[continued from previous message]   
      
   > kind of built in that the assumption works.  Hmm.   
   >   
   >   
   >> Any halting decisions predicated on such a comparison are   
   >> suspicious. Doign the comparison properly is as hard as deciding   
   >> halting; therefore no halting decider may rely on the answer to "are   
   >> these two functions equivalent". That amounts to begging the question.   
   >>   
   >>> What's left is whether the address comparison issue "is still a   
   >>> problem".  I don't see it as   
   >>> relevant to anything much, but you see it as important in some way.   
   >>> I've explained why I don't   
   >>> consider it relevant, but that doesn't convince you, so I'll leave it   
   >>> at that.   
   >>   
   >> Even if function address comparison is used, it will catch some cases of   
   >> recursive loops where functions are incorrectly concluded to be   
   >> different like:   
   >>   
   >>    f() { g(); }   
   >>    g() { h(); }   
   >>    h() { k(); }   
   >>    k() { g(); } // k is same as f   
   >>   
   >   
   > I'm not clear on what this is meant to show.  I see k and f are   
   > "equivalent" in their functionality.  PO's (unsound) rule would   
   > presumably see f-->g-->h-->k-->g, and g has been called recursively -   
   > but it would NOT match at this point because although the target g has   
   > been called twice, the calls are from different addresses viz somewhere   
   > inside f vs somewhere inside k.   
   >   
   > That's not a problem.  PO does not claim his decider catches all   
   > recursive simulations.   
   >   
      
   It does catch recursive chains. Line 1307   
   https://github.com/plolcott/x86utm/blob/master/Halt7.c   
      
   > Continuing, g-->h, and now PO's rule will match, because h has been   
   > called twice from the same address somewhere in g.  So HHH correctly   
   > decides non-halting.   
   >   
   > All fine, unless you're saying the rule /should/ have matched earlier   
   > when g was called.  That would simply be a different rule.  The decider   
   > here isn't "incorrectly concluding that k and f are different" - it's   
   > simply agnostic on that point.   
   >   
   > That's an argument that a decider /isn't/ required to identify that k is   
   > the same as f.   
   >   
   > BTW, don't you think that this scenario is exactly why PO is so   
   > convinced his "infinite recursion" pattern is sound, simply being unable   
   > to understand the differences between this example (Call) and his HHH/DD   
   > example (Simulation)?   
   >   
   >   
   > Mike.   
   >   
      
   I have always been correct on this and the only   
   "errors" are mere lack of sufficient comprehension   
   by my reviewers.   
      
   --   
   Copyright 2025 Olcott   
      
   My 28 year goal has been to make   
   "true on the basis of meaning" computable.   
      
   This required establishing a new foundation   
   for correct reasoning.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)