by FrediFizzx » Thu Oct 10, 2019 9:58 am
Joy Christian wrote:FrediFizzx wrote:
Anyways, the direction of the spin vector is very non-trivial due to 3-sphere topology. And it looks like the detectors are involved in the topology also. Is that correct, Joy?
The detectors, which are represented by unit bivectors, are indeed involved in the topology, and that looks problematic from the flatland perspective. That is what Albert Jan was saying:
Joy Christian wrote:ajw wrote:
I have expressed earlier my feeling that this is still the flatlanders simulation of the model. It works on the 3 computer setup discussed elsewhere, but only if one takes the number of events to be the amount of particle pairs received at the detectors, not the amount of particle pairs sent to the filters. So one has to ignore a fair amount of single 'clicks' in the result set, because the measurement function on the opposite side has set the result to 'no state'.
I think I know what Albert Jan is saying. But such a "non-flatland" simulation would be very difficult, if not impossible to do. Not that I am an expert in programming. Quite the opposite.
??? It is not impossible to do at all. Complete states is the way to do it. If we know
a,
s,
and
, we can predict with certainty the outcome at station A. And if we know
b,
s,
and
, we can predict with certainty the outcome at station B.
The problem is that the singlet spin vector's direction is highly non-trivial due to the 3-sphere topology. This "bangs" it in. It is definitely wrong in the code to use points on a 2-sphere as a starting point.
Here is a recent paper I found that highlights the problem.
https://arxiv.org/abs/1501.00693"Do Spins Have Directions?"
.
[quote="Joy Christian"][quote="FrediFizzx"]
Anyways, the direction of the spin vector is very non-trivial due to 3-sphere topology. And it looks like the detectors are involved in the topology also. Is that correct, Joy?
[/quote]
The detectors, which are represented by unit bivectors, are indeed involved in the topology, and that looks problematic from the flatland perspective. That is what Albert Jan was saying:
[quote="Joy Christian"][quote="ajw"]
I have expressed earlier my feeling that this is still the flatlanders simulation of the model. It works on the 3 computer setup discussed elsewhere, but only if one takes the number of events to be the amount of particle pairs received at the detectors, not the amount of particle pairs sent to the filters. So one has to ignore a fair amount of single 'clicks' in the result set, because the measurement function on the opposite side has set the result to 'no state'.[/quote]
I think I know what Albert Jan is saying. But such a "non-flatland" simulation would be very difficult, if not impossible to do. Not that I am an expert in programming. Quite the opposite.[/quote][/quote]
??? It is not impossible to do at all. Complete states is the way to do it. If we know [b]a[/b], [b]s[/b], [tex]\eta[/tex] and [tex]\lambda[/tex], we can predict with certainty the outcome at station A. And if we know [b]b[/b], [b]s[/b], [tex]\eta[/tex] and [tex]\lambda[/tex], we can predict with certainty the outcome at station B.
The problem is that the singlet spin vector's direction is highly non-trivial due to the 3-sphere topology. This "bangs" it in. It is definitely wrong in the code to use points on a 2-sphere as a starting point.
Here is a recent paper I found that highlights the problem.
https://arxiv.org/abs/1501.00693
"Do Spins Have Directions?"
.