XPost: comp.theory, sci.logic, sci.math   
   From: 643-408-1753@kylheku.com   
      
   On 2025-11-18, dart200  wrote:   
   > On 11/17/25 7:07 PM, Kaz Kylheku wrote:   
   >> On 2025-11-18, dart200  wrote:   
   >>> On 11/17/25 4:31 PM, olcott wrote:   
   >>>> On 11/17/2025 6:06 PM, dart200 wrote:   
   >>>>> On 11/17/25 3:35 PM, olcott wrote:   
   >>>>>> The halting problem is requiring deciders to   
   >>>>>> compute information that is not contained in   
   >>>>>> their input.   
   >>>>>   
   >>>>> ur agreeing with turing and the halting problem:   
   >>>>>   
   >>>>> one cannot compute whether a machine halts or not from the string   
   >>>>> describing the machine   
   >>>>>   
   >>>>   
   >>>> That the halting problem limits computation   
   >>>> is like this very extreme example:   
   >>>>   
   >>>> Predict who the next president of the United States   
   >>>> will be entirely on the basis of √2 (square root of 2).   
   >>>> That cannot be derived from the input.   
   >>>   
   >>> bruh, ur agreeing with the halting problem:   
   >>>   
   >>> one cannot take the string describing the machine, and use it to compute   
   >>> whether the machine described halts   
   >>   
   >> But that isn't true; you certainly can do that. Just not using one   
   >> unified algorithm that works for absolutely all such strings.   
   >>   
   >> When it /does/ work, it's certainly not based on any input other than   
   >> the string.   
   >   
   > yes i meant generally   
      
   But polocott means something else. He keeps insisting (without any   
   rational justification) that the conventional halting problem,   
   when "H" is presented with the diagonal "D" case, is asking   
   "H" to decide something which is not the finite string input.   
      
   He believes that D literally calls the same instance of H in   
   the same program image, which is the only way it can be,   
   and thus D is H's caller. And thus H is being asked to decide   
   about its caller. But the caller is not the parameter D, but   
   an activated procedure. Therefore H is being asked to decide something   
   about an activated procedure and not its finite string parameter.   
      
   The "reasoning" if it can be called that, is completely   
   disconnected from rationality; it's eaxctly like the witch   
   scene in Monty Python and The Holy Grail.   
      
   Witches burn, and wood also burns proving that witches are made of wood;   
   wood floats; a duck also floats so it must be made of wood; so if the   
   woman weighs as much as a duck, she must be witch.   
      
   This is computer science according to olcott:   
      
   1. The standard halting problem stupidly forgets to restrict   
      decider inputs to finite machine restrictions, sometimes   
      requiring them to decide on their callers.   
      
   2. D calls H, and so D is H's caller.   
      
   3. A caller cannot be an input.   
      
   4. But H clearly does have an input D in the expression H(D)   
      and D is its caller.   
      
   5. Since the caller cannot be an input, there must be two D's:   
      the caller D and the input D.   
      
   6. It is the caller D that is nonterminating, and the Halting Problem is   
      wrongly asking about that one, rather than the input.   
      
   7. The input D is nonterminating. (Proof: when H simulates it,   
      it gets into some kind of recursive tizzy that Olcott poorly   
      understands. Anyway, because of that H is correct to call its   
      input nonterminating and return 0.)   
      
   8. Deciders other than H can report 1 because D is not /their/ caller,   
      and so to them, the caller D and input D are the same.   
      (Proof: when olcott makes a an exact copy of H under the name H1,   
      it is found that H1(D) returns 1. The only difference is that   
      D calls H and not H1: D is not H1's caller, and so H1 decides the   
      terminating D as required by the halting problem.)   
      
   Problem is:   
      
   In (1) the halting problem does not forget to restrict decider   
   inputs to finite machine descriptions.   
      
   In (7) the recursion detecting conditions olcott came up with   
   and tested in the x86utm/Halt7 are bogus. They actually detect   
   the emergence of simulation tower, plus have some other issues   
   due to cheating with static, mutable state.   
      
   In (8), the business with H1(D) and H(D) returning a different value has   
   to do with invalid comparison of functions. H1 and H want to be the same   
   function according to the math, but the abort test uses address   
   equivalance to conclude they are not the same function.   
   That test then /makes/ them be different functions.   
   But because they have the same body, that speaks something to Olcott,   
   through is massive confirmation bias; he takes it as evidence that   
   is caller versus input hypothesis is correct.   
      
   --   
   TXR Programming Language: http://nongnu.org/txr   
   Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal   
   Mastodon: @Kazinator@mstdn.ca   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)