Breaking the Shannon Channel Capacity Limit

On 02/11/2017 10:23 PM, Tim Wescott wrote:
> On Sat, 11 Feb 2017 17:46:49 -0800, yanli0008 wrote:
>
>> 在 2015年7月29日星期三 UTC+8上午5:29:17，ChesterW写道：
>>> These guys claim a new modulation method that exceeds the data
>>> transmission rate set by the Shannon limit. Any opinions? Opinions on
>>> this topic that is, I KNOW you all have considerable opinions on other
>>> issues ;)
>>>
>>> http://www.astrapi-corp.com
>>>
>>> ChesterW
>>
>> Hi all
>>
>> Has anyone read the paper "A Brief Introduction on Shannon's Information
>> Theory" by Ricky Chen, 	arXiv:1612.09316 [cs.IT]? I did not find
> any
>> error in the arguments there. So, mathematically, it seems there is a
>> little chance that Shannon's limit can be broken in certain channels.
>> Any idea?
>
> No, but I have read one of Shannon's papers from the 1940's.  Not only
> did I find it quite understandable, I think he made his point very well.
>
> If there are channels where the Shannon Capacity can be "broken", it's
> more likely that the channel is not one that's covered by the capacity
> theorem than that the theorem itself is faulty.
>
> On the other hand, the claims that I have seen about "breaking" the
> Shannon capacity theorem read more like arguments for perpetual motion
> machines than like serious mathematical treatises.
>

There isn't any "Shannon Limit." You can carry infinite information in a 
finite bandwidth no problem at all; the only tough part is to find a 
channel with -infinity dB of noise. ;-)

The reason that the claims seem like arguments for perpetual motion 
machines is that at the core the noisy channel theorem is isomorphic to 
the second law of thermodynamics. If you could find a way to bust the 
former then the latter is certainly busted as well