cube in the intuitive way - imagine that the number pad is the front face of the cube. I hope this isn't a disgusting method of controlling the cube, because I had fun coming up with it.

Please try it and tell me what I can improve on (in every aspect - graphics, user interface, speed, anything). Also, I haven't tested it on an actual calculator, only on PindurTI, so I may have missed things.

Edit: fixed a typo which caused the calculator to cheat and switch stickers.

Edit: additional controls: holding down MODE for a couple seconds resets the cube, while holding down XTΘn for a couple of seconds scrambles it.

[tr1p1ea]

Looks pretty sweet DarkerLine, top job i say . How are you handling the cube, with sprite sections or some other approach?

[Harrierfalcon]

This should be interesting . Looks good so far, DarkerLine!

Any chance of grayscale?

[DarkerLine]

[Liazon]

[justusdude]

I would suggest adding in scrambling.

[DarkerLine]

Done

[elfprince13]

[DarkerLine]

[luby]

would this help?

[DarkerLine]

Possibly, but what I'm considering is a form of cheating. It seems wasteful to write an algorithm that will solve each position independently. Instead, the algorithm could maintain a solution sequence, and update it with each new move. I don't know if this will be any better than the other way, but it's worth looking at.

[magicdanw]

That would work, but imagine someone solves it correctly up until the last few steps that they can't remember, so they click solve to see how it's done. The cube would then proceed to re-scramble itself before unscrambling in an atypical fashion. I think that would be very annoying to the person who wanted to learn how to solve it themselves.

[DarkerLine]

[AaroneusTheGreat]

It is possible and has been done.

Rubot II (ONE)

Another Rubix robot.

[DarkerLine]

Well, obviously it's possible to solve the Rubik's cube. That is not what I am curious about. I want to know if it's possible to find a shortcut to figuring out how to solve it, using the method I outlined...

[simplethinker]

Are you talking about something like God's algorithm? If the computer is powerful enough, I think it would be possible for it to come up with a few (hundred-thousand, maybe more) ways to do it (based on some pre-programmed strategy) and then test to see which ones were the shortest, so you would get practically the best way to solve it, but it would take a while to test them all.

[DarkerLine]

Of course, I could just implement a conventional algorithm, but that would be no fun.

I mean, I'm probably not going to implement any sort of Rubik's cube solving at first, that will have to wait for a new version.

[elfprince13]

a student in one of my dad's classes a few years ago wrote an interesting applet that does it step by step, which actually got picked up by the Rubiks company, and is available somewhere on their website....he's also one of the worlds fastest rubiks cube solvers. anyhow, I doubt the source is publicly available, but there are some fairly effective java decompilers available.