News for hackers. Stuff that matters.
Re:Has James Gates Discovered Computer Code in String Theory... by sp332 on Thursday October 03, 2013 @08:19AM (#6488371)
Nope, and here's why: If you take a valid code and flip one bit, you get an invalid code. But if you take a particle and (somehow) change one attribute, you get a completely valid particle. It's impossible to change just one attribute at a time, but you can wander around the graph and eventually arrive at every point.
Re:Has James Gates Discovered Computer Code in String Theory... by ofthecaribbean on Thursday October 03, 2013 @08:33AM (#6488416)
Like https://en.wikipedia.org/wiki/Lights_Out_%28game%29 ?

But I don't see how your explanation is compatible with their claim that found a neat structure like a "doubly even self-dual linear" ECC. If their claim is that all transitions maintain that structure, then you are guaranteed to stay inside the structure no matter how long you wander around the graph.

Re:Has James Gates Discovered Computer Code in String Theory... by chiph on Thursday October 03, 2013 @08:54AM (#6488512)
I was thinking more like a Drunkard's Walk[1]. Eventually, every possible location (state) will be visited.

[1] http://en.wikipedia.org/wiki/Random_walk

Re:Has James Gates Discovered Computer Code in String Theory... by ofthecaribbean on Thursday October 03, 2013 @09:14AM (#6488628)
If you have an integer random walk and the step sizes are multiples of 3, you cannot reach a location that is not a multiple of 3. The ECC structure they described is similar to the set of multiples of 3: for example, an analog computer trying to simulate such a walk would periodically round the location to the nearest multiple of 3 to counter the effects of noise. The rounding does not change correct computations (becase they would already be multiples of 3), but it increases the likelihood that the result of the computation is correct.
Re:Has James Gates Discovered Computer Code in String Theory... by sp332 on Thursday October 03, 2013 @10:24AM (#6489080)
Right, but here, each half of the graph is a different code. If you screw up a "white dot", you just get a "black dot".