2600 3D Rubik's Cube Discovered!

[Curt Vendel]

Len,

 

Here ya go:

 

http://www.atarimuseum.com/arcade/fullsize/pong.jpg

 

 

 

 

Curt

[Nukey Shay]

One can only wonder if the caption "Video Skill Game" would have still been used had a vector monitor been used. I think it would have (simply because the public knew the word "video" more often than "vector").

[Curt Vendel]

Well, the actual PONG's were shipping with B&W TV's installed inside the coin-op cabinets with black cardboard cut out around them.

 

 

Curt

[rolenta]

One can only wonder if the caption "Video Skill Game" would have still been used had a vector monitor been used.  I think it would have (simply because the public knew the word "video" more often than "vector").

 

Actually, the word VIDEO was not in vogue in 1972. Remember, that was when Sony introduced the Betamax, which was called a Video Tape Recorder as opposed to an audio tape recorder. Although videotape has been around since the late fifties, the average public never had any experience with it, and hence, the word Video was not part of the common vernacular of the day.

 

I think it was MTV that made Video a household word. They're the ones who began calling their videos "Videos", a term I've always hated since you don't call a song on the radio an "Audio".

 

And while I'm on the subject, do you see how the film industry promotes their video releases as being on Video and DVD? Disney started this and the other companies followed suit. Why can't they correctly say a movie is being released on Tape and DVD; or at least VHS and DVD, since both formats are examples of Video!

[david.winter]

Talking about the "video game" term, I noticed an amazing french translation. Univox made several games, of which the most common are the 41 and 61 models. Their box show a large "JEUX DE VIDEO" title, which translates as GAMES OF VIDEO !!!!!!!!

 

This said, I must admit I never tried to find the oldest "video game" statement, yet I think a 1975 game shows it. Or 1976, can't recall exactly. Note that I only have the color PONG flyer, so if anyone could send me a yellow for the shipping costs that would be great

[Nukey Shay]

That's video in the sense of "video unit". What I meant was that the word itself was better known than the word "vector". I mean, was there any 60's kid who didn't know who Captian Video was?

[Eckhard Stolberg]

That's a good question, here's the answer: you can put as many sides of the same color as you want. However, the game will refuse to solve until all faces are put back to their original color. To test this, I changed some colors, rotated some slices and tried the solve mode, which failed and put a yellow background color. Once I put the correct colors back, the game would solve again. Therefore, each face has a dual-color memory: the original color (for solving), and the altered one. I couldn't find a mode to restore the original colors.

 

Does the game really require the original setup to be restored before it would continue with the solve feature, or would any colour setup that would be possible on a real Rubik's Cube work? For example, could you exchange the two colours of an edge piece with the two colours from a different edge piece and the solving routine would still not continue? Given the limited RAM of the 2600 it seems unlikely to me that the game keeps two colour setups stored. It seems more plausible that the game checks if the colour setup you entered is possible at all.

 

BTW, if you really find anything in your Rubik's Cude prototype that works differently in z26 than on a real 2600, It would be nice, if you'd let me know, so that we can try to improve the emulator.

 

 

Ciao, Eckhard Stolberg

[Nukey Shay]

That would be easy enough to check...just manually change the colors to what they would be if you turned a side, and then see if it refuses to solve.

[david.winter]

Thanks for the heads up guys: I forgot to talk about that !

You can't swap the colors. You MUST keep the original configuration to use the "Solve" option.

 

You can cheat, rotate etc, but to get the solve option working, you must put the right colors back. If for example you put all the right colors back except two which are just swapped, the solve option won't work.

 

The game DOES store two colors sets (2 colors a face). We have 6 colors, which takes 3 bits, so 6 bits per face. Total memory required is 3 x 3 x 6 bytes, so 54 bytes, which is more than enough for the 2600.

 

The 2600 can also use the second color set temporarily for the cheat mode: a second color set is created when you start cheating. When you want to use the solve option, the 2600 compares the two color sets. If there's a difference, it refuses to solve the cube and keeps the second color set, thus allowing you to continue cheating or restoring the colors. Now, if the two color sets are same, then the 2600 can get rid of the second and start solving the cube. To use that feature in an relevant manner, the program needs to store two faces per byte, so a whole color set of 27 bytes. In cheat mode, 54 bytes would be used. In normal mode, the second set is useless, so only 27 bytes are used, leaving 101 bytes to solve the cube.

 

All of this is well feasible, even with only 128 bytes of RAM.

[Thomas Jentzsch]

You can't swap the colors. You MUST keep the original configuration to use the "Solve" option.

Is the "original" configuration randomly generated? Or is it always the same?

 

You can cheat, rotate etc, but to get the solve option working, you must put the right colors back. If for example you put all the right colors back except two which are just swapped, the solve option won't work.

Then I guess there is no general solving algortithm built into.

 

If the "original" configuration is not randomly generated, then it's just a precalculated and stored sequence of moves. Else the game probably just records the moves which created this configuration, starting from an ordered cube. Storing one move will easily fit into one byte. And with 20..30 moves, you'll get a pretty random pattern.

The game DOES store two colors sets (2 colors a face). We have 6 colors, which takes 3 bits, so 6 bits per face. Total memory required is 3 x 3 x 6 bytes, so 54 bytes, which is more than enough for the 2600.

Did you look at the (disassembled) code or are you just guessing this?

To use that feature in an relevant manner, the program needs to store two faces per byte, so a whole color set of 27 bytes. In cheat mode, 54 bytes would be used.

The second set is not necessary. If the "original" configuration is randomly generated, then the game just needs to reverse the recorded sequence and check if the cube is ordered then (and play the recorded sequence again to return to the current configuration). And if it's always the same, the game just has to compare it with the stored one.

All of this is well feasible, even with only 128 bytes of RAM.

That's for sure.

[david.winter]

Is the "original" configuration randomly generated? Or is it always the same?

 

I don't know exactly. I can't really disassemble the code to determine if the cube is randomly generated or if the rotations to generate the cube are hard-coded (in which case they must start from a random position as I don't get the same cube each time I reset the game).

 

 

Then I guess there is no general solving algortithm built into.  

 

If the "original" configuration is not randomly generated, then it's just a precalculated and stored sequence of moves. Else the game probably just records the moves which created this configuration, starting from an ordered cube. Storing one move will easily fit into one byte. And with 20..30 moves, you'll get a pretty random pattern.

 

There must be a solving algorithm, because you can rotate as many times as you want and the game will still solve the cube.

 

 

Did you look at the (disassembled) code or are you just guessing this?

...  

The second set is not necessary. If the "original" configuration is randomly generated, then the game just needs to reverse the recorded sequence and check if the cube is ordered then (and play the recorded sequence again to return to the current configuration). And if it's always the same, the game just has to compare it with the stored one.

 

The program will solve the cube even if you do 1000 rotations, so it is useless to record the rotations done to generate the cube.

 

I'm quite convinced there are two color sets in cheat mode (just used to determine whether the cube has been altered before solving), and that there is a solving algorithm.

[Thomas Jentzsch]

There must be a solving algorithm, because you can rotate as many times as you want and the game will still solve the cube.

Ah, now I understand! You can puzzle the cube yourself (by rotating sides) and the game will sove it then, right?

 

And if you can rotate it more than ~200 times, then there is a solving algortihm build into. Cool!

I'm quite convinced there are two color sets in cheat mode (just used to determine whether the cube has been altered before solving), and that there is a solving algorithm.

Yup, with the new information, I have to agree.

 

BTW: Does the 2D Atari Video Cube have such a solving feature, too

[david.winter]

BTW: Does the 2D Atari Video Cube have such a solving feature, too  

 

From what I read on the web, it does, but I'm not definitely sure. I just got to the page where you can get $8 repro carts. They have a page where they list 8 rare games per month. This month lists Rubik's Cube (2D version) and they say it has the solve option. However, I don't know how this woks, especially if you can also change the colors in the 2D version.

[Xot]

BTW: Does the 2D Atari Video Cube have such a solving feature, too  

 

From what I read on the web, it does, but I'm not definitely sure. I just got to the page where you can get $8 repro carts. They have a page where they list 8 rare games per month. This month lists Rubik's Cube (2D version) and they say it has the solve option. However, I don't know how this woks, especially if you can also change the colors in the 2D version.

 

 

Atari Video Cube does have an automatic play option, where the computer shows you the minimum number of moves required to solve a particular cube.

[Retro Rogue]

On a sidenote, three weeks ago, Entertainment Weekly, a magazine put out by Time, did a cover story about videogames. They mentioned in their timeline that Ralph did the first videogame console. I wrote them to thank them and told them that this was the first time I saw Ralph's acheievement acknowledged in a mainstream magazine. They wrote back and said they were considering printing my letter and if so, it should be out today.

 

Congrats Leonard, hope they print it. On the other side of the coin, I sent them something congratulating them on the article itself but ranting about the inaccuracy of the timeline. Some examples: The Channel F being released in '76 as the first cartridge system (when in fact it was the Video Entertainment System, and was the first PROGRAMMABLE system, not first cartridge system). They also have Atari's Asteroids "Hitting the stores" in 1979 (methinks they can't tell the difference between an arcade coin-op and an Atari 2600 cartridge over there). I also told them about the importance of having technical information right in an a magazine with that wide of a distribution, as people will take that material and reference it as fact.

 

For some reason, I don't think they'll print it though.

[david.winter]

You're right and it's always the same damned thing. You take some significant time to provide some information, but they never manage to understand that you provide something accurate. They prefer getting more infos around, and put what THEY think to be right. If there was a law that forced them to publish corrections, I think there would be much more accurate stuff in the non-VG press (yet some of the VG press also put false things, but that's much less than the rest).

[Marco(2)]

It is necessary, if one wants to proceed toward clarity of perception, to understand the shallow and synthetic nature of "knowledge" in any field and therefore the thinness of expertise which any authority on the subject would have.  

 

If you consider the subject of history, for example, knowledge consists of a collection of supposed "facts", including dates, descriptions, and interpretations of events significant to the flow of culture. But those descriptions, while often given de facto consideration as reliable if commonly accepted in a particular culture, originated only ever in personal accounts, either passed on by verbal description or written some time subsequent to the event. These accounts are always oriented and colored to some degree -- and possibly extensively -- by the bias of the observer, and pick up subsequent distortion by any later turnover through translation or compilation. As cultural attitudes and values evolve, such accounts will be again revised to suit newer interpretations.  

 

History, in other words, is not the carved-in-granite collection of objectively gleaned accounts and factual depictions it would seem. It is a collection of passed-on, subjectively perceived and interpreted experiences, often retold numerous times and invariably revised according to the latest reteller, thus subjected to his/her bias and currently fashionable historical perspectives.  

 

What I'm saying is, history is rather an illusion of established fact, based on accepted sources. You will note that historical accounts vary widely from country to country according to the perspective, and, again, change regularly in any given culture based on current values. It's not just that different people see history differently, i.e., interpret facts differently -- it's that subjective accounts are all that really exist! There is no absolute truth about history, no objective record of events, no real "facts", only accounts.  

 

Try to grasp that fully before you proceed. History, as a recording of facts, is an illusion. It is a compilation of subjective accounts, nothing more.

 

Taken from: The Essence of Reality

 

Cheers,

Marco

[Eckhard Stolberg]

You can cheat, rotate etc, but to get the solve option working, you must put the right colors back. If for example you put all the right colors back except two which are just swapped, the solve option won't work.

 

Swapping just two colours would most likely result in a setup that is impossible on a real Rubik's Cube. The Cube has 6 center pieces (1 colour), 8 corner pieces (3 colours) and 12 edge pieces (2 colours). For the solving algorithm to be able to work, you can change the positions of the corner and edge pieces, but the colours on all the pieces itself must stay intact.

 

So if you have one edge piece with a blue and a red face and one with a blue and a yellow face, could you swap the red and the yellow faces, or would that break the solving routine? If this were possible, you could probably use the colour selection feature to input the setup of a real scrambled cube, and the 2600 would help you solve it.

 

 

Ciao, Eckhard Stolberg

[david.winter]

For the solving algorithm to be able to work, you can change the positions of the corner and edge pieces, but the colours on all the pieces itself must stay intact.

 

Therotically yes, but the program won't let you do this. You have to restore the colors. Even if you leave a solvable configuration that differs from the real one, the game won't work.

[david.winter]

Sound in the game !

 

I also forgot to mention the sound "effects" in the game:

 

A single (different) note is produced when you rotate the cube and slices.

An amazing little song is played once the cube is solved.

A crash-like sound is produced when the Solve option fails (cheats).

[Thomas Jentzsch]

Therotically yes, but the program won't let you do this. You have to restore the colors. Even if you leave a solvable configuration that differs from the real one, the game won't work.

That's probably something that would have been fixed in the final release. Or the cheat mode would have been removed completely.

[Nukey Shay]

The instructions to solve any scrambled cube (that is solvable) could be programmed in well under 1k of rom...depending on the type of solution used. Though an unsolvable puzzle (where one square was re-oriented manually) could potentially crash or hang a game that included those solutions.

[david.winter]

That's probably something that would have been fixed in the final release. Or the cheat mode would have been removed completely.

 

Supposing the game is unfinished, it is possible that the programers would have improved it.

However, the whole game fits in 4K, so why not leaving the cheat mode for the fun of it ?

 

The problem is that even if you suppose whateter makes sense, we can't be sure whether the game is finished or not unless somebody finds some original docs stating so, or manages to contact the programers who still remember if the game is finished or not.

 

It would be amazing to find the programer(s) and get more infos from them, especially concerning the real history of that game. We still don't know why the 3D version was not released.

 

My opinion is that Rubik's Cube was the original name and that the game was originally designed in 3D (which makes more sense). For some reasons, Atari got in trouble and replaced the game with the 2D version. They kept the original title until they changed it to Video Cube for other reasons we don't know. That could be copyrights litigations, game features not accepted, irrelevant title compared to the gameplay, etc. If the game is finished, there are good chances that it be due to copyright litigations, since the crash didn't happen yet (as of 1982).

[Thomas Jentzsch]

Supposing the game is unfinished, it is possible that the programers would have improved it.

However, the whole game fits in 4K, so why not leaving the cheat mode for the fun of it ?

 

The problem is that even if you suppose whateter makes sense, we can't be sure whether the game is finished or not unless somebody finds some original docs stating so, or manages to contact the programers who still remember if the game is finished or not.

Agreed, we can only speculate here. When/if you release the ROM someday, a look into the code might help to answer some questions.

 

My opinion is that Rubik's Cube was the original name and that the game was originally designed in 3D (which makes more sense). For some reasons, Atari got in trouble and replaced the game with the 2D version. They kept the original title until they changed it to Video Cube for other reasons we don't know. That could be copyrights litigations, game features not accepted, irrelevant title compared to the gameplay, etc. If the game is finished, there are good chances that it be due to copyright litigations, since the crash didn't happen yet (as of 1982).

But why did they had to change the graphics? They kept the complete gameplay and the cube, just displaying it in 2D. But you could still see that it was a three dimensional cube you are playing with. Strange!

 

Hm... Are there any problems with the user interface in the 3D version? Or maybe somebody else already had bought the right for a 3D version?

[rolenta]

David,

 

Just wondering since you're in France. Is the game in NTSC or PAL (or SECAM?)