Hello, fin_81, you wrote: _> Hello, kochetkov.vladimir, you wrote: KV>> to Explain? I am not simple so I climb, that is why that just she and explains, whence undertake mentioned above an exponent. _> if exhibitors, it already a theorem refutation climb. Certainly is not present) As we cannot analyze something, equivalent to the machine of Turing, we should reduce it something to a computational model of the smaller order. Generally it is possible, through the description of all possible configurations and passages in between in the form of the finite state machine. However, the number of states in it will be infinite for the account of infinity of tape and finite such automatic machine will not be (differently it really would refute the theorem), hence, we should approximate to a type which allows to come to finite (and reasonable) to number of states of the automatic machine describing it. In practice it expresses in the restriction of number of iterations of cycles mentioned above with unknown variables in their invariants, restriction of lengths of lines, arrays, the storage accessible round pointers, etc. And the number of states of the received automatic machine grows on an exhibitor from a polynomial which variables are limits of approximation of the initial program. And, as it is necessary for analyzer to sort out hardly less, than all these states, we and receive EXPTIME its operations. KV>> it in normal languages. And in a case with JS - we neigh, how abnormal and desperately we gesticulate _> Normal language is, that which not tjuring-full? It is possible and so. For example rules in aircraft and astronautics (https://yurichev.com/mirrors/C/JPL_Codi … dard_C.pdf) just also describe how to make the code not so Turing complete, at least in the part. In the general case, normal language is what possesses the abstractions shifting responsibility for approximation of the code from the analyzer on the developer, writing this code. Strict static typification is one of such abstractions. As well as pure functions with lazy calculations and dependent types (it gave the chance births of languages Coq, Idris, Agda, F *, allowing formally to prove at least special cases of properties of programs).