local wrote:gill1109 wrote: At least, it further reduces the set of “Joy Christian believers” to a tiny cargo cult.Of course, maybe they are right!
This stuff is why you have zero credibility. You acknowledge that “Joy Christian believers” (already a slur) may be right, but you have no problem dismissing it as a "tiny cargo cult". Shame on you!
But if Joy Christian talks about “Bell believers” that is not a slur? If he talks disparagingly about statisticians, that is not a slur? Don’t feel insulted. Don’t take it personally. Wear the epithet as a badge of honour!
Karl Hess discovered that though there are many people who believe that Bell was wrong, they cannot agree with one another. They each go off on their own track. That’s just the “state of play”. Time for something new. We should take a look at what Tim Palmer and Sabine Hossenfelder are doing. Understand how it is related to what Bell’s work tells us.
But anyway - this thread is about computer science (distributed computing, with classical computers). Gull, and earlier Gill, argued that a certain distributed computing task was doomed to failure. The fastest way to prove us wrong is to come up with a computer simulation of two separated computers, each receiving inputs (setting angles) from two outside experimentalists. Computer A receives an input file (list of setting angles), and generates an output file (list of +/-1’s). Computer B, idem. After the two computers have independently done their work, the output data is analysed. Note: inputs come from outside. Data analysis is done outside.
You may ask: where is the source of the pairs of photons or electrons or whatever? Where are the local hidden variables? My answer is, they can be written into the programs of both computers A and B, at the outset. Suppose for instance you have figured out separate physical mechanisms involving separate, local, random processes at a source, in the transmission lines from source to A and B, at the detector A, and at the detector B. You’ll program these using pseudo random generators. Well: put all the needed pseudo random generators into one program. Fix the parameters of the generators, and the initial seeds of the generators, as constants, in the initial definitions and declarations of the program. You run the same, completely deterministic, program, on both computers A and B. You just flip the outputs of computer B so that with the same list of input setting angles, the two computers produce lists of equal and opposite outcomes, instead of equal outcomes.
If you want to see different outcomes for the same inputs, you write a different RNG seed into the header of the program, and you recompile it.
So to prove Gill and Gull wrong, you don’t even need to do complicated simulations of computer networks. Just write one program, which we will use twice: once to simulate source plus detector A, once to simulate source plus detector B. The second time, we’ll flip the outcomes. We won’t tell you in advance what the two input lists will be. We reserve the right to run the same program many times with varying inputs. We will check that it satisfies the specifications. For each n, n’th output depends only on n and on the n’th input. Same inputs -> same outputs.
Look how complicated it was for Perelman to solve the Poincaré conjecture. It is just a very complex problem to solve but someone will eventually shoot down your theory.