[E-Lang] Quantum computing and capabilities

Ralph Hartley hartley@aic.nrl.navy.mil
Tue, 06 Feb 2001 15:48:36 -0500


hal@finney.org wrote:

> Ralph Hartley, <hartley@aic.nrl.navy.mil>, writes:
> 
>> hal@finney.org wrote:
>> 
>>> Also, I believe quantum key exchange algorithms require a round trip
>>> (or maybe two) before the sides can start sending data.  This would be
>>> a big problem at interstellar distances.
>> 
>> No.
>> 
>> Suppose Bob wants to send a private message to Alice.
>> For each bit of his key he does the following:
>> 
>> Obtains an "Alice qbit" (a member of an EPR pair Alice has the other 
>> half of).
>> Selects an one of his own qbits such that the corresponding Bob qbit is 
>> convenient to Alice.
>> Selects a direction.
> 
> 
> How do these qbits get shared, in the case of interstellar communication
> without trusted couriers?  Wouldn't they have to get transmitted?

Yes, of course. The key has to be transmitted even with a trusted 
courier. But trusted couriers must travel at less than the speed of light.

> If Alice sent them to Bob, and he is sending back results that depend
> on them, that is a round trip.

In that sense, any kind of cryptography requires a round trip, be it 
public key, quantum, or one time pads. The quantum protocol requires pre 
distribution of the qbits, but only to within easy reach of the parties. 
You don't need to know in advance who on alpha centuari you want to 
communicate with.

> Or can Bob send all the necessary shared qbits to Alice at the same
> time as his classical information, or slightly earlier?  In that case
> you would avoid a round trip.

If he did that someone could intercept the whole bundle, just like they 
could with a courier or a one time pad.

> It seems like it's risky, though, to send a qbit (or two) in entangled
> states, and then send classical information that tells exactly what basis
> to measure with in order to get reliable information out.  Won't it be
> harder to detect an eavesdropper if she knows what basis to use to avoid
> perturbing the data?

I'm pretty sure that it makes no difference in the protocol I described. 
If I made a mistake, it should be easy to fix. Actually, I think the 
result of the measurement Bob uses to get his key is independent of the 
basis, so he doesn't need to send it, but it doesn't give Eve any 
information either, so it doesn't hurt.

Ralph Hartley