h a l f b a k e r y"Not baked goods, Professor; baked bads!" -- The Tick
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
I thought of this in 1991 as one of several ideas to start a business. In the end I helped set up a dial-up ISP. With hindsight I think I picked the wrong opportunity :-)
If you're familiar with Conway's Game of Life, it's usually played on a computer screen with a program calculating the generations.
Halfbakers have suggested other implementations such as on solar-powered wallpaper.
My invention is far less grandiose, and is quite buildable with current technology. I was reminded of it when I recently saw a toy called Cube World, in which stick figures interact with their neighbors. The cubes in that game looked very similar to what I designed - they can connect to any other cube above, below, or to the side, and have three electronic connections on each face which in my design at least, provided data in, data out, and a clock. You could probably also power them over the same three connections rather than have a battery in each cube, because in my toy you're going to want a LOT of cubes.
The Game of Life was a major draw in its day; everyone in our Computer Science department was playing it for months - it was far more popular than Rubik's Cube. This invention would not only allow a new generation of kids to have that experience, it would also free it from the straightjacket of a computer.
From the marketing point of view it's an open-ended sale, like Legos - you can keep buying more cubes as you explore larger worlds.
You could also sell boards whose main purpose is to wrap around the signals on the edges.
A world is initialised by pressing the face of each cube to be illuminated; the faces are touch sensitive as well as having a LED or LCD for display. Holding any one button down for a specific length of time would kick off the generations.
(?) Cube World
https://www.dealext...details.dx/sku.5571
My cubes look a lot like these Cube World cubes... [gtoal, Feb 10 2006, last modified Dec 15 2007]
Modular_20neural_20network#1124062430
I had a similar idea... [xaviergisz, Feb 10 2006]
Register Transfer Modules
http://www.brouhaha...p16/rtm_modules.pdf
Needs a circuit something like the evoke/arm in these [gtoal, Feb 17 2006]
DIY life hardware project
http://www.sparetim...dware/Life_Game.htm
microcontroller, array of LEDs [gtoal, Mar 26 2007]
tilable boards rather than cubes
http://www.ladyada.net/make/conway/
courtesy of Lady Ada. Not to be confused with Lady Gaga, except perhaps on Google Plus. [gtoal, Sep 02 2011]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
Give it colored lights and sell it as a toy, and you'll make a mint. |
|
| |
Can someone tell me why Americans tend to pluralize (sic) 'Lego' to 'Legos' whereas most European nations don't ? |
|
| |
Game systems like this are a key plot element in Brin's "Glory Season" |
|
| |
neilp: I wouldn't know, I'm British. And was brought up on Lego and Mecanno. However when I said Legos I was using it as a contraction of "Lego bricks". Maybe that's how Americans think of them too, as individual units of consumption rather than as a mass entity. Or it may it's somehow tied up in the way that the Americans say "The BBC is" vs how we say "The BBC are". (We do. don't we? I must admit I haven't seen that construct as much nowadays as I did in the 60's & 70's when people were more bothered about such things) |
|
| |
// what goes in the cubes? //
state machine logic for a single cell of a Life array.
A LED or LCD that can be on or off.
Clock logic (needs a 2-phase clock synchronised across all cubies).
a button on the surface for use in setting up the initial state and kicking off the generations.
Possibly a power source.
It may *all* be possible on a single chip actually. |
|
| |
//why is that Americans pluralise Maths//
I think that Maths actually is a contraction, of mathematics. |
|
| |
bigsleep - true, in the abstract case, but if we're genuinely trying to make this work as a toy, then the cost per cubie has to be no more than say 50c each, which means a manufacturing cost of under 10c. You don't get much of a computer for 10c nowadays - that's about the cost of a very cheap watch. And about the same complexity. |
|
| |
[gtol] As far as I can tell this game relies on iterations or generations. If one were to build these cubes and place them side by side how do you envision the trigger for the iterations or generations. |
|
| |
Would it be automatic or would it be user invoked? |
|
| |
What would be more fun and interesting for the Player? |
|
| |
Would you envision this game toy working 3 dimentionaly? |
|
| |
Would the cubes need to push each other apart or just light up. |
|
| |
I kind of like the idea of an interactive toy I'll give at plus. |
|
| |
//As far as I can tell this game relies on iterations or generations. If one were to build these cubes and place them side by side how do you envision the trigger for the iterations or generations. // |
|
| |
You may allow an initial single-step for the curious to learn the game, but it would rapidly get boring. The point of Life sims is to run as many generations as fast as possible. Search on the net for Life code and you'll find some *very* optimised tweaking. |
|
| |
As I mentioned, the trigger can be a simple extended button-press on the face. It needs a button anyway, to set up the initial states. Single-step mode could be as simple as just keeping it pressed indefinitely and not letting up. |
|
| |
// Would you envision this game toy working 3 dimentionaly? // |
|
| |
Unlikely in true 3D as you wouldn't see the cubes inside the solid. But maybe as the surface of a larger cube. But that wasn't my plan, I saw it as being planar. |
|
| |
// Would the cubes need to push each other apart or just light up. // |
|
| |
Just light up. They need to be in contact for the circuits to work. Either stuck together magnetically, or interlocked, or simply plugged into a Lego baseboard if made to Lego-compatible standards. (Which is a last resort due to licensing rights :-) ) |
|
| |
For the mechanics of the cubes I would suggest tge following. |
|
| |
Ok we build cubes with 1 metal ring on 4 sides and 1 inner metal dot in the center of each metal circle.
these will provide the electrical contacts no matter what way they are put together. |
|
| |
Inside the cube there are are 4 small spherical NdFeB magnets in loose fitting spherical plastic pockets one on each side where the metal contacts are. These will attract eachother no matter how you arrange them. The magnets will cause the contacts to make contact. |
|
| |
The magnets can be manufactured for about 10 cents each. |
|
| |
This is where cost comes in. |
|
| |
The way I see it logically there are up to 5 inputs to keep it simple 2 output signals. |
|
| |
1 button or touch sensor and 4 inputs for each of the 4 contact sides. |
|
| |
The 4 of the input signals can be infered one of two ways. |
|
| |
1. If you use an RS-485 signal each cube can have a 'mac address' This may prove to be kind of costly both power and money. The chip would have to be cheap and not a power hog. However these were manufactured in bulk the dollar cost could be brought down. Research needs to be done as far as what chip to use. Maybe one of you guys can suggest a lower power chip and or communication bus |
|
| |
2. TTL logic could be used this uses and depending on the complexity of logic there need not be any programing. This option would need to use some signal across the contacts so that the cube with the contact pressed would get an address of 1 and each successive cubes recognised address increases. This could have the advantage of only being on when 1 of the buttons is pressed. |
|
| |
I think I prefer option 1 if it can be made cheap enough. |
|
| |
That's all I have for now maybe I'll get a chance to look for a chip sometime. |
|
| |
// If you use an RS-485 signal each cube can have a 'mac address' ... Maybe one of you guys can suggest a lower power chip and or communication bus// |
|
| |
Ed, there's no need for any sort of bus or addressing, you only need to look at the state of your adjacent neighbours. The only global signal is a 2-phase clock - sample neighbours on phase I (or rising edge), make change on phase II (or falling edge). |
|
| |
You could do away with a synchronous clock if say the lower-left corner cubie sent out a 'sample now' evoke signal which propogated upwards and rightwards at its own speed; then when the wave of evoke signals hit the top-right corner, that cubie would send back a 'ok, make the change' acknowlegement signal in the opposite direction. The whole layout would be self-timed, slowing down to match the propogation delays. (Anyone remember Gordon Bell's PDP16 Register Transfer Modules/flip chips? - something like that... I'll add a link because they're fascinating reading if you're too young to have used them at University) |
|
| |
By the way, apropos of nothing much in particular, I discovered today that you can have interesting triangular FSMs too. Do a google search for this exact string: "Strange Universe is a triangular cellular automaton". Makes me wonder what a Hex grid might do. Hmmm.. yup, that's a "me too". Search for "As far as I know, no one else has come up with a hexagonal version of life" :-) Cool! I made a hex glider on my first try! |
|
| |
Just a thought in passing, this seems like a perfect application to use optical links ( just a low-powered LED and a phototransistor on each side ) between the elements rather than metallic contacts ...
Also, if there was an optical receiver element on the top surface, they could all be programmed by a powerful emitter connected to a laptop, etc.
( this could also provide the timing signal ).
I like the idea of each unit having a 'mac address'.
Anyway, fantastic idea ! [+] ! |
|
| |
I think this could really work out. |
|
| |
You can get quite alot of MCU processing power for a dollar these days. |
|
| |
A 16-Bit ARM mcu would work well and some even have networking built into the chips themselves. |
|
| |
What about a light-pen for input and configuration? |
|
| |
I'd figure these would have to be made small, keychain friendly too perhaps... |
|
| |
batou: it occured to me that to cost-engineer this, you'ld want to reduce *everything* to a single chip, and have no wiring at all (with the battery on the chip). So your idea of using optical links would fit pretty nicely. Whether a little laser on a chip is cheaper than a LED on a chip, I've no idea. I do know that a laser on a chip is possible. If a decent output LED were also possible then the chip would be self-sufficient, by driving the 'display' LED as well as the communications links. |
|
| |
Anyone remember making EPROMS light up by turning up the juice? :-) (unfortunately along with the light you also got quite a bit of heat...) |
|
| |