My question points to the fact that Turing machines are isolated by definition. But what if they can send and receive information from/to other Turing machines? What if they can be "interrupted" at any time by communication with other Turing machines?
I think this problem is different, and perhaps gives different answer to the halting problem. But I don't know.
Is there any work on this? Has this equivalence been proven?
If someone found a single example where a non-computable algorithm (with a single universal Turing machine) becomes computable with a network, could it mean that perhaps a Turing network may solve the halting problem?