• Re: richard the lying chatbot

    From dart200@[email protected] to comp.theory on Sun Nov 2 12:45:08 2025
    From Newsgroup: comp.theory

    On 11/2/25 11:10 AM, Richard Damon wrote:
    On 11/1/25 12:37 PM, dart200 wrote:
    On 11/1/25 6:19 AM, Richard Damon wrote:
    On 11/1/25 2:18 AM, dart200 wrote:
    On 10/31/25 9:33 PM, Richard Damon wrote:
    On 10/31/25 11:52 PM, dart200 wrote:
    On 10/31/25 8:33 PM, Richard Damon wrote:
    On 10/31/25 11:16 PM, dart200 wrote:
    On 10/31/25 7:59 PM, Richard Damon wrote:
    On 10/31/25 3:29 PM, dart200 wrote:
    On 10/31/25 9:18 AM, Richard Damon wrote:
    On 10/30/25 2:27 PM, dart200 wrote:
    On 10/29/25 8:42 PM, olcott wrote:
    On 10/29/2025 10:23 PM, dart200 wrote:
    On 10/29/25 6:45 PM, Richard Damon wrote:
    On 10/29/25 12:24 PM, dart200 wrote:
    On 10/29/25 4:29 AM, Richard Damon wrote:
    On 10/28/25 11:23 PM, dart200 wrote:
    On 10/28/25 6:55 PM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>> On 10/28/25 9:41 PM, dart200 wrote:
    On 10/28/25 6:34 PM, Chris M. Thomasson wrote: >>>>>>>>>>>>>>>>>>>>> On 10/28/2025 5:43 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>> [...]
    not my argument u fucking useless boomer >>>>>>>>>>>>>>>>>>>>>
    Well, how do you solve the halting problem? >>>>>>>>>>>>>>>>>>>>>

    with reflection, something u lack as a useless >>>>>>>>>>>>>>>>>>>> fucking boomer


    Which you can't say how to do, because your only >>>>>>>>>>>>>>>>>>> answer is to define that you can.

    we haven't even gotten the point of working out the >>>>>>>>>>>>>>>>>> ramifications of what i've defined,

    Because you haven't actually defined what you are doing. >>>>>>>>>>>>>>>>
    blatant lie

    Really, show where you have ACTUALLY fully defined your >>>>>>>>>>>>>>> computation machine to a level you can actually analyize >>>>>>>>>>>>>>> one.

    TMs with a modification, specifically adding full >>>>>>>>>>>>>> mechanical reflection to TMs in order to produce
    Reflective TMs (RTMs)

    adding full mechanical reflection just means it gets a >>>>>>>>>>>>>> single instruction that dumps *all* this information to >>>>>>>>>>>>>> the tape, at the head, when that instruction is called: >>>>>>>>>>>>>>
    1) machine description
    2) initial tape state
    3) current instruction
    4) current tape state

    call it REFLECT, if u will... this information allows a >>>>>>>>>>>>>> computation to know exactly where it is in the overall >>>>>>>>>>>>>> runtime of computation, and with that it can subvert >>>>>>>>>>>>>> semantic paradoxes like the halting problem.

    you have to understand this is a *mechanical* modification >>>>>>>>>>>>>> to turing machines, not a computational one. much like how >>>>>>>>>>>>>> RTMs can move a head left/right, or write 0/1 to the >>>>>>>>>>>>>> tape ... they mechanically just do this operation when the >>>>>>>>>>>>>> instruction is called.

    *these mechanics are (relatively) simple and self- >>>>>>>>>>>>>> evidentially possible and therefore mathematically feasible.* >>>>>>>>>>>>>>
    BUT, let me try a step deeper, and quote turing at you >>>>>>>>>>>>>> from his original paper on computable numbers:

    /We may compare a man in the process of computing a real >>>>>>>>>>>>>> number to a
    machine which is only capable of a finite number of >>>>>>>>>>>>>> conditions/ [Tur36]

    THE ENTIRE JUSTIFICATION FOR TURING MACHINES RESTS >>>>>>>>>>>>>> SQUARELY ON THE NOTION OF "CAN A MAN MECHANICALLY >>>>>>>>>>>>>> UNDERTAKE THESE STATE TRANSITIONS"

    1) if a man is running thru the steps of a computation, >>>>>>>>>>>>>> then obviously he has access the full machine description, >>>>>>>>>>>>>> or how else the fuck is he doing any of these state >>>>>>>>>>>>>> transitions??? therefore he obviously can dump the machine >>>>>>>>>>>>>> description, in whatever format he has it in, to the tape >>>>>>>>>>>>>> when REFLECT is encountered.

    2) if a man is running thru the steps of a computation, >>>>>>>>>>>>>> then obviously he can to the side save the initial tape in >>>>>>>>>>>>>> a "buffer" that than cannot be overwritten. with that he >>>>>>>>>>>>>> can can obviously dump the initial tape state to the tape >>>>>>>>>>>>>> whenever REFLECT is encountered.

    3) if a man is running thru the steps of a computation, >>>>>>>>>>>>>> then obviously he has access input of a particular >>>>>>>>>>>>>> transition, and therefore can dump that input to the tape >>>>>>>>>>>>>> when REFLECT is encountered, in whatever format he >>>>>>>>>>>>>> encounters it in.

    4) if a man is running thru the steps of a computation ... >>>>>>>>>>>>>> he can obviously duplicate the tape when REFLECT is >>>>>>>>>>>>>> called. or more precisely he can save it into a buffer, >>>>>>>>>>>>>> and dump from the buffer after putting 1-3 on the tape. >>>>>>>>>>>>>>
    THAT IS THE SAME LEVEL OF FORMAT JUSTIFICATION TURING >>>>>>>>>>>>>> ULTIMATELY GIVES FOR ANY COMPUTING MACHINE. SERIOUSLY READ >>>>>>>>>>>>>> HIS FUCKING ESSAY YOU DORK

    please tell me where u are confused, not how u think i'm >>>>>>>>>>>>>> wrong


    I can finally see the gist of what you are saying
    and it seems to be very similar to my idea.

    like i've said polcott, i consider you having a half-way >>>>>>>>>>>> solution. u have something that's executable for sure, and >>>>>>>>>>>> that is a definite improvement over total undecidability. >>>>>>>>>>>>
    but it's not totally epistemically useful, it still
    doesn't / effectively compute/ the halting function in a >>>>>>>>>>>> *general* manner.

    i'm proposing something that's both executable *and*
    epistemically useful for that purpose.

    the RTM i'm proposing here is just the underlying
    theoretical mechanisms that can be used to produce a
    theoretically coherent context-dependent decider,

    which can then be used to /effectively compute/ the halting >>>>>>>>>>>> function irrespective of semantic paradoxes ...

    something that is in total a step beyond the mere
    executability that you are proposing.


    No, the problem you have is that you make the same error and >>>>>>>>>>> try to redefine the question, and allow your machine to give >>>>>>>>>>> wrong answers.

    it's not a "wrong" answer if it contains no information to be >>>>>>>>>> wrong about, and it's only given in a situation where there's >>>>>>>>>> no "right" answer.

    But their is a right answer for every actual D based on every >>>>>>>>> actual H.

    It just isn't the answer that H gives.


    how is it reasonable to expect a decider to respond truthfully >>>>>>>>>> to a situation that would only make that truth untrue

    Because deciders are just deterministic processes with a fixed >>>>>>>>> algorithm.

    right, and context can be an input to the determinism

    console.trace() isn't nondeterministic, it's context dependent >>>>>>>>

    Note, the system DOES understand that some problems turn out to >>>>>>>>> be uncomputable, so no decider exists that can be correct for >>>>>>>>> all inputs. Nothing wrong with that, it just means that any >>>>>>>>> decider that wants to try to claim to solve it will be wrong >>>>>>>>> for some input.

    The key point here is the question can't be asked until after a >>>>>>>>> decider steps up to make the claim, as the prove-it-wrong input >>>>>>>>> is different for every possible decider, and thus, by the time >>>>>>>>> the question can be asked, the deciders result for that input >>>>>>>>> was already determined by its algorithm. For the decider it >>>>>>>>> isn't a matter of being "truthful", it is just doing what it >>>>>>>>> was programmed to do.

    It becomes a matter of correctness, not truthfulness.

    It is the creator of the algoritm to which truthfulness is
    applied, and claiming that it will give the right answer when >>>>>>>>> it doesn't shows that his claim wasn't truthful.


    there's no "correct" answer to a pathological question of
    halts(und) on this line:

    und = () -> halts(und) && loop_forever()

    Which isn't a computation if your symbology means that halts is >>>>>>> taken from the context that the symbol und is used in.

    Sorry, you are just showing you don't know the meaning of the words. >>>>>>>

    and the resolution is just to give a non-answer, so i split the >>>>>>>> decider into two halves and made false a non-answer. it cannot >>>>>>>> be a contradiction because it doesn't give information /in that >>>>>>>> moment/ about und() to contradict.

    But since you are starting with the same category error of
    Olcott, you get the same nonsense he does.

    The category for the input to a Halt Decider is the
    representation of a Computation.

    By DEFINITION, this has a FULLY defined algorithm, and thus NOT a >>>>>>> closure.

    If you want to try to define your own theory of computation that >>>>>>> uses a broader defintion, go ahead, the problem is you will find >>>>>>> you can do litle with it, and that the semantic paradox you are >>>>>>> complaining about is totally due to you altered definition of a >>>>>>> computation.

    right why do u keep asserting ur existential fuckup is more
    correct than a working decider on the matter???

    Because it doesn't give the right answer per the original problem.

    it gives you the right answer in all the places that the information
    of that answer is actually usable, and it does so generally.

    In other words, your logic says it is ok to be wrong in some cases,
    if you can give a "good excuse" why it is ok to be wrong there.

    All that does is destroy the meaning of "Correct".


    you are holding onto some notion of wanting a "right" answer where
    no right answer can exist, and that's just bizarre because with that
    desire u lose the ability to generally get a "right" answer in first
    place.

    Because it *DOES* have a proper definition.

    ALWAYS means ALWAYS, not just where I want it.

    i can't believe ur a chief engineer asking for correctness where

    a) it has no practical usecase

    b) it ends up neutering ur theoretical ability to generally handle the
    problem

    ur demanding to be able to shoot urself in ur foot over some decider
    that can't even exist.

    weird fetish bro

    I guess you just think that lies are valid, and truth doesn't matter.

    Note, a real halt decider, if it existed, would have a LOT of actually practical uses, and the fact it can't exist is important knowledge.




    ur throwing the fucking baby out with the bathwater dude,

    it's fucking stupid

    Nope, because for actual problems like Busy Beaver, which again asks
    about for results from ALL possible machihes with specific criteria,
    that pattern might just coencidentally occur.



    If you want to be clear that you aren't working on that problem, or >>>>> even working under that system, go ahead, just make sure you fully
    define what you are doing and try to show where it is useful.


    i'm not building a separate form of computing, i'm upgrading what we
    have

    In other words, you are admitting you are lying.

    It is a VIOLATION of the meaning of the words to change the meaning
    of the word as already used.

    You "upgrade" a system by adding new ideas without removing or
    changing any of the old.

    If you need to remove or change something, you have a new system.
    DEFINITION.

    like adding a new operation BY DEFINITION???

    Sure, it it can't be done within the limits of the previously defined system.


    honestly idk if ur just in denial, too senior to take criticism
    anymore, or just shoved a stick too far ur butthole,

    Nope, it is just sticking up for truth against stupid people who don't understand what they are talking about.


    but suggesting i'm not changing anything when i've spent pages trying
    to define a very particular change, even given it's a label RTM is
    just completely disingenuous of you.

    So, how does the machine write out information that it doesn't have stored?


    i'm insulted at the lack of consideration i'm getting here, clearly u
    have no willingness to engaging in genuine discussion, cause ur just
    blatantly lying about my position.

    Then you are being insulted by your own errors.

    YOU are the one who has decided not to engage in answering question put
    to you to explain.


    how can u offer me any meaningful feedback if ur just gunna blatantly
    lie about my position???

    What LIE have I said?

    To show that what I have said can't be done, just needs a showing how
    you claim it actually could be done within the restrictions of the
    framework you are in.

    Part of your problem, is it seems you haven't bothered to learn what the frameworks actually is.



    Again, all you are doing is admitting that you system is based on you
    claiming the right to lie about what you are doing, by ignoring the
    rules of the system.





    yes this does mean maybe /false/ could be returned everywhere, >>>>>>>> but that's not fucking interesting ... nor does it give /
    effective computability/ in regards to the halting question.

    Since your input is a category error for the actual Computation >>>>>>> Theory as classical defined, your results don't apply.

    classic computing theory doesn't have a halt decider defined
    tho ... u only have a halting function, not a decider. the decider >>>>>> is *disproven* and therefore cannot be defined as per consensus
    theory.

    u fucking retard.

    u keep asserting SOMETHING THAT DOESN'T EVEN EXIST is somehow more >>>>>> correct than SOMETHING THAT CAN EXIST

    The DEFINITION of a Halting Decider exists.

    The fact that the definiton is not realizable, proves the theory.

    To call something a Halt Decider that doesn't meet the definition,
    is just a lie, making the person making the claim just a liar.

    just like there's more than one way to ask the same question,
    there's more than one way to define a halting decider, which is the
    interface into an algorithm, not the algorithm itself.

    Right, but if they are semantically different, it isn't the same
    question.


    the one i'm proposing can actually generally exist, unlike the
    classical definition.

    But only is a Computation System that is broken, because programs
    don't reliable produce the same answer when run.

    Once you allow things to lie, you have destroyed the concept of truth.


    it's really that simple.

    Yes, allowing lying seems simple until you need to count on something.



    We can easily DEFINE things that don't exist.>


    which is the point of the context-dependent decider. halts() is >>>>>>>> required to return /true/ only whereever both /possible/ and / >>>>>>>> truthful/ to do so.

    So, if you want to try to talk about this, you first need to try >>>>>>> to develop a basic theory about context dependent Computations, >>>>>>> to see if it can actually be useful for anything.

    i fucking already did. u just aren't reading cause ur too busy
    trying to show off how big a fucking retard u can be.

    No you haven't, unless you consider your claim that is the
    equivalent of defining that 1 == 2 is a definition.

    DESCRIBING what you want to do is NOT the same as DEFINING how to
    do it.

    ur special pleading is just so mindnumbingly boring. i DEFINED my
    mechanics in MORE detail that turing did his, and yet u arbitrarily
    reject it as "describing" which seems like a made up distinction here.

    NO YOU HAVEN'T.

    You have PRESUMED an ability on the machine that can not actually be
    implemented on it without breaking it.

    i'm intelligent enough to write down the machine description to the
    tape when i see the operation REFLECT, and therefore accept the
    defined ability

    But is it the RIGHT machine description for the program in question?

    And where does it get the contents of the original tape?


    i get that ur apparently too fucking retarded to do that, and
    therefore don't accept it.

    Nope, it seems you are too retarded to actually think about what other
    say about your system. You suffer from the fallacy that you assume you
    are right and thus ignore the definitions of what you are talking about.



    Your "Definition" is the equivalent of defining that for this work, I
    will consider 1 to be equal to 2.


    sometimes i really do feel like just buying a gun a shooting myself,
    cause i really am stuck on a planet of fucking idiots.

    y must u endlessly torture me so with ur fucking idiocracy?


    I could say the same to you. Why are you so stuck on insisting on
    things that are not.

    Computations, as defined, can't change their answers for the same
    input. PERIOD.

    To insist they can is just to admit you are just a lunatic.

    right, printing stack traces actually isn't part of computing. every
    time we step thru debugging we're doing theoretically indescribable
    magic, according to richard


    And what does that have to do with what I said?

    And yes, a routine looking at who called it is NOT part of "Computing"
    per computation theory, as it can't matter.

    It seems part of your problem is you don't actually know what you are talking about.

    Remember, Computing isn't directly about "Computers" as we currently
    know them, but is more fundamental.

    It seems you have forgotten that "Computing" as a term existed BEFORE
    the computers you want to talk about.


    blows my mind that u think a computing function that prints a stack
    trace does something which cannot be done on and explained by a turing
    machine

    i'm not sure y u think this mindnumbing gishgallop u shit out is worth a continued response to

    can u explain what my argument even is?

    or are you basically just a fucking chatbot keying off one sentence at a
    time?
    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Tristan Wibberley@[email protected] to comp.theory on Sun Nov 2 20:57:33 2025
    From Newsgroup: comp.theory

    On 02/11/2025 19:10, Richard Damon wrote:
    So, how does the machine write out information that it doesn't have stored?

    A TM (at least typically) has a current machine state from a finite list
    of machine states, a reflecting extension can be defined to have a representation for each that it writes out upon interpreting a symbol at
    the tape head for each machine state (or for a subset of machine states,
    etc) instead of a simple move-left or move-right.

    For clarity I think dart200 is talking about an extended TM, extended to
    have one or more machine-state write-out transitions defined.

    For additional clarity, "machine state" does not normally refer to the
    tape contents nor to the current tape-cell index but the state of the
    machine (like register values, for machines defined to have those). You
    can't include the tape cell index in the machine state because then it
    would have an infinite number of machine states. I understand that would
    not be a TM, strictly.

    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From olcott@[email protected] to comp.theory on Sun Nov 2 14:59:01 2025
    From Newsgroup: comp.theory

    On 11/2/2025 2:45 PM, dart200 wrote:
    On 11/2/25 11:10 AM, Richard Damon wrote:
    On 11/1/25 12:37 PM, dart200 wrote:
    On 11/1/25 6:19 AM, Richard Damon wrote:
    On 11/1/25 2:18 AM, dart200 wrote:
    On 10/31/25 9:33 PM, Richard Damon wrote:
    On 10/31/25 11:52 PM, dart200 wrote:
    On 10/31/25 8:33 PM, Richard Damon wrote:
    On 10/31/25 11:16 PM, dart200 wrote:
    On 10/31/25 7:59 PM, Richard Damon wrote:
    On 10/31/25 3:29 PM, dart200 wrote:
    On 10/31/25 9:18 AM, Richard Damon wrote:
    On 10/30/25 2:27 PM, dart200 wrote:
    On 10/29/25 8:42 PM, olcott wrote:
    On 10/29/2025 10:23 PM, dart200 wrote:
    On 10/29/25 6:45 PM, Richard Damon wrote:
    On 10/29/25 12:24 PM, dart200 wrote:
    On 10/29/25 4:29 AM, Richard Damon wrote:
    On 10/28/25 11:23 PM, dart200 wrote:
    On 10/28/25 6:55 PM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>>> On 10/28/25 9:41 PM, dart200 wrote:
    On 10/28/25 6:34 PM, Chris M. Thomasson wrote: >>>>>>>>>>>>>>>>>>>>>> On 10/28/2025 5:43 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>> [...]
    not my argument u fucking useless boomer >>>>>>>>>>>>>>>>>>>>>>
    Well, how do you solve the halting problem? >>>>>>>>>>>>>>>>>>>>>>

    with reflection, something u lack as a useless >>>>>>>>>>>>>>>>>>>>> fucking boomer


    Which you can't say how to do, because your only >>>>>>>>>>>>>>>>>>>> answer is to define that you can.

    we haven't even gotten the point of working out the >>>>>>>>>>>>>>>>>>> ramifications of what i've defined,

    Because you haven't actually defined what you are doing. >>>>>>>>>>>>>>>>>
    blatant lie

    Really, show where you have ACTUALLY fully defined your >>>>>>>>>>>>>>>> computation machine to a level you can actually analyize >>>>>>>>>>>>>>>> one.

    TMs with a modification, specifically adding full >>>>>>>>>>>>>>> mechanical reflection to TMs in order to produce >>>>>>>>>>>>>>> Reflective TMs (RTMs)

    adding full mechanical reflection just means it gets a >>>>>>>>>>>>>>> single instruction that dumps *all* this information to >>>>>>>>>>>>>>> the tape, at the head, when that instruction is called: >>>>>>>>>>>>>>>
    1) machine description
    2) initial tape state
    3) current instruction
    4) current tape state

    call it REFLECT, if u will... this information allows a >>>>>>>>>>>>>>> computation to know exactly where it is in the overall >>>>>>>>>>>>>>> runtime of computation, and with that it can subvert >>>>>>>>>>>>>>> semantic paradoxes like the halting problem.

    you have to understand this is a *mechanical*
    modification to turing machines, not a computational one. >>>>>>>>>>>>>>> much like how RTMs can move a head left/right, or write >>>>>>>>>>>>>>> 0/1 to the tape ... they mechanically just do this >>>>>>>>>>>>>>> operation when the instruction is called.

    *these mechanics are (relatively) simple and self- >>>>>>>>>>>>>>> evidentially possible and therefore mathematically >>>>>>>>>>>>>>> feasible.*

    BUT, let me try a step deeper, and quote turing at you >>>>>>>>>>>>>>> from his original paper on computable numbers:

    /We may compare a man in the process of computing a real >>>>>>>>>>>>>>> number to a
    machine which is only capable of a finite number of >>>>>>>>>>>>>>> conditions/ [Tur36]

    THE ENTIRE JUSTIFICATION FOR TURING MACHINES RESTS >>>>>>>>>>>>>>> SQUARELY ON THE NOTION OF "CAN A MAN MECHANICALLY >>>>>>>>>>>>>>> UNDERTAKE THESE STATE TRANSITIONS"

    1) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>> then obviously he has access the full machine
    description, or how else the fuck is he doing any of >>>>>>>>>>>>>>> these state transitions??? therefore he obviously can >>>>>>>>>>>>>>> dump the machine description, in whatever format he has >>>>>>>>>>>>>>> it in, to the tape when REFLECT is encountered.

    2) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>> then obviously he can to the side save the initial tape >>>>>>>>>>>>>>> in a "buffer" that than cannot be overwritten. with that >>>>>>>>>>>>>>> he can can obviously dump the initial tape state to the >>>>>>>>>>>>>>> tape whenever REFLECT is encountered.

    3) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>> then obviously he has access input of a particular >>>>>>>>>>>>>>> transition, and therefore can dump that input to the tape >>>>>>>>>>>>>>> when REFLECT is encountered, in whatever format he >>>>>>>>>>>>>>> encounters it in.

    4) if a man is running thru the steps of a
    computation ... he can obviously duplicate the tape when >>>>>>>>>>>>>>> REFLECT is called. or more precisely he can save it into >>>>>>>>>>>>>>> a buffer, and dump from the buffer after putting 1-3 on >>>>>>>>>>>>>>> the tape.

    THAT IS THE SAME LEVEL OF FORMAT JUSTIFICATION TURING >>>>>>>>>>>>>>> ULTIMATELY GIVES FOR ANY COMPUTING MACHINE. SERIOUSLY >>>>>>>>>>>>>>> READ HIS FUCKING ESSAY YOU DORK

    please tell me where u are confused, not how u think i'm >>>>>>>>>>>>>>> wrong


    I can finally see the gist of what you are saying
    and it seems to be very similar to my idea.

    like i've said polcott, i consider you having a half-way >>>>>>>>>>>>> solution. u have something that's executable for sure, and >>>>>>>>>>>>> that is a definite improvement over total undecidability. >>>>>>>>>>>>>
    but it's not totally epistemically useful, it still >>>>>>>>>>>>> doesn't / effectively compute/ the halting function in a >>>>>>>>>>>>> *general* manner.

    i'm proposing something that's both executable *and* >>>>>>>>>>>>> epistemically useful for that purpose.

    the RTM i'm proposing here is just the underlying
    theoretical mechanisms that can be used to produce a >>>>>>>>>>>>> theoretically coherent context-dependent decider,

    which can then be used to /effectively compute/ the halting >>>>>>>>>>>>> function irrespective of semantic paradoxes ...

    something that is in total a step beyond the mere
    executability that you are proposing.


    No, the problem you have is that you make the same error and >>>>>>>>>>>> try to redefine the question, and allow your machine to give >>>>>>>>>>>> wrong answers.

    it's not a "wrong" answer if it contains no information to be >>>>>>>>>>> wrong about, and it's only given in a situation where there's >>>>>>>>>>> no "right" answer.

    But their is a right answer for every actual D based on every >>>>>>>>>> actual H.

    It just isn't the answer that H gives.


    how is it reasonable to expect a decider to respond
    truthfully to a situation that would only make that truth untrue >>>>>>>>>>
    Because deciders are just deterministic processes with a fixed >>>>>>>>>> algorithm.

    right, and context can be an input to the determinism

    console.trace() isn't nondeterministic, it's context dependent >>>>>>>>>

    Note, the system DOES understand that some problems turn out >>>>>>>>>> to be uncomputable, so no decider exists that can be correct >>>>>>>>>> for all inputs. Nothing wrong with that, it just means that >>>>>>>>>> any decider that wants to try to claim to solve it will be >>>>>>>>>> wrong for some input.

    The key point here is the question can't be asked until after >>>>>>>>>> a decider steps up to make the claim, as the prove-it-wrong >>>>>>>>>> input is different for every possible decider, and thus, by >>>>>>>>>> the time the question can be asked, the deciders result for >>>>>>>>>> that input was already determined by its algorithm. For the >>>>>>>>>> decider it isn't a matter of being "truthful", it is just >>>>>>>>>> doing what it was programmed to do.

    It becomes a matter of correctness, not truthfulness.

    It is the creator of the algoritm to which truthfulness is >>>>>>>>>> applied, and claiming that it will give the right answer when >>>>>>>>>> it doesn't shows that his claim wasn't truthful.


    there's no "correct" answer to a pathological question of
    halts(und) on this line:

    und = () -> halts(und) && loop_forever()

    Which isn't a computation if your symbology means that halts is >>>>>>>> taken from the context that the symbol und is used in.

    Sorry, you are just showing you don't know the meaning of the >>>>>>>> words.


    and the resolution is just to give a non-answer, so i split the >>>>>>>>> decider into two halves and made false a non-answer. it cannot >>>>>>>>> be a contradiction because it doesn't give information /in that >>>>>>>>> moment/ about und() to contradict.

    But since you are starting with the same category error of
    Olcott, you get the same nonsense he does.

    The category for the input to a Halt Decider is the
    representation of a Computation.

    By DEFINITION, this has a FULLY defined algorithm, and thus NOT >>>>>>>> a closure.

    If you want to try to define your own theory of computation that >>>>>>>> uses a broader defintion, go ahead, the problem is you will find >>>>>>>> you can do litle with it, and that the semantic paradox you are >>>>>>>> complaining about is totally due to you altered definition of a >>>>>>>> computation.

    right why do u keep asserting ur existential fuckup is more
    correct than a working decider on the matter???

    Because it doesn't give the right answer per the original problem.

    it gives you the right answer in all the places that the
    information of that answer is actually usable, and it does so
    generally.

    In other words, your logic says it is ok to be wrong in some cases,
    if you can give a "good excuse" why it is ok to be wrong there.

    All that does is destroy the meaning of "Correct".


    you are holding onto some notion of wanting a "right" answer where
    no right answer can exist, and that's just bizarre because with
    that desire u lose the ability to generally get a "right" answer in >>>>> first place.

    Because it *DOES* have a proper definition.

    ALWAYS means ALWAYS, not just where I want it.

    i can't believe ur a chief engineer asking for correctness where

    a) it has no practical usecase

    b) it ends up neutering ur theoretical ability to generally handle
    the problem

    ur demanding to be able to shoot urself in ur foot over some decider
    that can't even exist.

    weird fetish bro

    I guess you just think that lies are valid, and truth doesn't matter.

    Note, a real halt decider, if it existed, would have a LOT of actually
    practical uses, and the fact it can't exist is important knowledge.




    ur throwing the fucking baby out with the bathwater dude,

    it's fucking stupid

    Nope, because for actual problems like Busy Beaver, which again asks
    about for results from ALL possible machihes with specific criteria,
    that pattern might just coencidentally occur.



    If you want to be clear that you aren't working on that problem,
    or even working under that system, go ahead, just make sure you
    fully define what you are doing and try to show where it is useful. >>>>>>

    i'm not building a separate form of computing, i'm upgrading what
    we have

    In other words, you are admitting you are lying.

    It is a VIOLATION of the meaning of the words to change the meaning
    of the word as already used.

    You "upgrade" a system by adding new ideas without removing or
    changing any of the old.

    If you need to remove or change something, you have a new system.
    DEFINITION.

    like adding a new operation BY DEFINITION???

    Sure, it it can't be done within the limits of the previously defined
    system.


    honestly idk if ur just in denial, too senior to take criticism
    anymore, or just shoved a stick too far ur butthole,

    Nope, it is just sticking up for truth against stupid people who don't
    understand what they are talking about.


    but suggesting i'm not changing anything when i've spent pages trying
    to define a very particular change, even given it's a label RTM is
    just completely disingenuous of you.

    So, how does the machine write out information that it doesn't have
    stored?


    i'm insulted at the lack of consideration i'm getting here, clearly u
    have no willingness to engaging in genuine discussion, cause ur just
    blatantly lying about my position.

    Then you are being insulted by your own errors.

    YOU are the one who has decided not to engage in answering question
    put to you to explain.


    how can u offer me any meaningful feedback if ur just gunna blatantly
    lie about my position???

    What LIE have I said?

    To show that what I have said can't be done, just needs a showing how
    you claim it actually could be done within the restrictions of the
    framework you are in.

    Part of your problem, is it seems you haven't bothered to learn what
    the frameworks actually is.



    Again, all you are doing is admitting that you system is based on
    you claiming the right to lie about what you are doing, by ignoring
    the rules of the system.





    yes this does mean maybe /false/ could be returned everywhere, >>>>>>>>> but that's not fucking interesting ... nor does it give /
    effective computability/ in regards to the halting question.

    Since your input is a category error for the actual Computation >>>>>>>> Theory as classical defined, your results don't apply.

    classic computing theory doesn't have a halt decider defined
    tho ... u only have a halting function, not a decider. the
    decider is *disproven* and therefore cannot be defined as per
    consensus theory.

    u fucking retard.

    u keep asserting SOMETHING THAT DOESN'T EVEN EXIST is somehow
    more correct than SOMETHING THAT CAN EXIST

    The DEFINITION of a Halting Decider exists.

    The fact that the definiton is not realizable, proves the theory.

    To call something a Halt Decider that doesn't meet the definition, >>>>>> is just a lie, making the person making the claim just a liar.

    just like there's more than one way to ask the same question,
    there's more than one way to define a halting decider, which is the >>>>> interface into an algorithm, not the algorithm itself.

    Right, but if they are semantically different, it isn't the same
    question.


    the one i'm proposing can actually generally exist, unlike the
    classical definition.

    But only is a Computation System that is broken, because programs
    don't reliable produce the same answer when run.

    Once you allow things to lie, you have destroyed the concept of truth. >>>>

    it's really that simple.

    Yes, allowing lying seems simple until you need to count on something. >>>>


    We can easily DEFINE things that don't exist.>


    which is the point of the context-dependent decider. halts() is >>>>>>>>> required to return /true/ only whereever both /possible/ and / >>>>>>>>> truthful/ to do so.

    So, if you want to try to talk about this, you first need to try >>>>>>>> to develop a basic theory about context dependent Computations, >>>>>>>> to see if it can actually be useful for anything.

    i fucking already did. u just aren't reading cause ur too busy
    trying to show off how big a fucking retard u can be.

    No you haven't, unless you consider your claim that is the
    equivalent of defining that 1 == 2 is a definition.

    DESCRIBING what you want to do is NOT the same as DEFINING how to >>>>>> do it.

    ur special pleading is just so mindnumbingly boring. i DEFINED my
    mechanics in MORE detail that turing did his, and yet u arbitrarily >>>>> reject it as "describing" which seems like a made up distinction here. >>>>
    NO YOU HAVEN'T.

    You have PRESUMED an ability on the machine that can not actually be
    implemented on it without breaking it.

    i'm intelligent enough to write down the machine description to the
    tape when i see the operation REFLECT, and therefore accept the
    defined ability

    But is it the RIGHT machine description for the program in question?

    And where does it get the contents of the original tape?


    i get that ur apparently too fucking retarded to do that, and
    therefore don't accept it.

    Nope, it seems you are too retarded to actually think about what other
    say about your system. You suffer from the fallacy that you assume you
    are right and thus ignore the definitions of what you are talking about.



    Your "Definition" is the equivalent of defining that for this work,
    I will consider 1 to be equal to 2.


    sometimes i really do feel like just buying a gun a shooting
    myself, cause i really am stuck on a planet of fucking idiots.

    y must u endlessly torture me so with ur fucking idiocracy?


    I could say the same to you. Why are you so stuck on insisting on
    things that are not.

    Computations, as defined, can't change their answers for the same
    input. PERIOD.

    To insist they can is just to admit you are just a lunatic.

    right, printing stack traces actually isn't part of computing. every
    time we step thru debugging we're doing theoretically indescribable
    magic, according to richard


    And what does that have to do with what I said?

    And yes, a routine looking at who called it is NOT part of "Computing"
    per computation theory, as it can't matter.

    It seems part of your problem is you don't actually know what you are
    talking about.

    Remember, Computing isn't directly about "Computers" as we currently
    know them, but is more fundamental.

    It seems you have forgotten that "Computing" as a term existed BEFORE
    the computers you want to talk about.


    blows my mind that u think a computing function that prints a stack
    trace does something which cannot be done on and explained by a turing machine

    i'm not sure y u think this mindnumbing gishgallop u shit out is worth a continued response to

    can u explain what my argument even is?

    or are you basically just a fucking chatbot keying off one sentence at a time?


    Richard is an intelligent reasonably knowledgeable
    chatbot stuck in rebuttal mode with a mean streak.

    If he would utterly give up his mean streak and
    break out of rebuttal mode he might be a pretty good
    reviewer.

    Claude AI has proven a depth of understanding much
    more than ten-fold greater than the best reviewer
    here. With no ego to defend Claude does honest
    dialogue much much better than anyone here.
    --
    Copyright 2025 Olcott "Talent hits a target no one else can hit; Genius
    hits a target no one else can see." Arthur Schopenhauer
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From olcott@[email protected] to comp.theory on Sun Nov 2 15:00:37 2025
    From Newsgroup: comp.theory

    On 11/2/2025 2:57 PM, Tristan Wibberley wrote:
    On 02/11/2025 19:10, Richard Damon wrote:
    So, how does the machine write out information that it doesn't have stored?

    A TM (at least typically) has a current machine state from a finite list
    of machine states, a reflecting extension can be defined to have a representation for each that it writes out upon interpreting a symbol at
    the tape head for each machine state (or for a subset of machine states,
    etc) instead of a simple move-left or move-right.

    For clarity I think dart200 is talking about an extended TM, extended to
    have one or more machine-state write-out transitions defined.

    For additional clarity, "machine state" does not normally refer to the
    tape contents nor to the current tape-cell index but the state of the
    machine (like register values, for machines defined to have those). You
    can't include the tape cell index in the machine state because then it
    would have an infinite number of machine states. I understand that would
    not be a TM, strictly.

    --
    Tristan Wibberley


    That seems to show a pretty good depth of understanding.

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --
    Copyright 2025 Olcott "Talent hits a target no one else can hit; Genius
    hits a target no one else can see." Arthur Schopenhauer
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Kaz Kylheku@[email protected] to comp.theory on Sun Nov 2 22:35:51 2025
    From Newsgroup: comp.theory

    On 2025-11-02, olcott <[email protected]> wrote:
    On 11/2/2025 2:57 PM, Tristan Wibberley wrote:
    [ ... whatever ... ]
    That seems to show a pretty good depth of understanding.

    I'm picturing a bunch of flat earthers sitting around a table,
    acknowledging the new guy.

    "Yes Dave, you got it! Ships fall off the earth when they go over the
    horizon. Your depth of understanding is superb!"
    --
    TXR Programming Language: http://nongnu.org/txr
    Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
    Mastodon: @[email protected]
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Sun Nov 2 17:05:44 2025
    From Newsgroup: comp.theory

    On 11/1/25 1:26 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 11/1/25 9:37 AM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 10/31/25 9:20 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    there's no "correct" answer to a pathological question of halts(und) on >>>>>> this line:

    und = () -> halts(und) && loop_forever()

    Yes there is; the correct answer is ¬halts(und). The logical negation >>>>> of whatever halts(und) decides.

    the correct answer is to give the opposite of it's own answer???

    The correct answer isn't /to give/ anything; it just /is/

    halts() is a total decider, ur now just ignoring what i defined()

    I'm obviously aware of how you defined it.

    If you initially impose the definition "halts is the name of a total decider", then you arrive at a contradiction: halts cannot be
    deciding und, as required of a total decider.

    So then, the definition is a kind of proposition, which is shown false;
    it does not hold.

    But what is the logical opposite of the proposition "halts is the name
    of a total decider?" It is not: "halts does not exist", but simply:
    "halts is /not/ the name of a total decider".

    A halts can exist just fine, just not as a total decider.

    Stop insisting that halts is a the total halting decider which cannot
    exist, allowing it to be a partial halting decider which cannot be
    total, and the apparent paradox goes away. und then exists and has a
    defnite halting status, which is opposite to whatever that partial
    decider says.

    great now ur arguing polcott is right.

    Not really; he has never agreed with the above!

    he might not agree with the wording, but it is kinda what he's proposing


    the problem with a set of partial deciders is that the resolution isn't
    /effectively computable/. with "partial" deciders then the question
    becomes which partial decider is correct: u've just kicked the can down
    the road

    Yes; because no combination of partial deciders gives you a total
    decider.

    We can trivially form a set of partial deciders which has complete
    coverage:

    H_0 := (P) -> 0

    H_1 := (P) -> 1

    H_0 correctly decides all non-terminating inputs. H_1 correctly
    decides all terminating inputs. So one of these two is always
    correct! (Notice how I'm not saying anything more informative
    than "a machine always halts or does not halt").

    We don't know which decider is the right one for which input.

    We need a third decider H whch decides whether to use H_0
    or H_1.

    This example may look silly: because what is H, but simply
    a decider which has to determine whether to returh 0 or 1
    (which are only trivially disguised as H_0(P) and H_1(P))?

    But this idea generalizes. No matter what kind of partial deciders we
    have, or how many, deciding which one is correct for which input is equivalent to deciding halting.

    The result of combining partial deciders will always be a decider
    that can be shown not to be total.

    The can is indeed always kicked down the road.

    great! so ur own answered question of why i'm not particularly excited
    by partial deciders, and want a way to /effectively compute/ whether an
    input machine halts not

    do we need to have this discussion again kaz, or???


    yes this does mean maybe /false/ could be returned everywhere, but >>>>>> that's not fucking interesting ... nor does it give /effective
    computability/ in regards to the halting question.

    which is the point of the context-dependent decider. halts() is required >>>>>> to return /true/ only whereever both /possible/ and /truthful/ to do so. >>>>>>
    if u don't agree with this, that's fine...

    but you actually need to construct a contradiction, you can't just >>>>>> continually make bare assertions that it's a contradiction.

    If contexts are represenable in a Turing substrate, then everything
    succumbs to Turing incomputability of halting.

    If contexts are not representable in Turing, then the
    model-with-contexts is either effectievely computable non-Turing, or >>>>> else contexts are just imaginary, like magic oracles.

    false dichotomy

    Well, trichotomy, by my count.

    - contexts are Turing (1)
    - contexts are not Turing:
    - somehow effectively computable (2)
    - not computable (3)

    What possibility have I unintentionally left out?

    like i'm even open to being wrong, but like u need to work with the
    system in order to show a contradiction with the system itself, just not >>>> assume the contradiction will happen.

    If they are imaginary, you only have a thought experiment.
    (In which Turing-plus-contexts machines can decide Turing
    halting, but so what?)

    REFLECT is completely mechanically demonstrable operation with a pen and >>>> paper, i demonstrated it in a post here already, justifying it's
    theoretical existence

    OK, that points to it being embeddable in Turing
    computation. Which then points to undecidabiity.

    or ur just assuming a false dichotomy

    If your mechanical pencil-and-paper calculations are shown to be Turing, anything rendered using those calculations is Turing. It is not
    a dichotomy, true or false.

    what i'm proposing obviously will look very much like TMs in many if not
    most computatation. in fact: any computation can run on a TM will in
    fact be run the same on an RTM.

    the differences in RTMs only come up in computations that aren't
    directly expressible on TMs


    the TM limitation is mechanical, not computational. classical theory
    only supposes about computational limitations, there's no notion of
    mechanical limitation.

    Are you talking about analog computation with continuous quantities,
    and such?

    no


    So your pencil-and-paper calculations are actually perhaps geometric,
    like compass-and-ruler derivations?

    r u intentionally trying to be obtuse?


    but the mechanical, not computational, limitation means:

    - TMs can simulate RTMs
    - RTMs don't fix the problems with TMs
    - it doesn't matter that they don't because RTMs can compute
    everything TMs do + have increased power

    If TMs can simulate RTMs then they do not have increased power.

    no. TMs have the computational power to handle what RTMs compute ... but
    the lack of the REFLECT operation mechanically limits them from
    dependable access to the required runtime information to actually be
    RTMs directly.

    this relationship just can't be pigeonholed into what u'd like, you are overgeneralizing

    this doesn't solve TM limitation because again the problem is TMs lack mechanical access to not be limited. RTMs instead *trivialize* the TM
    halting problem because RTMs can express everything TMs + overcome
    attempts at constructing semantic paradoxes with RTMs.


    They may have increased "expressive power", which is something only
    relevant to how "nice" they are to write applications in.

    A different Turing model can also have better efficiency; i.e the best
    way of solving a problem in one kind of machine may have exponential
    time whereas another can express a solution that is polynomial.

    Since

    - ugly code that correctly solves a problem halts,
    - inefficient code that correctly solves a problem halts,

    neither expressive power nor efficiency are relevant in halting.

    what i'm doing is trivializing TMs undecidability, not resolving it. but
    this overall resolves the halting problem by providing a mechanical
    improvement that can allow theory to escape limitations stemming from
    the liar's paradox being encoded into machines

    (If you want to talk Olcott, /that/'s a great narrative:
    "halting problem implies liar paradox.")


    REFLECT gives a semantic decider the ability to know it's calling
    context, which gives it the ability to return a context dependent answer. >>>>
    If you find some computational model with contexts which is effective, >>>>> but beyond Turing, then /that/ is the big deal; delegate the halting >>>>> cruft to others. You are important because you've done what nobody else >>>>> has done before: you found an effective method of computation that is >>>>> not Turing! So you go to conference sall over the world where you are >>>>> flocked by groupies. Fuck halting, at that point, you know?

    it's a mild extension of TMs,

    If it is mechanically demonstrable with pen and paper, then it isn't
    an extension of TM.

    no idea what u mean here, TMs are likewise also demonstrable with a pen
    and paper, and that was turing's original justification for their
    feasibility

    (What I mean that it's just TM, and not an extension.)

    Functions can depend on their context. That context can make some
    instances of them halting and others not, even with all the explicit
    arguments being the same. Thus we cannot ask whether a function halts;
    it has to be, does it halt /in what calling context/.

    only if it involves a REFLECT command, otherwise it's not possible to be
    dependent on context.

    So if only deciders are allowed to use REFLECT, but the inputs
    are not allowed to use REFLECT, then indeed it is impossible for an
    input to contradict a decider which uses REFLECT.

    But then deciders are not deciding everything; they are deciding
    only those inputs that don't use REFLECT.

    inputs are allowed to use REFLECT. this doesn't make it possible to
    construct a halting paradox because it does not stop deciders from using /false/ to escape any demonstrable paradox.
    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Chris M. Thomasson@[email protected] to comp.theory on Sun Nov 2 17:39:20 2025
    From Newsgroup: comp.theory

    On 11/2/2025 5:05 PM, dart200 wrote:
    On 11/1/25 1:26 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 11/1/25 9:37 AM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 10/31/25 9:20 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    there's no "correct" answer to a pathological question of
    halts(und) on
    this line:

    und = () -> halts(und) && loop_forever()

    Yes there is; the correct answer is ¬halts(und). The logical negation >>>>>> of whatever halts(und) decides.

    the correct answer is to give the opposite of it's own answer???

    The correct answer isn't /to give/ anything; it just /is/

    halts() is a total decider, ur now just ignoring what i defined()

    I'm obviously aware of how you defined it.

    If you initially impose the definition "halts is the name of a total
    decider", then you arrive at a contradiction: halts cannot be
    deciding und, as required of a total decider.

    So then, the definition is a kind of proposition, which is shown false;
    it does not hold.

    But what is the logical opposite of the proposition "halts is the name
    of a total decider?" It is not: "halts does not exist", but simply:
    "halts is /not/ the name of a total decider".

    A halts can exist just fine, just not as a total decider.

    Stop insisting that halts is a the total halting decider which cannot
    exist, allowing it to be a partial halting decider which cannot be
    total, and the apparent paradox goes away.  und then exists and has a >>>> defnite halting status, which is opposite to whatever that partial
    decider says.

    great now ur arguing polcott is right.

    Not really; he has never agreed with the above!

    he might not agree with the wording, but it is kinda what he's proposing


    the problem with a set of partial deciders is that the resolution isn't
    /effectively computable/. with "partial" deciders then the question
    becomes which partial decider is correct: u've just kicked the can down
    the road

    Yes; because no combination of partial deciders gives you a total
    decider.

    We can trivially form a set of partial deciders which has complete
    coverage:

       H_0 := (P) -> 0

       H_1 := (P) -> 1

    H_0 correctly decides all non-terminating inputs. H_1 correctly
    decides all terminating inputs. So one of these two is always
    correct! (Notice how I'm not saying anything more informative
    than "a machine always halts or does not halt").

    We don't know which decider is the right one for which input.

    We need a third decider H whch decides whether to use H_0
    or H_1.

    This example may look silly: because what is H, but simply
    a decider which has to determine whether to returh 0 or 1
    (which are only trivially disguised as H_0(P) and H_1(P))?

    But this idea generalizes. No matter what kind of partial deciders we
    have, or how many, deciding which one is correct for which input is
    equivalent to deciding halting.

    The result of combining partial deciders will always be a decider
    that can be shown not to be total.

    The can is indeed always kicked down the road.

    great! so ur own answered question of why i'm not particularly excited
    by partial deciders, and want a way to /effectively compute/ whether an input machine halts not

    do we need to have this discussion again kaz, or???


    yes this does mean maybe /false/ could be returned everywhere, but >>>>>>> that's not fucking interesting ... nor does it give /effective
    computability/ in regards to the halting question.

    which is the point of the context-dependent decider. halts() is >>>>>>> required
    to return /true/ only whereever both /possible/ and /truthful/ to >>>>>>> do so.

    if u don't agree with this, that's fine...

    but you actually need to construct a contradiction, you can't just >>>>>>> continually make bare assertions that it's a contradiction.

    If contexts are represenable in a Turing substrate, then everything >>>>>> succumbs to Turing incomputability of halting.

    If contexts are not representable in Turing, then the
    model-with-contexts is either effectievely computable non-Turing, or >>>>>> else contexts are just imaginary, like magic oracles.

    false dichotomy

    Well, trichotomy, by my count.

    - contexts are Turing (1)
    - contexts are not Turing:
        - somehow effectively computable (2)
        - not computable (3)

    What possibility have I unintentionally left out?

    like i'm even open to being wrong, but like u need to work with the
    system in order to show a contradiction with the system itself,
    just not
    assume the contradiction will happen.

    If they are imaginary, you only have a thought experiment.
    (In which Turing-plus-contexts machines can decide Turing
    halting, but so what?)

    REFLECT is completely mechanically demonstrable operation with a
    pen and
    paper, i demonstrated it in a post here already, justifying it's
    theoretical existence

    OK, that points to it being embeddable in Turing
    computation. Which then points to undecidabiity.

    or ur just assuming a false dichotomy

    If your mechanical pencil-and-paper calculations are shown to be Turing,
    anything rendered using those calculations is Turing. It is not
    a dichotomy, true or false.

    what i'm proposing obviously will look very much like TMs in many if not most computatation. in fact: any computation can run on a TM will in
    fact be run the same on an RTM.

    the differences in RTMs only come up in computations that aren't
    directly expressible on TMs


    the TM limitation is mechanical, not computational. classical theory
    only supposes about computational limitations, there's no notion of
    mechanical limitation.

    Are you talking about analog computation with continuous quantities,
    and such?

    no


    So your pencil-and-paper calculations are actually perhaps geometric,
    like compass-and-ruler derivations?

    r u intentionally trying to be obtuse?


    but the mechanical, not computational, limitation means:

    - TMs can simulate RTMs
    - RTMs don't fix the problems with TMs
    - it doesn't matter that they don't because RTMs can compute
        everything TMs do + have increased power

    If TMs can simulate RTMs then they do not have increased power.

    no. TMs have the computational power to handle what RTMs compute ... but
    the lack of the REFLECT operation mechanically limits them from
    dependable access to the required runtime information to actually be
    RTMs directly.

    this relationship just can't be pigeonholed into what u'd like, you are overgeneralizing

    this doesn't solve TM limitation because again the problem is TMs lack mechanical access to not be limited. RTMs instead *trivialize* the TM halting problem because RTMs can express everything TMs + overcome
    attempts at constructing semantic paradoxes with RTMs.


    They may have increased "expressive power", which is something only
    relevant to how "nice" they are to write applications in.

    A different Turing model can also have better efficiency; i.e the best
    way of solving a problem in one kind of machine may have exponential
    time whereas another can express a solution that is polynomial.

    Since

    - ugly code that correctly solves a problem halts,
    - inefficient code that correctly solves a problem halts,

    neither expressive power nor efficiency are relevant in halting.

    what i'm doing is trivializing TMs undecidability, not resolving it. but >>> this overall resolves the halting problem by providing a mechanical
    improvement that can allow theory to escape limitations stemming from
    the liar's paradox being encoded into machines

    (If you want to talk Olcott, /that/'s a great narrative:
    "halting problem implies liar paradox.")


    REFLECT gives a semantic decider the ability to know it's calling
    context, which gives it the ability to return a context dependent
    answer.

    If you find some computational model with contexts which is
    effective,
    but beyond Turing, then /that/ is the big deal; delegate the halting >>>>>> cruft to others. You are important because you've done what nobody >>>>>> else
    has done before: you found an effective method of computation that is >>>>>> not Turing! So you go to conference sall over the world where you are >>>>>> flocked by groupies. Fuck halting, at that point, you know?

    it's a mild extension of TMs,

    If it is mechanically demonstrable with pen and paper, then it isn't
    an extension of TM.

    no idea what u mean here, TMs are likewise also demonstrable with a pen
    and paper, and that was turing's original justification for their
    feasibility

    (What I mean that it's just TM, and not an extension.)

    Functions can depend on their context. That context can make some
    instances of them halting and others not, even with all the explicit
    arguments being the same. Thus we cannot ask whether a function halts; >>>> it has to be, does it halt /in what calling context/.

    only if it involves a REFLECT command, otherwise it's not possible to be >>> dependent on context.

    So if only deciders are allowed to use REFLECT, but the inputs
    are not allowed to use REFLECT, then indeed it is impossible for an
    input to contradict a decider which uses REFLECT.

    But then deciders are not deciding everything; they are deciding
    only those inputs that don't use REFLECT.

    inputs are allowed to use REFLECT. this doesn't make it possible to construct a halting paradox because it does not stop deciders from
    using /false/ to escape any demonstrable paradox.



    Code up a simulation of your concept. You can do it!
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Sun Nov 2 19:25:43 2025
    From Newsgroup: comp.theory

    On 11/2/25 2:35 PM, Kaz Kylheku wrote:
    On 2025-11-02, olcott <[email protected]> wrote:
    On 11/2/2025 2:57 PM, Tristan Wibberley wrote:
    [ ... whatever ... ]
    That seems to show a pretty good depth of understanding.

    I'm picturing a bunch of flat earthers sitting around a table,
    acknowledging the new guy.

    "Yes Dave, you got it! Ships fall off the earth when they go over the horizon. Your depth of understanding is superb!"


    skeptics and flat-earther analogies: name a more iconic duo
    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Sun Nov 2 19:29:45 2025
    From Newsgroup: comp.theory

    On 11/2/25 12:57 PM, Tristan Wibberley wrote:
    On 02/11/2025 19:10, Richard Damon wrote:
    So, how does the machine write out information that it doesn't have stored?

    A TM (at least typically) has a current machine state from a finite list
    of machine states, a reflecting extension can be defined to have a representation for each that it writes out upon interpreting a symbol at
    the tape head for each machine state (or for a subset of machine states,
    etc) instead of a simple move-left or move-right.

    For clarity I think dart200 is talking about an extended TM, extended to
    have one or more machine-state write-out transitions defined.

    For additional clarity, "machine state" does not normally refer to the
    tape contents nor to the current tape-cell index but the state of the
    machine (like register values, for machines defined to have those). You
    can't include the tape cell index in the machine state because then it
    would have an infinite number of machine states. I understand that would
    not be a TM, strictly.

    further more RTMs themselves don't just then solve the halting problem.
    RTMs allow for context-aware computing, which we can use to propose a context-aware decider, that can proactively avoid succumbing to semantic paradoxes


    G0 = () -> {
    if ( halts(G0) ) // false, as true would cause input to loop
    loop_forever()
    if ( loops(G0) ) // false, as true would cause input to halt
    halt

    if ( halts(G0) ) // true, as input does halt
    print "halts!"
    if ( loops(G0) ) // false, as input does not loop
    print "does not halt!"

    halt
    }

    --
    Tristan Wibberley

    ur not a boomer eh?


    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Richard Damon@[email protected] to comp.theory on Sun Nov 2 23:08:10 2025
    From Newsgroup: comp.theory

    On 11/2/25 3:45 PM, dart200 wrote:
    On 11/2/25 11:10 AM, Richard Damon wrote:
    On 11/1/25 12:37 PM, dart200 wrote:
    On 11/1/25 6:19 AM, Richard Damon wrote:
    On 11/1/25 2:18 AM, dart200 wrote:
    On 10/31/25 9:33 PM, Richard Damon wrote:
    On 10/31/25 11:52 PM, dart200 wrote:
    On 10/31/25 8:33 PM, Richard Damon wrote:
    On 10/31/25 11:16 PM, dart200 wrote:
    On 10/31/25 7:59 PM, Richard Damon wrote:
    On 10/31/25 3:29 PM, dart200 wrote:
    On 10/31/25 9:18 AM, Richard Damon wrote:
    On 10/30/25 2:27 PM, dart200 wrote:
    On 10/29/25 8:42 PM, olcott wrote:
    On 10/29/2025 10:23 PM, dart200 wrote:
    On 10/29/25 6:45 PM, Richard Damon wrote:
    On 10/29/25 12:24 PM, dart200 wrote:
    On 10/29/25 4:29 AM, Richard Damon wrote:
    On 10/28/25 11:23 PM, dart200 wrote:
    On 10/28/25 6:55 PM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>>> On 10/28/25 9:41 PM, dart200 wrote:
    On 10/28/25 6:34 PM, Chris M. Thomasson wrote: >>>>>>>>>>>>>>>>>>>>>> On 10/28/2025 5:43 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>> [...]
    not my argument u fucking useless boomer >>>>>>>>>>>>>>>>>>>>>>
    Well, how do you solve the halting problem? >>>>>>>>>>>>>>>>>>>>>>

    with reflection, something u lack as a useless >>>>>>>>>>>>>>>>>>>>> fucking boomer


    Which you can't say how to do, because your only >>>>>>>>>>>>>>>>>>>> answer is to define that you can.

    we haven't even gotten the point of working out the >>>>>>>>>>>>>>>>>>> ramifications of what i've defined,

    Because you haven't actually defined what you are doing. >>>>>>>>>>>>>>>>>
    blatant lie

    Really, show where you have ACTUALLY fully defined your >>>>>>>>>>>>>>>> computation machine to a level you can actually analyize >>>>>>>>>>>>>>>> one.

    TMs with a modification, specifically adding full >>>>>>>>>>>>>>> mechanical reflection to TMs in order to produce >>>>>>>>>>>>>>> Reflective TMs (RTMs)

    adding full mechanical reflection just means it gets a >>>>>>>>>>>>>>> single instruction that dumps *all* this information to >>>>>>>>>>>>>>> the tape, at the head, when that instruction is called: >>>>>>>>>>>>>>>
    1) machine description
    2) initial tape state
    3) current instruction
    4) current tape state

    call it REFLECT, if u will... this information allows a >>>>>>>>>>>>>>> computation to know exactly where it is in the overall >>>>>>>>>>>>>>> runtime of computation, and with that it can subvert >>>>>>>>>>>>>>> semantic paradoxes like the halting problem.

    you have to understand this is a *mechanical*
    modification to turing machines, not a computational one. >>>>>>>>>>>>>>> much like how RTMs can move a head left/right, or write >>>>>>>>>>>>>>> 0/1 to the tape ... they mechanically just do this >>>>>>>>>>>>>>> operation when the instruction is called.

    *these mechanics are (relatively) simple and self- >>>>>>>>>>>>>>> evidentially possible and therefore mathematically >>>>>>>>>>>>>>> feasible.*

    BUT, let me try a step deeper, and quote turing at you >>>>>>>>>>>>>>> from his original paper on computable numbers:

    /We may compare a man in the process of computing a real >>>>>>>>>>>>>>> number to a
    machine which is only capable of a finite number of >>>>>>>>>>>>>>> conditions/ [Tur36]

    THE ENTIRE JUSTIFICATION FOR TURING MACHINES RESTS >>>>>>>>>>>>>>> SQUARELY ON THE NOTION OF "CAN A MAN MECHANICALLY >>>>>>>>>>>>>>> UNDERTAKE THESE STATE TRANSITIONS"

    1) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>> then obviously he has access the full machine
    description, or how else the fuck is he doing any of >>>>>>>>>>>>>>> these state transitions??? therefore he obviously can >>>>>>>>>>>>>>> dump the machine description, in whatever format he has >>>>>>>>>>>>>>> it in, to the tape when REFLECT is encountered.

    2) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>> then obviously he can to the side save the initial tape >>>>>>>>>>>>>>> in a "buffer" that than cannot be overwritten. with that >>>>>>>>>>>>>>> he can can obviously dump the initial tape state to the >>>>>>>>>>>>>>> tape whenever REFLECT is encountered.

    3) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>> then obviously he has access input of a particular >>>>>>>>>>>>>>> transition, and therefore can dump that input to the tape >>>>>>>>>>>>>>> when REFLECT is encountered, in whatever format he >>>>>>>>>>>>>>> encounters it in.

    4) if a man is running thru the steps of a
    computation ... he can obviously duplicate the tape when >>>>>>>>>>>>>>> REFLECT is called. or more precisely he can save it into >>>>>>>>>>>>>>> a buffer, and dump from the buffer after putting 1-3 on >>>>>>>>>>>>>>> the tape.

    THAT IS THE SAME LEVEL OF FORMAT JUSTIFICATION TURING >>>>>>>>>>>>>>> ULTIMATELY GIVES FOR ANY COMPUTING MACHINE. SERIOUSLY >>>>>>>>>>>>>>> READ HIS FUCKING ESSAY YOU DORK

    please tell me where u are confused, not how u think i'm >>>>>>>>>>>>>>> wrong


    I can finally see the gist of what you are saying
    and it seems to be very similar to my idea.

    like i've said polcott, i consider you having a half-way >>>>>>>>>>>>> solution. u have something that's executable for sure, and >>>>>>>>>>>>> that is a definite improvement over total undecidability. >>>>>>>>>>>>>
    but it's not totally epistemically useful, it still >>>>>>>>>>>>> doesn't / effectively compute/ the halting function in a >>>>>>>>>>>>> *general* manner.

    i'm proposing something that's both executable *and* >>>>>>>>>>>>> epistemically useful for that purpose.

    the RTM i'm proposing here is just the underlying
    theoretical mechanisms that can be used to produce a >>>>>>>>>>>>> theoretically coherent context-dependent decider,

    which can then be used to /effectively compute/ the halting >>>>>>>>>>>>> function irrespective of semantic paradoxes ...

    something that is in total a step beyond the mere
    executability that you are proposing.


    No, the problem you have is that you make the same error and >>>>>>>>>>>> try to redefine the question, and allow your machine to give >>>>>>>>>>>> wrong answers.

    it's not a "wrong" answer if it contains no information to be >>>>>>>>>>> wrong about, and it's only given in a situation where there's >>>>>>>>>>> no "right" answer.

    But their is a right answer for every actual D based on every >>>>>>>>>> actual H.

    It just isn't the answer that H gives.


    how is it reasonable to expect a decider to respond
    truthfully to a situation that would only make that truth untrue >>>>>>>>>>
    Because deciders are just deterministic processes with a fixed >>>>>>>>>> algorithm.

    right, and context can be an input to the determinism

    console.trace() isn't nondeterministic, it's context dependent >>>>>>>>>

    Note, the system DOES understand that some problems turn out >>>>>>>>>> to be uncomputable, so no decider exists that can be correct >>>>>>>>>> for all inputs. Nothing wrong with that, it just means that >>>>>>>>>> any decider that wants to try to claim to solve it will be >>>>>>>>>> wrong for some input.

    The key point here is the question can't be asked until after >>>>>>>>>> a decider steps up to make the claim, as the prove-it-wrong >>>>>>>>>> input is different for every possible decider, and thus, by >>>>>>>>>> the time the question can be asked, the deciders result for >>>>>>>>>> that input was already determined by its algorithm. For the >>>>>>>>>> decider it isn't a matter of being "truthful", it is just >>>>>>>>>> doing what it was programmed to do.

    It becomes a matter of correctness, not truthfulness.

    It is the creator of the algoritm to which truthfulness is >>>>>>>>>> applied, and claiming that it will give the right answer when >>>>>>>>>> it doesn't shows that his claim wasn't truthful.


    there's no "correct" answer to a pathological question of
    halts(und) on this line:

    und = () -> halts(und) && loop_forever()

    Which isn't a computation if your symbology means that halts is >>>>>>>> taken from the context that the symbol und is used in.

    Sorry, you are just showing you don't know the meaning of the >>>>>>>> words.


    and the resolution is just to give a non-answer, so i split the >>>>>>>>> decider into two halves and made false a non-answer. it cannot >>>>>>>>> be a contradiction because it doesn't give information /in that >>>>>>>>> moment/ about und() to contradict.

    But since you are starting with the same category error of
    Olcott, you get the same nonsense he does.

    The category for the input to a Halt Decider is the
    representation of a Computation.

    By DEFINITION, this has a FULLY defined algorithm, and thus NOT >>>>>>>> a closure.

    If you want to try to define your own theory of computation that >>>>>>>> uses a broader defintion, go ahead, the problem is you will find >>>>>>>> you can do litle with it, and that the semantic paradox you are >>>>>>>> complaining about is totally due to you altered definition of a >>>>>>>> computation.

    right why do u keep asserting ur existential fuckup is more
    correct than a working decider on the matter???

    Because it doesn't give the right answer per the original problem.

    it gives you the right answer in all the places that the
    information of that answer is actually usable, and it does so
    generally.

    In other words, your logic says it is ok to be wrong in some cases,
    if you can give a "good excuse" why it is ok to be wrong there.

    All that does is destroy the meaning of "Correct".


    you are holding onto some notion of wanting a "right" answer where
    no right answer can exist, and that's just bizarre because with
    that desire u lose the ability to generally get a "right" answer in >>>>> first place.

    Because it *DOES* have a proper definition.

    ALWAYS means ALWAYS, not just where I want it.

    i can't believe ur a chief engineer asking for correctness where

    a) it has no practical usecase

    b) it ends up neutering ur theoretical ability to generally handle
    the problem

    ur demanding to be able to shoot urself in ur foot over some decider
    that can't even exist.

    weird fetish bro

    I guess you just think that lies are valid, and truth doesn't matter.

    Note, a real halt decider, if it existed, would have a LOT of actually
    practical uses, and the fact it can't exist is important knowledge.




    ur throwing the fucking baby out with the bathwater dude,

    it's fucking stupid

    Nope, because for actual problems like Busy Beaver, which again asks
    about for results from ALL possible machihes with specific criteria,
    that pattern might just coencidentally occur.



    If you want to be clear that you aren't working on that problem,
    or even working under that system, go ahead, just make sure you
    fully define what you are doing and try to show where it is useful. >>>>>>

    i'm not building a separate form of computing, i'm upgrading what
    we have

    In other words, you are admitting you are lying.

    It is a VIOLATION of the meaning of the words to change the meaning
    of the word as already used.

    You "upgrade" a system by adding new ideas without removing or
    changing any of the old.

    If you need to remove or change something, you have a new system.
    DEFINITION.

    like adding a new operation BY DEFINITION???

    Sure, it it can't be done within the limits of the previously defined
    system.


    honestly idk if ur just in denial, too senior to take criticism
    anymore, or just shoved a stick too far ur butthole,

    Nope, it is just sticking up for truth against stupid people who don't
    understand what they are talking about.


    but suggesting i'm not changing anything when i've spent pages trying
    to define a very particular change, even given it's a label RTM is
    just completely disingenuous of you.

    So, how does the machine write out information that it doesn't have
    stored?


    i'm insulted at the lack of consideration i'm getting here, clearly u
    have no willingness to engaging in genuine discussion, cause ur just
    blatantly lying about my position.

    Then you are being insulted by your own errors.

    YOU are the one who has decided not to engage in answering question
    put to you to explain.


    how can u offer me any meaningful feedback if ur just gunna blatantly
    lie about my position???

    What LIE have I said?

    To show that what I have said can't be done, just needs a showing how
    you claim it actually could be done within the restrictions of the
    framework you are in.

    Part of your problem, is it seems you haven't bothered to learn what
    the frameworks actually is.



    Again, all you are doing is admitting that you system is based on
    you claiming the right to lie about what you are doing, by ignoring
    the rules of the system.





    yes this does mean maybe /false/ could be returned everywhere, >>>>>>>>> but that's not fucking interesting ... nor does it give /
    effective computability/ in regards to the halting question.

    Since your input is a category error for the actual Computation >>>>>>>> Theory as classical defined, your results don't apply.

    classic computing theory doesn't have a halt decider defined
    tho ... u only have a halting function, not a decider. the
    decider is *disproven* and therefore cannot be defined as per
    consensus theory.

    u fucking retard.

    u keep asserting SOMETHING THAT DOESN'T EVEN EXIST is somehow
    more correct than SOMETHING THAT CAN EXIST

    The DEFINITION of a Halting Decider exists.

    The fact that the definiton is not realizable, proves the theory.

    To call something a Halt Decider that doesn't meet the definition, >>>>>> is just a lie, making the person making the claim just a liar.

    just like there's more than one way to ask the same question,
    there's more than one way to define a halting decider, which is the >>>>> interface into an algorithm, not the algorithm itself.

    Right, but if they are semantically different, it isn't the same
    question.


    the one i'm proposing can actually generally exist, unlike the
    classical definition.

    But only is a Computation System that is broken, because programs
    don't reliable produce the same answer when run.

    Once you allow things to lie, you have destroyed the concept of truth. >>>>

    it's really that simple.

    Yes, allowing lying seems simple until you need to count on something. >>>>


    We can easily DEFINE things that don't exist.>


    which is the point of the context-dependent decider. halts() is >>>>>>>>> required to return /true/ only whereever both /possible/ and / >>>>>>>>> truthful/ to do so.

    So, if you want to try to talk about this, you first need to try >>>>>>>> to develop a basic theory about context dependent Computations, >>>>>>>> to see if it can actually be useful for anything.

    i fucking already did. u just aren't reading cause ur too busy
    trying to show off how big a fucking retard u can be.

    No you haven't, unless you consider your claim that is the
    equivalent of defining that 1 == 2 is a definition.

    DESCRIBING what you want to do is NOT the same as DEFINING how to >>>>>> do it.

    ur special pleading is just so mindnumbingly boring. i DEFINED my
    mechanics in MORE detail that turing did his, and yet u arbitrarily >>>>> reject it as "describing" which seems like a made up distinction here. >>>>
    NO YOU HAVEN'T.

    You have PRESUMED an ability on the machine that can not actually be
    implemented on it without breaking it.

    i'm intelligent enough to write down the machine description to the
    tape when i see the operation REFLECT, and therefore accept the
    defined ability

    But is it the RIGHT machine description for the program in question?

    And where does it get the contents of the original tape?


    i get that ur apparently too fucking retarded to do that, and
    therefore don't accept it.

    Nope, it seems you are too retarded to actually think about what other
    say about your system. You suffer from the fallacy that you assume you
    are right and thus ignore the definitions of what you are talking about.



    Your "Definition" is the equivalent of defining that for this work,
    I will consider 1 to be equal to 2.


    sometimes i really do feel like just buying a gun a shooting
    myself, cause i really am stuck on a planet of fucking idiots.

    y must u endlessly torture me so with ur fucking idiocracy?


    I could say the same to you. Why are you so stuck on insisting on
    things that are not.

    Computations, as defined, can't change their answers for the same
    input. PERIOD.

    To insist they can is just to admit you are just a lunatic.

    right, printing stack traces actually isn't part of computing. every
    time we step thru debugging we're doing theoretically indescribable
    magic, according to richard


    And what does that have to do with what I said?

    And yes, a routine looking at who called it is NOT part of "Computing"
    per computation theory, as it can't matter.

    It seems part of your problem is you don't actually know what you are
    talking about.

    Remember, Computing isn't directly about "Computers" as we currently
    know them, but is more fundamental.

    It seems you have forgotten that "Computing" as a term existed BEFORE
    the computers you want to talk about.


    blows my mind that u think a computing function that prints a stack
    trace does something which cannot be done on and explained by a turing machine

    Well, try to define a Turing Machinne that does that.

    Your first problem will be that Turing Machines don't have a stack to dump.


    i'm not sure y u think this mindnumbing gishgallop u shit out is worth a continued response to

    can u explain what my argument even is?

    or are you basically just a fucking chatbot keying off one sentence at a time?


    No, but it seems you are just a idiot that refuses to learn the meaning
    of the words you are using.

    As I have pointed out several times, trying to relate a "Computation"
    per COmputation Theory to behavior of modern computers is just an error.

    A "stack dump" routine, that looks beyond the stack of the suppsoed computation it is part of, make that thing not a computation, by definition.

    Parts of computer programs can fail to be computations, and this is one
    thing that makes them hard to analylize with computation theory.

    And, you continued attempts to make such a comparison just proves tnat
    you are too stupid to understand that fact.

    Computations are not about the behavior of Computers, Computers are a
    tool that allow us to more easily do computations, but they can do other things as well, and that ability, when used, makes them very hard to
    analyize.

    If you want to make a theory based on the modern computer, go ahead,
    your first step is to define what you actually mean.

    Your second step is to see what basic properties you can actually define.

    The fact that your base unit of computation can give different answers
    for the same "input" based on the "context" you are in, is going to make
    thing VERY hard to work with, which is why it isn't done that way.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Richard Damon@[email protected] to comp.theory on Sun Nov 2 23:13:17 2025
    From Newsgroup: comp.theory

    On 11/2/25 3:57 PM, Tristan Wibberley wrote:
    On 02/11/2025 19:10, Richard Damon wrote:
    So, how does the machine write out information that it doesn't have stored?

    A TM (at least typically) has a current machine state from a finite list
    of machine states, a reflecting extension can be defined to have a representation for each that it writes out upon interpreting a symbol at
    the tape head for each machine state (or for a subset of machine states,
    etc) instead of a simple move-left or move-right.

    But their isn't a unique representation for that, and what good is
    dumping the information in a format that the program doesn't expect.

    And dumping the code isn't the biggest problem, although it does mean
    such a Turing Macine no longer has the property of always being a
    computation under the operation of composition of sub-proggrams.

    The bigger issue is recovering the inital tape as defined.


    For clarity I think dart200 is talking about an extended TM, extended to
    have one or more machine-state write-out transitions defined.

    And such an extended machine means it fails to have the essential
    property of a standard Turing Machine of all programs and subprograms
    designed for it are actual computations.

    That property is what make it a computation machine.>
    For additional clarity, "machine state" does not normally refer to the
    tape contents nor to the current tape-cell index but the state of the
    machine (like register values, for machines defined to have those). You
    can't include the tape cell index in the machine state because then it
    would have an infinite number of machine states. I understand that would
    not be a TM, strictly.

    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.


    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Richard Damon@[email protected] to comp.theory on Sun Nov 2 23:22:40 2025
    From Newsgroup: comp.theory

    On 11/2/25 10:29 PM, dart200 wrote:
    On 11/2/25 12:57 PM, Tristan Wibberley wrote:
    On 02/11/2025 19:10, Richard Damon wrote:
    So, how does the machine write out information that it doesn't have
    stored?

    A TM (at least typically) has a current machine state from a finite list
    of machine states, a reflecting extension can be defined to have a
    representation for each that it writes out upon interpreting a symbol at
    the tape head for each machine state (or for a subset of machine states,
    etc) instead of a simple move-left or move-right.

    For clarity I think dart200 is talking about an extended TM, extended to
    have one or more machine-state write-out transitions defined.

    For additional clarity, "machine state" does not normally refer to the
    tape contents nor to the current tape-cell index but the state of the
    machine (like register values, for machines defined to have those). You
    can't include the tape cell index in the machine state because then it
    would have an infinite number of machine states. I understand that would
    not be a TM, strictly.

    further more RTMs themselves don't just then solve the halting problem.
    RTMs allow for context-aware computing, which we can use to propose a context-aware decider, that can proactively avoid succumbing to semantic paradoxes

    But they CAN'T solve the Halting Problem as defined, as the input for
    the problem is to be a computation, and the decider needs to be a
    computation that always gives the right answer.

    If the decider isn't a computation, it can't be right, as the question
    *IS* about a strict input to output mapping, as that *IS* the mapping of
    the behavor of a program.



    G0 = () -> {
      if ( halts(G0) )      // false, as true would cause input to loop
        loop_forever()
      if ( loops(G0) )      // false, as true would cause input to halt
        halt

      if ( halts(G0) )      // true, as input does halt
        print "halts!"
      if ( loops(G0) )      // false, as input does not loop
        print "does not halt!"

      halt
    }

    It seems you have two different deciders.

    And G0, as defined halts.

    but halts(G0) is somehow "correct" to return false even though that
    answer is wrong.

    That just means you "logic" is based on the idea that wrong is sometimes right.

    If you haven't defined that the decider always gives the right answer,
    how can you know that the answer it gave you is actually correct and not
    one of its allowed lies?

    When dealing with absolute truth, there are no "unimportant" cases.

    All you are doing is demonstrating that you logic is just internally inconsistant.



    --
    Tristan Wibberley

    ur not a boomer eh?


    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.



    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Richard Damon@[email protected] to comp.theory on Sun Nov 2 23:29:58 2025
    From Newsgroup: comp.theory

    On 11/2/25 8:05 PM, dart200 wrote:
    On 11/1/25 1:26 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 11/1/25 9:37 AM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 10/31/25 9:20 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    there's no "correct" answer to a pathological question of
    halts(und) on
    this line:

    und = () -> halts(und) && loop_forever()

    Yes there is; the correct answer is ¬halts(und). The logical negation >>>>>> of whatever halts(und) decides.

    the correct answer is to give the opposite of it's own answer???

    The correct answer isn't /to give/ anything; it just /is/

    halts() is a total decider, ur now just ignoring what i defined()

    I'm obviously aware of how you defined it.

    If you initially impose the definition "halts is the name of a total
    decider", then you arrive at a contradiction: halts cannot be
    deciding und, as required of a total decider.

    So then, the definition is a kind of proposition, which is shown false;
    it does not hold.

    But what is the logical opposite of the proposition "halts is the name
    of a total decider?" It is not: "halts does not exist", but simply:
    "halts is /not/ the name of a total decider".

    A halts can exist just fine, just not as a total decider.

    Stop insisting that halts is a the total halting decider which cannot
    exist, allowing it to be a partial halting decider which cannot be
    total, and the apparent paradox goes away.  und then exists and has a >>>> defnite halting status, which is opposite to whatever that partial
    decider says.

    great now ur arguing polcott is right.

    Not really; he has never agreed with the above!

    he might not agree with the wording, but it is kinda what he's proposing


    the problem with a set of partial deciders is that the resolution isn't
    /effectively computable/. with "partial" deciders then the question
    becomes which partial decider is correct: u've just kicked the can down
    the road

    Yes; because no combination of partial deciders gives you a total
    decider.

    We can trivially form a set of partial deciders which has complete
    coverage:

       H_0 := (P) -> 0

       H_1 := (P) -> 1

    H_0 correctly decides all non-terminating inputs. H_1 correctly
    decides all terminating inputs. So one of these two is always
    correct! (Notice how I'm not saying anything more informative
    than "a machine always halts or does not halt").

    We don't know which decider is the right one for which input.

    We need a third decider H whch decides whether to use H_0
    or H_1.

    This example may look silly: because what is H, but simply
    a decider which has to determine whether to returh 0 or 1
    (which are only trivially disguised as H_0(P) and H_1(P))?

    But this idea generalizes. No matter what kind of partial deciders we
    have, or how many, deciding which one is correct for which input is
    equivalent to deciding halting.

    The result of combining partial deciders will always be a decider
    that can be shown not to be total.

    The can is indeed always kicked down the road.

    great! so ur own answered question of why i'm not particularly excited
    by partial deciders, and want a way to /effectively compute/ whether an input machine halts not

    do we need to have this discussion again kaz, or???


    yes this does mean maybe /false/ could be returned everywhere, but >>>>>>> that's not fucking interesting ... nor does it give /effective
    computability/ in regards to the halting question.

    which is the point of the context-dependent decider. halts() is >>>>>>> required
    to return /true/ only whereever both /possible/ and /truthful/ to >>>>>>> do so.

    if u don't agree with this, that's fine...

    but you actually need to construct a contradiction, you can't just >>>>>>> continually make bare assertions that it's a contradiction.

    If contexts are represenable in a Turing substrate, then everything >>>>>> succumbs to Turing incomputability of halting.

    If contexts are not representable in Turing, then the
    model-with-contexts is either effectievely computable non-Turing, or >>>>>> else contexts are just imaginary, like magic oracles.

    false dichotomy

    Well, trichotomy, by my count.

    - contexts are Turing (1)
    - contexts are not Turing:
        - somehow effectively computable (2)
        - not computable (3)

    What possibility have I unintentionally left out?

    like i'm even open to being wrong, but like u need to work with the
    system in order to show a contradiction with the system itself,
    just not
    assume the contradiction will happen.

    If they are imaginary, you only have a thought experiment.
    (In which Turing-plus-contexts machines can decide Turing
    halting, but so what?)

    REFLECT is completely mechanically demonstrable operation with a
    pen and
    paper, i demonstrated it in a post here already, justifying it's
    theoretical existence

    OK, that points to it being embeddable in Turing
    computation. Which then points to undecidabiity.

    or ur just assuming a false dichotomy

    If your mechanical pencil-and-paper calculations are shown to be Turing,
    anything rendered using those calculations is Turing. It is not
    a dichotomy, true or false.

    what i'm proposing obviously will look very much like TMs in many if not most computatation. in fact: any computation can run on a TM will in
    fact be run the same on an RTM.

    the differences in RTMs only come up in computations that aren't
    directly expressible on TMs


    the TM limitation is mechanical, not computational. classical theory
    only supposes about computational limitations, there's no notion of
    mechanical limitation.

    Are you talking about analog computation with continuous quantities,
    and such?

    no


    So your pencil-and-paper calculations are actually perhaps geometric,
    like compass-and-ruler derivations?

    r u intentionally trying to be obtuse?


    but the mechanical, not computational, limitation means:

    - TMs can simulate RTMs
    - RTMs don't fix the problems with TMs
    - it doesn't matter that they don't because RTMs can compute
        everything TMs do + have increased power

    If TMs can simulate RTMs then they do not have increased power.

    no. TMs have the computational power to handle what RTMs compute ... but
    the lack of the REFLECT operation mechanically limits them from
    dependable access to the required runtime information to actually be
    RTMs directly.

    this relationship just can't be pigeonholed into what u'd like, you are overgeneralizing

    this doesn't solve TM limitation because again the problem is TMs lack mechanical access to not be limited. RTMs instead *trivialize* the TM halting problem because RTMs can express everything TMs + overcome
    attempts at constructing semantic paradoxes with RTMs.


    They may have increased "expressive power", which is something only
    relevant to how "nice" they are to write applications in.

    A different Turing model can also have better efficiency; i.e the best
    way of solving a problem in one kind of machine may have exponential
    time whereas another can express a solution that is polynomial.

    Since

    - ugly code that correctly solves a problem halts,
    - inefficient code that correctly solves a problem halts,

    neither expressive power nor efficiency are relevant in halting.

    what i'm doing is trivializing TMs undecidability, not resolving it. but >>> this overall resolves the halting problem by providing a mechanical
    improvement that can allow theory to escape limitations stemming from
    the liar's paradox being encoded into machines

    (If you want to talk Olcott, /that/'s a great narrative:
    "halting problem implies liar paradox.")


    REFLECT gives a semantic decider the ability to know it's calling
    context, which gives it the ability to return a context dependent
    answer.

    If you find some computational model with contexts which is
    effective,
    but beyond Turing, then /that/ is the big deal; delegate the halting >>>>>> cruft to others. You are important because you've done what nobody >>>>>> else
    has done before: you found an effective method of computation that is >>>>>> not Turing! So you go to conference sall over the world where you are >>>>>> flocked by groupies. Fuck halting, at that point, you know?

    it's a mild extension of TMs,

    If it is mechanically demonstrable with pen and paper, then it isn't
    an extension of TM.

    no idea what u mean here, TMs are likewise also demonstrable with a pen
    and paper, and that was turing's original justification for their
    feasibility

    (What I mean that it's just TM, and not an extension.)

    Functions can depend on their context. That context can make some
    instances of them halting and others not, even with all the explicit
    arguments being the same. Thus we cannot ask whether a function halts; >>>> it has to be, does it halt /in what calling context/.

    only if it involves a REFLECT command, otherwise it's not possible to be >>> dependent on context.

    So if only deciders are allowed to use REFLECT, but the inputs
    are not allowed to use REFLECT, then indeed it is impossible for an
    input to contradict a decider which uses REFLECT.

    But then deciders are not deciding everything; they are deciding
    only those inputs that don't use REFLECT.

    inputs are allowed to use REFLECT. this doesn't make it possible to construct a halting paradox because it does not stop deciders from
    using /false/ to escape any demonstrable paradox.



    The issue is that the Halting Problem doesn't presume the existance of
    such a total decider, just askes if such a thing is possible, and it
    turns out it isn't.

    It defines the properties such a machine would need to have to be one,
    and it just turns out that nothing can meet them all at once.

    The some PROOFS, use the valid logical technique of proof by
    contraciction, where we put up a trial presuposition of the existance of
    such a thing to see what cmes out of it, and the results being a
    contradiction means that that initial presupposition must be wrong (or
    the whole logic system was inconsistant before making that
    presuposition, which we hope isn't the case).

    There are other proofs, that don't need that assumption, but show that
    for every possible decider we might make, we can make an input that it
    will give the wrong answer for.

    These proofs are a bit longer, and do require being able to think of
    working through an infinite set of all deciders. But avoid the use of
    proof by contradiction, that causes some people these problems.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Sun Nov 2 20:42:14 2025
    From Newsgroup: comp.theory

    On 11/2/25 8:08 PM, Richard Damon wrote:
    On 11/2/25 3:45 PM, dart200 wrote:
    On 11/2/25 11:10 AM, Richard Damon wrote:
    On 11/1/25 12:37 PM, dart200 wrote:
    On 11/1/25 6:19 AM, Richard Damon wrote:
    On 11/1/25 2:18 AM, dart200 wrote:
    On 10/31/25 9:33 PM, Richard Damon wrote:
    On 10/31/25 11:52 PM, dart200 wrote:
    On 10/31/25 8:33 PM, Richard Damon wrote:
    On 10/31/25 11:16 PM, dart200 wrote:
    On 10/31/25 7:59 PM, Richard Damon wrote:
    On 10/31/25 3:29 PM, dart200 wrote:
    On 10/31/25 9:18 AM, Richard Damon wrote:
    On 10/30/25 2:27 PM, dart200 wrote:
    On 10/29/25 8:42 PM, olcott wrote:
    On 10/29/2025 10:23 PM, dart200 wrote:
    On 10/29/25 6:45 PM, Richard Damon wrote:
    On 10/29/25 12:24 PM, dart200 wrote:
    On 10/29/25 4:29 AM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>> On 10/28/25 11:23 PM, dart200 wrote:
    On 10/28/25 6:55 PM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>>>> On 10/28/25 9:41 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>> On 10/28/25 6:34 PM, Chris M. Thomasson wrote: >>>>>>>>>>>>>>>>>>>>>>> On 10/28/2025 5:43 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>>> [...]
    not my argument u fucking useless boomer >>>>>>>>>>>>>>>>>>>>>>>
    Well, how do you solve the halting problem? >>>>>>>>>>>>>>>>>>>>>>>

    with reflection, something u lack as a useless >>>>>>>>>>>>>>>>>>>>>> fucking boomer


    Which you can't say how to do, because your only >>>>>>>>>>>>>>>>>>>>> answer is to define that you can.

    we haven't even gotten the point of working out the >>>>>>>>>>>>>>>>>>>> ramifications of what i've defined,

    Because you haven't actually defined what you are doing. >>>>>>>>>>>>>>>>>>
    blatant lie

    Really, show where you have ACTUALLY fully defined your >>>>>>>>>>>>>>>>> computation machine to a level you can actually >>>>>>>>>>>>>>>>> analyize one.

    TMs with a modification, specifically adding full >>>>>>>>>>>>>>>> mechanical reflection to TMs in order to produce >>>>>>>>>>>>>>>> Reflective TMs (RTMs)

    adding full mechanical reflection just means it gets a >>>>>>>>>>>>>>>> single instruction that dumps *all* this information to >>>>>>>>>>>>>>>> the tape, at the head, when that instruction is called: >>>>>>>>>>>>>>>>
    1) machine description
    2) initial tape state
    3) current instruction
    4) current tape state

    call it REFLECT, if u will... this information allows a >>>>>>>>>>>>>>>> computation to know exactly where it is in the overall >>>>>>>>>>>>>>>> runtime of computation, and with that it can subvert >>>>>>>>>>>>>>>> semantic paradoxes like the halting problem.

    you have to understand this is a *mechanical* >>>>>>>>>>>>>>>> modification to turing machines, not a computational >>>>>>>>>>>>>>>> one. much like how RTMs can move a head left/right, or >>>>>>>>>>>>>>>> write 0/1 to the tape ... they mechanically just do this >>>>>>>>>>>>>>>> operation when the instruction is called.

    *these mechanics are (relatively) simple and self- >>>>>>>>>>>>>>>> evidentially possible and therefore mathematically >>>>>>>>>>>>>>>> feasible.*

    BUT, let me try a step deeper, and quote turing at you >>>>>>>>>>>>>>>> from his original paper on computable numbers: >>>>>>>>>>>>>>>>
    /We may compare a man in the process of computing a real >>>>>>>>>>>>>>>> number to a
    machine which is only capable of a finite number of >>>>>>>>>>>>>>>> conditions/ [Tur36]

    THE ENTIRE JUSTIFICATION FOR TURING MACHINES RESTS >>>>>>>>>>>>>>>> SQUARELY ON THE NOTION OF "CAN A MAN MECHANICALLY >>>>>>>>>>>>>>>> UNDERTAKE THESE STATE TRANSITIONS"

    1) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>>> then obviously he has access the full machine >>>>>>>>>>>>>>>> description, or how else the fuck is he doing any of >>>>>>>>>>>>>>>> these state transitions??? therefore he obviously can >>>>>>>>>>>>>>>> dump the machine description, in whatever format he has >>>>>>>>>>>>>>>> it in, to the tape when REFLECT is encountered. >>>>>>>>>>>>>>>>
    2) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>>> then obviously he can to the side save the initial tape >>>>>>>>>>>>>>>> in a "buffer" that than cannot be overwritten. with that >>>>>>>>>>>>>>>> he can can obviously dump the initial tape state to the >>>>>>>>>>>>>>>> tape whenever REFLECT is encountered.

    3) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>>> then obviously he has access input of a particular >>>>>>>>>>>>>>>> transition, and therefore can dump that input to the >>>>>>>>>>>>>>>> tape when REFLECT is encountered, in whatever format he >>>>>>>>>>>>>>>> encounters it in.

    4) if a man is running thru the steps of a
    computation ... he can obviously duplicate the tape when >>>>>>>>>>>>>>>> REFLECT is called. or more precisely he can save it into >>>>>>>>>>>>>>>> a buffer, and dump from the buffer after putting 1-3 on >>>>>>>>>>>>>>>> the tape.

    THAT IS THE SAME LEVEL OF FORMAT JUSTIFICATION TURING >>>>>>>>>>>>>>>> ULTIMATELY GIVES FOR ANY COMPUTING MACHINE. SERIOUSLY >>>>>>>>>>>>>>>> READ HIS FUCKING ESSAY YOU DORK

    please tell me where u are confused, not how u think i'm >>>>>>>>>>>>>>>> wrong


    I can finally see the gist of what you are saying >>>>>>>>>>>>>>> and it seems to be very similar to my idea.

    like i've said polcott, i consider you having a half-way >>>>>>>>>>>>>> solution. u have something that's executable for sure, and >>>>>>>>>>>>>> that is a definite improvement over total undecidability. >>>>>>>>>>>>>>
    but it's not totally epistemically useful, it still >>>>>>>>>>>>>> doesn't / effectively compute/ the halting function in a >>>>>>>>>>>>>> *general* manner.

    i'm proposing something that's both executable *and* >>>>>>>>>>>>>> epistemically useful for that purpose.

    the RTM i'm proposing here is just the underlying >>>>>>>>>>>>>> theoretical mechanisms that can be used to produce a >>>>>>>>>>>>>> theoretically coherent context-dependent decider,

    which can then be used to /effectively compute/ the >>>>>>>>>>>>>> halting function irrespective of semantic paradoxes ... >>>>>>>>>>>>>>
    something that is in total a step beyond the mere >>>>>>>>>>>>>> executability that you are proposing.


    No, the problem you have is that you make the same error >>>>>>>>>>>>> and try to redefine the question, and allow your machine to >>>>>>>>>>>>> give wrong answers.

    it's not a "wrong" answer if it contains no information to >>>>>>>>>>>> be wrong about, and it's only given in a situation where >>>>>>>>>>>> there's no "right" answer.

    But their is a right answer for every actual D based on every >>>>>>>>>>> actual H.

    It just isn't the answer that H gives.


    how is it reasonable to expect a decider to respond
    truthfully to a situation that would only make that truth >>>>>>>>>>>> untrue

    Because deciders are just deterministic processes with a >>>>>>>>>>> fixed algorithm.

    right, and context can be an input to the determinism

    console.trace() isn't nondeterministic, it's context dependent >>>>>>>>>>

    Note, the system DOES understand that some problems turn out >>>>>>>>>>> to be uncomputable, so no decider exists that can be correct >>>>>>>>>>> for all inputs. Nothing wrong with that, it just means that >>>>>>>>>>> any decider that wants to try to claim to solve it will be >>>>>>>>>>> wrong for some input.

    The key point here is the question can't be asked until after >>>>>>>>>>> a decider steps up to make the claim, as the prove-it-wrong >>>>>>>>>>> input is different for every possible decider, and thus, by >>>>>>>>>>> the time the question can be asked, the deciders result for >>>>>>>>>>> that input was already determined by its algorithm. For the >>>>>>>>>>> decider it isn't a matter of being "truthful", it is just >>>>>>>>>>> doing what it was programmed to do.

    It becomes a matter of correctness, not truthfulness.

    It is the creator of the algoritm to which truthfulness is >>>>>>>>>>> applied, and claiming that it will give the right answer when >>>>>>>>>>> it doesn't shows that his claim wasn't truthful.


    there's no "correct" answer to a pathological question of >>>>>>>>>> halts(und) on this line:

    und = () -> halts(und) && loop_forever()

    Which isn't a computation if your symbology means that halts is >>>>>>>>> taken from the context that the symbol und is used in.

    Sorry, you are just showing you don't know the meaning of the >>>>>>>>> words.


    and the resolution is just to give a non-answer, so i split >>>>>>>>>> the decider into two halves and made false a non-answer. it >>>>>>>>>> cannot be a contradiction because it doesn't give
    information /in that moment/ about und() to contradict.

    But since you are starting with the same category error of
    Olcott, you get the same nonsense he does.

    The category for the input to a Halt Decider is the
    representation of a Computation.

    By DEFINITION, this has a FULLY defined algorithm, and thus NOT >>>>>>>>> a closure.

    If you want to try to define your own theory of computation >>>>>>>>> that uses a broader defintion, go ahead, the problem is you >>>>>>>>> will find you can do litle with it, and that the semantic
    paradox you are complaining about is totally due to you altered >>>>>>>>> definition of a computation.

    right why do u keep asserting ur existential fuckup is more
    correct than a working decider on the matter???

    Because it doesn't give the right answer per the original problem. >>>>>>
    it gives you the right answer in all the places that the
    information of that answer is actually usable, and it does so
    generally.

    In other words, your logic says it is ok to be wrong in some cases, >>>>> if you can give a "good excuse" why it is ok to be wrong there.

    All that does is destroy the meaning of "Correct".


    you are holding onto some notion of wanting a "right" answer where >>>>>> no right answer can exist, and that's just bizarre because with
    that desire u lose the ability to generally get a "right" answer
    in first place.

    Because it *DOES* have a proper definition.

    ALWAYS means ALWAYS, not just where I want it.

    i can't believe ur a chief engineer asking for correctness where

    a) it has no practical usecase

    b) it ends up neutering ur theoretical ability to generally handle
    the problem

    ur demanding to be able to shoot urself in ur foot over some decider
    that can't even exist.

    weird fetish bro

    I guess you just think that lies are valid, and truth doesn't matter.

    Note, a real halt decider, if it existed, would have a LOT of
    actually practical uses, and the fact it can't exist is important
    knowledge.




    ur throwing the fucking baby out with the bathwater dude,

    it's fucking stupid

    Nope, because for actual problems like Busy Beaver, which again
    asks about for results from ALL possible machihes with specific
    criteria, that pattern might just coencidentally occur.



    If you want to be clear that you aren't working on that problem, >>>>>>> or even working under that system, go ahead, just make sure you >>>>>>> fully define what you are doing and try to show where it is useful. >>>>>>>

    i'm not building a separate form of computing, i'm upgrading what >>>>>> we have

    In other words, you are admitting you are lying.

    It is a VIOLATION of the meaning of the words to change the meaning >>>>> of the word as already used.

    You "upgrade" a system by adding new ideas without removing or
    changing any of the old.

    If you need to remove or change something, you have a new system.
    DEFINITION.

    like adding a new operation BY DEFINITION???

    Sure, it it can't be done within the limits of the previously defined
    system.


    honestly idk if ur just in denial, too senior to take criticism
    anymore, or just shoved a stick too far ur butthole,

    Nope, it is just sticking up for truth against stupid people who
    don't understand what they are talking about.


    but suggesting i'm not changing anything when i've spent pages
    trying to define a very particular change, even given it's a label
    RTM is just completely disingenuous of you.

    So, how does the machine write out information that it doesn't have
    stored?


    i'm insulted at the lack of consideration i'm getting here, clearly
    u have no willingness to engaging in genuine discussion, cause ur
    just blatantly lying about my position.

    Then you are being insulted by your own errors.

    YOU are the one who has decided not to engage in answering question
    put to you to explain.


    how can u offer me any meaningful feedback if ur just gunna
    blatantly lie about my position???

    What LIE have I said?

    To show that what I have said can't be done, just needs a showing how
    you claim it actually could be done within the restrictions of the
    framework you are in.

    Part of your problem, is it seems you haven't bothered to learn what
    the frameworks actually is.



    Again, all you are doing is admitting that you system is based on
    you claiming the right to lie about what you are doing, by ignoring >>>>> the rules of the system.





    yes this does mean maybe /false/ could be returned everywhere, >>>>>>>>>> but that's not fucking interesting ... nor does it give / >>>>>>>>>> effective computability/ in regards to the halting question. >>>>>>>>>
    Since your input is a category error for the actual Computation >>>>>>>>> Theory as classical defined, your results don't apply.

    classic computing theory doesn't have a halt decider defined
    tho ... u only have a halting function, not a decider. the
    decider is *disproven* and therefore cannot be defined as per >>>>>>>> consensus theory.

    u fucking retard.

    u keep asserting SOMETHING THAT DOESN'T EVEN EXIST is somehow >>>>>>>> more correct than SOMETHING THAT CAN EXIST

    The DEFINITION of a Halting Decider exists.

    The fact that the definiton is not realizable, proves the theory. >>>>>>>
    To call something a Halt Decider that doesn't meet the
    definition, is just a lie, making the person making the claim
    just a liar.

    just like there's more than one way to ask the same question,
    there's more than one way to define a halting decider, which is
    the interface into an algorithm, not the algorithm itself.

    Right, but if they are semantically different, it isn't the same
    question.


    the one i'm proposing can actually generally exist, unlike the
    classical definition.

    But only is a Computation System that is broken, because programs
    don't reliable produce the same answer when run.

    Once you allow things to lie, you have destroyed the concept of truth. >>>>>

    it's really that simple.

    Yes, allowing lying seems simple until you need to count on something. >>>>>


    We can easily DEFINE things that don't exist.>


    which is the point of the context-dependent decider. halts() >>>>>>>>>> is required to return /true/ only whereever both /possible/ >>>>>>>>>> and / truthful/ to do so.

    So, if you want to try to talk about this, you first need to >>>>>>>>> try to develop a basic theory about context dependent
    Computations, to see if it can actually be useful for anything. >>>>>>>>
    i fucking already did. u just aren't reading cause ur too busy >>>>>>>> trying to show off how big a fucking retard u can be.

    No you haven't, unless you consider your claim that is the
    equivalent of defining that 1 == 2 is a definition.

    DESCRIBING what you want to do is NOT the same as DEFINING how to >>>>>>> do it.

    ur special pleading is just so mindnumbingly boring. i DEFINED my >>>>>> mechanics in MORE detail that turing did his, and yet u
    arbitrarily reject it as "describing" which seems like a made up
    distinction here.

    NO YOU HAVEN'T.

    You have PRESUMED an ability on the machine that can not actually
    be implemented on it without breaking it.

    i'm intelligent enough to write down the machine description to the
    tape when i see the operation REFLECT, and therefore accept the
    defined ability

    But is it the RIGHT machine description for the program in question?

    And where does it get the contents of the original tape?


    i get that ur apparently too fucking retarded to do that, and
    therefore don't accept it.

    Nope, it seems you are too retarded to actually think about what
    other say about your system. You suffer from the fallacy that you
    assume you are right and thus ignore the definitions of what you are
    talking about.



    Your "Definition" is the equivalent of defining that for this work, >>>>> I will consider 1 to be equal to 2.


    sometimes i really do feel like just buying a gun a shooting
    myself, cause i really am stuck on a planet of fucking idiots.

    y must u endlessly torture me so with ur fucking idiocracy?


    I could say the same to you. Why are you so stuck on insisting on
    things that are not.

    Computations, as defined, can't change their answers for the same
    input. PERIOD.

    To insist they can is just to admit you are just a lunatic.

    right, printing stack traces actually isn't part of computing. every
    time we step thru debugging we're doing theoretically indescribable
    magic, according to richard


    And what does that have to do with what I said?

    And yes, a routine looking at who called it is NOT part of
    "Computing" per computation theory, as it can't matter.

    It seems part of your problem is you don't actually know what you are
    talking about.

    Remember, Computing isn't directly about "Computers" as we currently
    know them, but is more fundamental.

    It seems you have forgotten that "Computing" as a term existed BEFORE
    the computers you want to talk about.


    blows my mind that u think a computing function that prints a stack
    trace does something which cannot be done on and explained by a turing
    machine

    Well, try to define a Turing Machinne that does that.

    Your first problem will be that Turing Machines don't have a stack to dump.


    i'm not sure y u think this mindnumbing gishgallop u shit out is worth
    a continued response to

    can u explain what my argument even is?

    or are you basically just a fucking chatbot keying off one sentence at
    a time?


    No,  but it seems you are just a idiot that refuses to learn the meaning
    of the words you are using.

    As I have pointed out several times, trying to relate a "Computation"
    per COmputation Theory to behavior of modern computers is just an error.

    well, that's just fucking retarded.

    do you honestly think the behavior exhibiting within modern computing
    somehow exists outside the theory of computing???
    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Richard Damon@[email protected] to comp.theory on Sun Nov 2 23:53:15 2025
    From Newsgroup: comp.theory

    On 11/2/25 11:42 PM, dart200 wrote:
    On 11/2/25 8:08 PM, Richard Damon wrote:
    On 11/2/25 3:45 PM, dart200 wrote:
    On 11/2/25 11:10 AM, Richard Damon wrote:
    On 11/1/25 12:37 PM, dart200 wrote:
    On 11/1/25 6:19 AM, Richard Damon wrote:
    On 11/1/25 2:18 AM, dart200 wrote:
    On 10/31/25 9:33 PM, Richard Damon wrote:
    On 10/31/25 11:52 PM, dart200 wrote:
    On 10/31/25 8:33 PM, Richard Damon wrote:
    On 10/31/25 11:16 PM, dart200 wrote:
    On 10/31/25 7:59 PM, Richard Damon wrote:
    On 10/31/25 3:29 PM, dart200 wrote:
    On 10/31/25 9:18 AM, Richard Damon wrote:
    On 10/30/25 2:27 PM, dart200 wrote:
    On 10/29/25 8:42 PM, olcott wrote:
    On 10/29/2025 10:23 PM, dart200 wrote:
    On 10/29/25 6:45 PM, Richard Damon wrote:
    On 10/29/25 12:24 PM, dart200 wrote:
    On 10/29/25 4:29 AM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>>> On 10/28/25 11:23 PM, dart200 wrote:
    On 10/28/25 6:55 PM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>>>>> On 10/28/25 9:41 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>>> On 10/28/25 6:34 PM, Chris M. Thomasson wrote: >>>>>>>>>>>>>>>>>>>>>>>> On 10/28/2025 5:43 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>>>> [...]
    not my argument u fucking useless boomer >>>>>>>>>>>>>>>>>>>>>>>>
    Well, how do you solve the halting problem? >>>>>>>>>>>>>>>>>>>>>>>>

    with reflection, something u lack as a useless >>>>>>>>>>>>>>>>>>>>>>> fucking boomer


    Which you can't say how to do, because your only >>>>>>>>>>>>>>>>>>>>>> answer is to define that you can.

    we haven't even gotten the point of working out the >>>>>>>>>>>>>>>>>>>>> ramifications of what i've defined,

    Because you haven't actually defined what you are >>>>>>>>>>>>>>>>>>>> doing.

    blatant lie

    Really, show where you have ACTUALLY fully defined >>>>>>>>>>>>>>>>>> your computation machine to a level you can actually >>>>>>>>>>>>>>>>>> analyize one.

    TMs with a modification, specifically adding full >>>>>>>>>>>>>>>>> mechanical reflection to TMs in order to produce >>>>>>>>>>>>>>>>> Reflective TMs (RTMs)

    adding full mechanical reflection just means it gets a >>>>>>>>>>>>>>>>> single instruction that dumps *all* this information to >>>>>>>>>>>>>>>>> the tape, at the head, when that instruction is called: >>>>>>>>>>>>>>>>>
    1) machine description
    2) initial tape state
    3) current instruction
    4) current tape state

    call it REFLECT, if u will... this information allows a >>>>>>>>>>>>>>>>> computation to know exactly where it is in the overall >>>>>>>>>>>>>>>>> runtime of computation, and with that it can subvert >>>>>>>>>>>>>>>>> semantic paradoxes like the halting problem. >>>>>>>>>>>>>>>>>
    you have to understand this is a *mechanical* >>>>>>>>>>>>>>>>> modification to turing machines, not a computational >>>>>>>>>>>>>>>>> one. much like how RTMs can move a head left/right, or >>>>>>>>>>>>>>>>> write 0/1 to the tape ... they mechanically just do >>>>>>>>>>>>>>>>> this operation when the instruction is called. >>>>>>>>>>>>>>>>>
    *these mechanics are (relatively) simple and self- >>>>>>>>>>>>>>>>> evidentially possible and therefore mathematically >>>>>>>>>>>>>>>>> feasible.*

    BUT, let me try a step deeper, and quote turing at you >>>>>>>>>>>>>>>>> from his original paper on computable numbers: >>>>>>>>>>>>>>>>>
    /We may compare a man in the process of computing a >>>>>>>>>>>>>>>>> real number to a
    machine which is only capable of a finite number of >>>>>>>>>>>>>>>>> conditions/ [Tur36]

    THE ENTIRE JUSTIFICATION FOR TURING MACHINES RESTS >>>>>>>>>>>>>>>>> SQUARELY ON THE NOTION OF "CAN A MAN MECHANICALLY >>>>>>>>>>>>>>>>> UNDERTAKE THESE STATE TRANSITIONS"

    1) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>>>> then obviously he has access the full machine >>>>>>>>>>>>>>>>> description, or how else the fuck is he doing any of >>>>>>>>>>>>>>>>> these state transitions??? therefore he obviously can >>>>>>>>>>>>>>>>> dump the machine description, in whatever format he has >>>>>>>>>>>>>>>>> it in, to the tape when REFLECT is encountered. >>>>>>>>>>>>>>>>>
    2) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>>>> then obviously he can to the side save the initial tape >>>>>>>>>>>>>>>>> in a "buffer" that than cannot be overwritten. with >>>>>>>>>>>>>>>>> that he can can obviously dump the initial tape state >>>>>>>>>>>>>>>>> to the tape whenever REFLECT is encountered. >>>>>>>>>>>>>>>>>
    3) if a man is running thru the steps of a computation, >>>>>>>>>>>>>>>>> then obviously he has access input of a particular >>>>>>>>>>>>>>>>> transition, and therefore can dump that input to the >>>>>>>>>>>>>>>>> tape when REFLECT is encountered, in whatever format he >>>>>>>>>>>>>>>>> encounters it in.

    4) if a man is running thru the steps of a
    computation ... he can obviously duplicate the tape >>>>>>>>>>>>>>>>> when REFLECT is called. or more precisely he can save >>>>>>>>>>>>>>>>> it into a buffer, and dump from the buffer after >>>>>>>>>>>>>>>>> putting 1-3 on the tape.

    THAT IS THE SAME LEVEL OF FORMAT JUSTIFICATION TURING >>>>>>>>>>>>>>>>> ULTIMATELY GIVES FOR ANY COMPUTING MACHINE. SERIOUSLY >>>>>>>>>>>>>>>>> READ HIS FUCKING ESSAY YOU DORK

    please tell me where u are confused, not how u think >>>>>>>>>>>>>>>>> i'm wrong


    I can finally see the gist of what you are saying >>>>>>>>>>>>>>>> and it seems to be very similar to my idea.

    like i've said polcott, i consider you having a half-way >>>>>>>>>>>>>>> solution. u have something that's executable for sure, >>>>>>>>>>>>>>> and that is a definite improvement over total
    undecidability.

    but it's not totally epistemically useful, it still >>>>>>>>>>>>>>> doesn't / effectively compute/ the halting function in a >>>>>>>>>>>>>>> *general* manner.

    i'm proposing something that's both executable *and* >>>>>>>>>>>>>>> epistemically useful for that purpose.

    the RTM i'm proposing here is just the underlying >>>>>>>>>>>>>>> theoretical mechanisms that can be used to produce a >>>>>>>>>>>>>>> theoretically coherent context-dependent decider, >>>>>>>>>>>>>>>
    which can then be used to /effectively compute/ the >>>>>>>>>>>>>>> halting function irrespective of semantic paradoxes ... >>>>>>>>>>>>>>>
    something that is in total a step beyond the mere >>>>>>>>>>>>>>> executability that you are proposing.


    No, the problem you have is that you make the same error >>>>>>>>>>>>>> and try to redefine the question, and allow your machine >>>>>>>>>>>>>> to give wrong answers.

    it's not a "wrong" answer if it contains no information to >>>>>>>>>>>>> be wrong about, and it's only given in a situation where >>>>>>>>>>>>> there's no "right" answer.

    But their is a right answer for every actual D based on >>>>>>>>>>>> every actual H.

    It just isn't the answer that H gives.


    how is it reasonable to expect a decider to respond >>>>>>>>>>>>> truthfully to a situation that would only make that truth >>>>>>>>>>>>> untrue

    Because deciders are just deterministic processes with a >>>>>>>>>>>> fixed algorithm.

    right, and context can be an input to the determinism

    console.trace() isn't nondeterministic, it's context dependent >>>>>>>>>>>

    Note, the system DOES understand that some problems turn out >>>>>>>>>>>> to be uncomputable, so no decider exists that can be correct >>>>>>>>>>>> for all inputs. Nothing wrong with that, it just means that >>>>>>>>>>>> any decider that wants to try to claim to solve it will be >>>>>>>>>>>> wrong for some input.

    The key point here is the question can't be asked until >>>>>>>>>>>> after a decider steps up to make the claim, as the prove-it- >>>>>>>>>>>> wrong input is different for every possible decider, and >>>>>>>>>>>> thus, by the time the question can be asked, the deciders >>>>>>>>>>>> result for that input was already determined by its
    algorithm. For the decider it isn't a matter of being >>>>>>>>>>>> "truthful", it is just doing what it was programmed to do. >>>>>>>>>>>>
    It becomes a matter of correctness, not truthfulness.

    It is the creator of the algoritm to which truthfulness is >>>>>>>>>>>> applied, and claiming that it will give the right answer >>>>>>>>>>>> when it doesn't shows that his claim wasn't truthful.


    there's no "correct" answer to a pathological question of >>>>>>>>>>> halts(und) on this line:

    und = () -> halts(und) && loop_forever()

    Which isn't a computation if your symbology means that halts >>>>>>>>>> is taken from the context that the symbol und is used in.

    Sorry, you are just showing you don't know the meaning of the >>>>>>>>>> words.


    and the resolution is just to give a non-answer, so i split >>>>>>>>>>> the decider into two halves and made false a non-answer. it >>>>>>>>>>> cannot be a contradiction because it doesn't give
    information /in that moment/ about und() to contradict.

    But since you are starting with the same category error of >>>>>>>>>> Olcott, you get the same nonsense he does.

    The category for the input to a Halt Decider is the
    representation of a Computation.

    By DEFINITION, this has a FULLY defined algorithm, and thus >>>>>>>>>> NOT a closure.

    If you want to try to define your own theory of computation >>>>>>>>>> that uses a broader defintion, go ahead, the problem is you >>>>>>>>>> will find you can do litle with it, and that the semantic >>>>>>>>>> paradox you are complaining about is totally due to you
    altered definition of a computation.

    right why do u keep asserting ur existential fuckup is more >>>>>>>>> correct than a working decider on the matter???

    Because it doesn't give the right answer per the original problem. >>>>>>>
    it gives you the right answer in all the places that the
    information of that answer is actually usable, and it does so
    generally.

    In other words, your logic says it is ok to be wrong in some
    cases, if you can give a "good excuse" why it is ok to be wrong
    there.

    All that does is destroy the meaning of "Correct".


    you are holding onto some notion of wanting a "right" answer
    where no right answer can exist, and that's just bizarre because >>>>>>> with that desire u lose the ability to generally get a "right"
    answer in first place.

    Because it *DOES* have a proper definition.

    ALWAYS means ALWAYS, not just where I want it.

    i can't believe ur a chief engineer asking for correctness where

    a) it has no practical usecase

    b) it ends up neutering ur theoretical ability to generally handle
    the problem

    ur demanding to be able to shoot urself in ur foot over some
    decider that can't even exist.

    weird fetish bro

    I guess you just think that lies are valid, and truth doesn't matter.

    Note, a real halt decider, if it existed, would have a LOT of
    actually practical uses, and the fact it can't exist is important
    knowledge.




    ur throwing the fucking baby out with the bathwater dude,

    it's fucking stupid

    Nope, because for actual problems like Busy Beaver, which again
    asks about for results from ALL possible machihes with specific
    criteria, that pattern might just coencidentally occur.



    If you want to be clear that you aren't working on that problem, >>>>>>>> or even working under that system, go ahead, just make sure you >>>>>>>> fully define what you are doing and try to show where it is useful. >>>>>>>>

    i'm not building a separate form of computing, i'm upgrading what >>>>>>> we have

    In other words, you are admitting you are lying.

    It is a VIOLATION of the meaning of the words to change the
    meaning of the word as already used.

    You "upgrade" a system by adding new ideas without removing or
    changing any of the old.

    If you need to remove or change something, you have a new system. >>>>>> DEFINITION.

    like adding a new operation BY DEFINITION???

    Sure, it it can't be done within the limits of the previously
    defined system.


    honestly idk if ur just in denial, too senior to take criticism
    anymore, or just shoved a stick too far ur butthole,

    Nope, it is just sticking up for truth against stupid people who
    don't understand what they are talking about.


    but suggesting i'm not changing anything when i've spent pages
    trying to define a very particular change, even given it's a label
    RTM is just completely disingenuous of you.

    So, how does the machine write out information that it doesn't have
    stored?


    i'm insulted at the lack of consideration i'm getting here, clearly >>>>> u have no willingness to engaging in genuine discussion, cause ur
    just blatantly lying about my position.

    Then you are being insulted by your own errors.

    YOU are the one who has decided not to engage in answering question
    put to you to explain.


    how can u offer me any meaningful feedback if ur just gunna
    blatantly lie about my position???

    What LIE have I said?

    To show that what I have said can't be done, just needs a showing
    how you claim it actually could be done within the restrictions of
    the framework you are in.

    Part of your problem, is it seems you haven't bothered to learn what
    the frameworks actually is.



    Again, all you are doing is admitting that you system is based on >>>>>> you claiming the right to lie about what you are doing, by
    ignoring the rules of the system.





    yes this does mean maybe /false/ could be returned
    everywhere, but that's not fucking interesting ... nor does >>>>>>>>>>> it give / effective computability/ in regards to the halting >>>>>>>>>>> question.

    Since your input is a category error for the actual
    Computation Theory as classical defined, your results don't >>>>>>>>>> apply.

    classic computing theory doesn't have a halt decider defined >>>>>>>>> tho ... u only have a halting function, not a decider. the
    decider is *disproven* and therefore cannot be defined as per >>>>>>>>> consensus theory.

    u fucking retard.

    u keep asserting SOMETHING THAT DOESN'T EVEN EXIST is somehow >>>>>>>>> more correct than SOMETHING THAT CAN EXIST

    The DEFINITION of a Halting Decider exists.

    The fact that the definiton is not realizable, proves the theory. >>>>>>>>
    To call something a Halt Decider that doesn't meet the
    definition, is just a lie, making the person making the claim >>>>>>>> just a liar.

    just like there's more than one way to ask the same question,
    there's more than one way to define a halting decider, which is >>>>>>> the interface into an algorithm, not the algorithm itself.

    Right, but if they are semantically different, it isn't the same
    question.


    the one i'm proposing can actually generally exist, unlike the
    classical definition.

    But only is a Computation System that is broken, because programs >>>>>> don't reliable produce the same answer when run.

    Once you allow things to lie, you have destroyed the concept of
    truth.


    it's really that simple.

    Yes, allowing lying seems simple until you need to count on
    something.



    We can easily DEFINE things that don't exist.>


    which is the point of the context-dependent decider. halts() >>>>>>>>>>> is required to return /true/ only whereever both /possible/ >>>>>>>>>>> and / truthful/ to do so.

    So, if you want to try to talk about this, you first need to >>>>>>>>>> try to develop a basic theory about context dependent
    Computations, to see if it can actually be useful for anything. >>>>>>>>>
    i fucking already did. u just aren't reading cause ur too busy >>>>>>>>> trying to show off how big a fucking retard u can be.

    No you haven't, unless you consider your claim that is the
    equivalent of defining that 1 == 2 is a definition.

    DESCRIBING what you want to do is NOT the same as DEFINING how >>>>>>>> to do it.

    ur special pleading is just so mindnumbingly boring. i DEFINED my >>>>>>> mechanics in MORE detail that turing did his, and yet u
    arbitrarily reject it as "describing" which seems like a made up >>>>>>> distinction here.

    NO YOU HAVEN'T.

    You have PRESUMED an ability on the machine that can not actually >>>>>> be implemented on it without breaking it.

    i'm intelligent enough to write down the machine description to the >>>>> tape when i see the operation REFLECT, and therefore accept the
    defined ability

    But is it the RIGHT machine description for the program in question?

    And where does it get the contents of the original tape?


    i get that ur apparently too fucking retarded to do that, and
    therefore don't accept it.

    Nope, it seems you are too retarded to actually think about what
    other say about your system. You suffer from the fallacy that you
    assume you are right and thus ignore the definitions of what you are
    talking about.



    Your "Definition" is the equivalent of defining that for this
    work, I will consider 1 to be equal to 2.


    sometimes i really do feel like just buying a gun a shooting
    myself, cause i really am stuck on a planet of fucking idiots.

    y must u endlessly torture me so with ur fucking idiocracy?


    I could say the same to you. Why are you so stuck on insisting on >>>>>> things that are not.

    Computations, as defined, can't change their answers for the same >>>>>> input. PERIOD.

    To insist they can is just to admit you are just a lunatic.

    right, printing stack traces actually isn't part of computing.
    every time we step thru debugging we're doing theoretically
    indescribable magic, according to richard


    And what does that have to do with what I said?

    And yes, a routine looking at who called it is NOT part of
    "Computing" per computation theory, as it can't matter.

    It seems part of your problem is you don't actually know what you
    are talking about.

    Remember, Computing isn't directly about "Computers" as we currently
    know them, but is more fundamental.

    It seems you have forgotten that "Computing" as a term existed
    BEFORE the computers you want to talk about.


    blows my mind that u think a computing function that prints a stack
    trace does something which cannot be done on and explained by a
    turing machine

    Well, try to define a Turing Machinne that does that.

    Your first problem will be that Turing Machines don't have a stack to
    dump.


    i'm not sure y u think this mindnumbing gishgallop u shit out is
    worth a continued response to

    can u explain what my argument even is?

    or are you basically just a fucking chatbot keying off one sentence
    at a time?


    No,  but it seems you are just a idiot that refuses to learn the
    meaning of the words you are using.

    As I have pointed out several times, trying to relate a "Computation"
    per COmputation Theory to behavior of modern computers is just an error.

    well, that's just fucking retarded.

    do you honestly think the behavior exhibiting within modern computing somehow exists outside the theory of computing???


    Sure does, at least the way you are talking about it.

    Not that surprising when you remember that they didn't exist when the
    theory was developed.

    A modern computer is only within the standard theory of computing when
    it is treated as a whole, and its results are only dependent on the
    algorithm stored in the machine (which includes all the memory that is considered inside it) and the data provides from the outside as input.

    When you look internally and let information pass to the algorithm that
    isn't part of the input, it fails to be described by the thoery.

    Of course, the other view is that everything it looks at that was
    outside becomes the input, and thus the analysis must take it into account.

    Perhaps the more amazing part is that we haven't been able to come up
    with a "Computation" (a mechancaly generated mapping of an input to an
    output) that they can generate that can't also be done with the
    "primative" Turing Machine.

    They may be able to produce inconsistant results when doing their thing,
    but those just fail to be computations, whose definition requires
    consistancy.


    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Sun Nov 2 21:01:00 2025
    From Newsgroup: comp.theory

    On 11/2/25 8:13 PM, Richard Damon wrote:
    On 11/2/25 3:57 PM, Tristan Wibberley wrote:
    On 02/11/2025 19:10, Richard Damon wrote:
    So, how does the machine write out information that it doesn't have
    stored?

    A TM (at least typically) has a current machine state from a finite list
    of machine states, a reflecting extension can be defined to have a
    representation for each that it writes out upon interpreting a symbol at
    the tape head for each machine state (or for a subset of machine states,
    etc) instead of a simple move-left or move-right.

    But their isn't a unique representation for that, and what good is
    dumping the information in a format that the program doesn't expect.

    the info is dumped in the same format the machine description itself is
    coded in. obv the machine description can be made aware of it's own
    encoding format, so therefore it can make use of this info


    And dumping the code isn't the biggest problem, although it does mean
    such a Turing Macine no longer has the property of always being a computation under the operation of composition of sub-proggrams.

    real world programs already have this problem when sub-functions depend
    on global state

    turing machines can't inherently be context-free or we'd call them context-free grammars, eh???


    The bigger issue is recovering the inital tape as defined.

    it really isn't. demonstrating this with a pen and paper is simple: you
    write the initial tape on a separate peace of paper that then isn't part
    of the tape. this is self-evidently possible and really needs no further justification.

    tho you apparently suffer from an inability to envision something so
    simple. it's cray that ur apparently like a chief engineer or
    somethin??? what a fucking 🤡🌎



    For clarity I think dart200 is talking about an extended TM, extended to
    have one or more machine-state write-out transitions defined.

    And such an extended machine means it fails to have the essential
    property of a standard Turing Machine of all programs and subprograms designed for it are actual computations.

    That property is what make it a computation machine.>
    For additional clarity, "machine state" does not normally refer to the
    tape contents nor to the current tape-cell index but the state of the
    machine (like register values, for machines defined to have those). You
    can't include the tape cell index in the machine state because then it
    would have an infinite number of machine states. I understand that would
    not be a TM, strictly.

    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.


    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Sun Nov 2 21:03:53 2025
    From Newsgroup: comp.theory

    On 11/2/25 8:53 PM, Richard Damon wrote:
    On 11/2/25 11:42 PM, dart200 wrote:
    On 11/2/25 8:08 PM, Richard Damon wrote:
    On 11/2/25 3:45 PM, dart200 wrote:
    On 11/2/25 11:10 AM, Richard Damon wrote:
    On 11/1/25 12:37 PM, dart200 wrote:
    On 11/1/25 6:19 AM, Richard Damon wrote:
    On 11/1/25 2:18 AM, dart200 wrote:
    On 10/31/25 9:33 PM, Richard Damon wrote:
    On 10/31/25 11:52 PM, dart200 wrote:
    On 10/31/25 8:33 PM, Richard Damon wrote:
    On 10/31/25 11:16 PM, dart200 wrote:
    On 10/31/25 7:59 PM, Richard Damon wrote:
    On 10/31/25 3:29 PM, dart200 wrote:
    On 10/31/25 9:18 AM, Richard Damon wrote:
    On 10/30/25 2:27 PM, dart200 wrote:
    On 10/29/25 8:42 PM, olcott wrote:
    On 10/29/2025 10:23 PM, dart200 wrote:
    On 10/29/25 6:45 PM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>> On 10/29/25 12:24 PM, dart200 wrote:
    On 10/29/25 4:29 AM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>>>> On 10/28/25 11:23 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>> On 10/28/25 6:55 PM, Richard Damon wrote: >>>>>>>>>>>>>>>>>>>>>>> On 10/28/25 9:41 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>>>> On 10/28/25 6:34 PM, Chris M. Thomasson wrote: >>>>>>>>>>>>>>>>>>>>>>>>> On 10/28/2025 5:43 PM, dart200 wrote: >>>>>>>>>>>>>>>>>>>>>>>>> [...]
    not my argument u fucking useless boomer >>>>>>>>>>>>>>>>>>>>>>>>>
    Well, how do you solve the halting problem? >>>>>>>>>>>>>>>>>>>>>>>>>

    with reflection, something u lack as a useless >>>>>>>>>>>>>>>>>>>>>>>> fucking boomer


    Which you can't say how to do, because your only >>>>>>>>>>>>>>>>>>>>>>> answer is to define that you can. >>>>>>>>>>>>>>>>>>>>>>
    we haven't even gotten the point of working out >>>>>>>>>>>>>>>>>>>>>> the ramifications of what i've defined, >>>>>>>>>>>>>>>>>>>>>
    Because you haven't actually defined what you are >>>>>>>>>>>>>>>>>>>>> doing.

    blatant lie

    Really, show where you have ACTUALLY fully defined >>>>>>>>>>>>>>>>>>> your computation machine to a level you can actually >>>>>>>>>>>>>>>>>>> analyize one.

    TMs with a modification, specifically adding full >>>>>>>>>>>>>>>>>> mechanical reflection to TMs in order to produce >>>>>>>>>>>>>>>>>> Reflective TMs (RTMs)

    adding full mechanical reflection just means it gets a >>>>>>>>>>>>>>>>>> single instruction that dumps *all* this information >>>>>>>>>>>>>>>>>> to the tape, at the head, when that instruction is >>>>>>>>>>>>>>>>>> called:

    1) machine description
    2) initial tape state
    3) current instruction
    4) current tape state

    call it REFLECT, if u will... this information allows >>>>>>>>>>>>>>>>>> a computation to know exactly where it is in the >>>>>>>>>>>>>>>>>> overall runtime of computation, and with that it can >>>>>>>>>>>>>>>>>> subvert semantic paradoxes like the halting problem. >>>>>>>>>>>>>>>>>>
    you have to understand this is a *mechanical* >>>>>>>>>>>>>>>>>> modification to turing machines, not a computational >>>>>>>>>>>>>>>>>> one. much like how RTMs can move a head left/right, or >>>>>>>>>>>>>>>>>> write 0/1 to the tape ... they mechanically just do >>>>>>>>>>>>>>>>>> this operation when the instruction is called. >>>>>>>>>>>>>>>>>>
    *these mechanics are (relatively) simple and self- >>>>>>>>>>>>>>>>>> evidentially possible and therefore mathematically >>>>>>>>>>>>>>>>>> feasible.*

    BUT, let me try a step deeper, and quote turing at you >>>>>>>>>>>>>>>>>> from his original paper on computable numbers: >>>>>>>>>>>>>>>>>>
    /We may compare a man in the process of computing a >>>>>>>>>>>>>>>>>> real number to a
    machine which is only capable of a finite number of >>>>>>>>>>>>>>>>>> conditions/ [Tur36]

    THE ENTIRE JUSTIFICATION FOR TURING MACHINES RESTS >>>>>>>>>>>>>>>>>> SQUARELY ON THE NOTION OF "CAN A MAN MECHANICALLY >>>>>>>>>>>>>>>>>> UNDERTAKE THESE STATE TRANSITIONS"

    1) if a man is running thru the steps of a >>>>>>>>>>>>>>>>>> computation, then obviously he has access the full >>>>>>>>>>>>>>>>>> machine description, or how else the fuck is he doing >>>>>>>>>>>>>>>>>> any of these state transitions??? therefore he >>>>>>>>>>>>>>>>>> obviously can dump the machine description, in >>>>>>>>>>>>>>>>>> whatever format he has it in, to the tape when REFLECT >>>>>>>>>>>>>>>>>> is encountered.

    2) if a man is running thru the steps of a >>>>>>>>>>>>>>>>>> computation, then obviously he can to the side save >>>>>>>>>>>>>>>>>> the initial tape in a "buffer" that than cannot be >>>>>>>>>>>>>>>>>> overwritten. with that he can can obviously dump the >>>>>>>>>>>>>>>>>> initial tape state to the tape whenever REFLECT is >>>>>>>>>>>>>>>>>> encountered.

    3) if a man is running thru the steps of a >>>>>>>>>>>>>>>>>> computation, then obviously he has access input of a >>>>>>>>>>>>>>>>>> particular transition, and therefore can dump that >>>>>>>>>>>>>>>>>> input to the tape when REFLECT is encountered, in >>>>>>>>>>>>>>>>>> whatever format he encounters it in.

    4) if a man is running thru the steps of a >>>>>>>>>>>>>>>>>> computation ... he can obviously duplicate the tape >>>>>>>>>>>>>>>>>> when REFLECT is called. or more precisely he can save >>>>>>>>>>>>>>>>>> it into a buffer, and dump from the buffer after >>>>>>>>>>>>>>>>>> putting 1-3 on the tape.

    THAT IS THE SAME LEVEL OF FORMAT JUSTIFICATION TURING >>>>>>>>>>>>>>>>>> ULTIMATELY GIVES FOR ANY COMPUTING MACHINE. SERIOUSLY >>>>>>>>>>>>>>>>>> READ HIS FUCKING ESSAY YOU DORK

    please tell me where u are confused, not how u think >>>>>>>>>>>>>>>>>> i'm wrong


    I can finally see the gist of what you are saying >>>>>>>>>>>>>>>>> and it seems to be very similar to my idea.

    like i've said polcott, i consider you having a half-way >>>>>>>>>>>>>>>> solution. u have something that's executable for sure, >>>>>>>>>>>>>>>> and that is a definite improvement over total >>>>>>>>>>>>>>>> undecidability.

    but it's not totally epistemically useful, it still >>>>>>>>>>>>>>>> doesn't / effectively compute/ the halting function in a >>>>>>>>>>>>>>>> *general* manner.

    i'm proposing something that's both executable *and* >>>>>>>>>>>>>>>> epistemically useful for that purpose.

    the RTM i'm proposing here is just the underlying >>>>>>>>>>>>>>>> theoretical mechanisms that can be used to produce a >>>>>>>>>>>>>>>> theoretically coherent context-dependent decider, >>>>>>>>>>>>>>>>
    which can then be used to /effectively compute/ the >>>>>>>>>>>>>>>> halting function irrespective of semantic paradoxes ... >>>>>>>>>>>>>>>>
    something that is in total a step beyond the mere >>>>>>>>>>>>>>>> executability that you are proposing.


    No, the problem you have is that you make the same error >>>>>>>>>>>>>>> and try to redefine the question, and allow your machine >>>>>>>>>>>>>>> to give wrong answers.

    it's not a "wrong" answer if it contains no information to >>>>>>>>>>>>>> be wrong about, and it's only given in a situation where >>>>>>>>>>>>>> there's no "right" answer.

    But their is a right answer for every actual D based on >>>>>>>>>>>>> every actual H.

    It just isn't the answer that H gives.


    how is it reasonable to expect a decider to respond >>>>>>>>>>>>>> truthfully to a situation that would only make that truth >>>>>>>>>>>>>> untrue

    Because deciders are just deterministic processes with a >>>>>>>>>>>>> fixed algorithm.

    right, and context can be an input to the determinism

    console.trace() isn't nondeterministic, it's context dependent >>>>>>>>>>>>

    Note, the system DOES understand that some problems turn >>>>>>>>>>>>> out to be uncomputable, so no decider exists that can be >>>>>>>>>>>>> correct for all inputs. Nothing wrong with that, it just >>>>>>>>>>>>> means that any decider that wants to try to claim to solve >>>>>>>>>>>>> it will be wrong for some input.

    The key point here is the question can't be asked until >>>>>>>>>>>>> after a decider steps up to make the claim, as the prove- >>>>>>>>>>>>> it- wrong input is different for every possible decider, >>>>>>>>>>>>> and thus, by the time the question can be asked, the >>>>>>>>>>>>> deciders result for that input was already determined by >>>>>>>>>>>>> its algorithm. For the decider it isn't a matter of being >>>>>>>>>>>>> "truthful", it is just doing what it was programmed to do. >>>>>>>>>>>>>
    It becomes a matter of correctness, not truthfulness. >>>>>>>>>>>>>
    It is the creator of the algoritm to which truthfulness is >>>>>>>>>>>>> applied, and claiming that it will give the right answer >>>>>>>>>>>>> when it doesn't shows that his claim wasn't truthful. >>>>>>>>>>>>>

    there's no "correct" answer to a pathological question of >>>>>>>>>>>> halts(und) on this line:

    und = () -> halts(und) && loop_forever()

    Which isn't a computation if your symbology means that halts >>>>>>>>>>> is taken from the context that the symbol und is used in. >>>>>>>>>>>
    Sorry, you are just showing you don't know the meaning of the >>>>>>>>>>> words.


    and the resolution is just to give a non-answer, so i split >>>>>>>>>>>> the decider into two halves and made false a non-answer. it >>>>>>>>>>>> cannot be a contradiction because it doesn't give
    information /in that moment/ about und() to contradict. >>>>>>>>>>>
    But since you are starting with the same category error of >>>>>>>>>>> Olcott, you get the same nonsense he does.

    The category for the input to a Halt Decider is the
    representation of a Computation.

    By DEFINITION, this has a FULLY defined algorithm, and thus >>>>>>>>>>> NOT a closure.

    If you want to try to define your own theory of computation >>>>>>>>>>> that uses a broader defintion, go ahead, the problem is you >>>>>>>>>>> will find you can do litle with it, and that the semantic >>>>>>>>>>> paradox you are complaining about is totally due to you >>>>>>>>>>> altered definition of a computation.

    right why do u keep asserting ur existential fuckup is more >>>>>>>>>> correct than a working decider on the matter???

    Because it doesn't give the right answer per the original problem. >>>>>>>>
    it gives you the right answer in all the places that the
    information of that answer is actually usable, and it does so >>>>>>>> generally.

    In other words, your logic says it is ok to be wrong in some
    cases, if you can give a "good excuse" why it is ok to be wrong >>>>>>> there.

    All that does is destroy the meaning of "Correct".


    you are holding onto some notion of wanting a "right" answer
    where no right answer can exist, and that's just bizarre because >>>>>>>> with that desire u lose the ability to generally get a "right" >>>>>>>> answer in first place.

    Because it *DOES* have a proper definition.

    ALWAYS means ALWAYS, not just where I want it.

    i can't believe ur a chief engineer asking for correctness where

    a) it has no practical usecase

    b) it ends up neutering ur theoretical ability to generally handle >>>>>> the problem

    ur demanding to be able to shoot urself in ur foot over some
    decider that can't even exist.

    weird fetish bro

    I guess you just think that lies are valid, and truth doesn't matter. >>>>>
    Note, a real halt decider, if it existed, would have a LOT of
    actually practical uses, and the fact it can't exist is important
    knowledge.




    ur throwing the fucking baby out with the bathwater dude,

    it's fucking stupid

    Nope, because for actual problems like Busy Beaver, which again >>>>>>> asks about for results from ALL possible machihes with specific >>>>>>> criteria, that pattern might just coencidentally occur.



    If you want to be clear that you aren't working on that
    problem, or even working under that system, go ahead, just make >>>>>>>>> sure you fully define what you are doing and try to show where >>>>>>>>> it is useful.


    i'm not building a separate form of computing, i'm upgrading
    what we have

    In other words, you are admitting you are lying.

    It is a VIOLATION of the meaning of the words to change the
    meaning of the word as already used.

    You "upgrade" a system by adding new ideas without removing or
    changing any of the old.

    If you need to remove or change something, you have a new system. >>>>>>> DEFINITION.

    like adding a new operation BY DEFINITION???

    Sure, it it can't be done within the limits of the previously
    defined system.


    honestly idk if ur just in denial, too senior to take criticism
    anymore, or just shoved a stick too far ur butthole,

    Nope, it is just sticking up for truth against stupid people who
    don't understand what they are talking about.


    but suggesting i'm not changing anything when i've spent pages
    trying to define a very particular change, even given it's a label >>>>>> RTM is just completely disingenuous of you.

    So, how does the machine write out information that it doesn't have >>>>> stored?


    i'm insulted at the lack of consideration i'm getting here,
    clearly u have no willingness to engaging in genuine discussion,
    cause ur just blatantly lying about my position.

    Then you are being insulted by your own errors.

    YOU are the one who has decided not to engage in answering question >>>>> put to you to explain.


    how can u offer me any meaningful feedback if ur just gunna
    blatantly lie about my position???

    What LIE have I said?

    To show that what I have said can't be done, just needs a showing
    how you claim it actually could be done within the restrictions of
    the framework you are in.

    Part of your problem, is it seems you haven't bothered to learn
    what the frameworks actually is.



    Again, all you are doing is admitting that you system is based on >>>>>>> you claiming the right to lie about what you are doing, by
    ignoring the rules of the system.





    yes this does mean maybe /false/ could be returned
    everywhere, but that's not fucking interesting ... nor does >>>>>>>>>>>> it give / effective computability/ in regards to the halting >>>>>>>>>>>> question.

    Since your input is a category error for the actual
    Computation Theory as classical defined, your results don't >>>>>>>>>>> apply.

    classic computing theory doesn't have a halt decider defined >>>>>>>>>> tho ... u only have a halting function, not a decider. the >>>>>>>>>> decider is *disproven* and therefore cannot be defined as per >>>>>>>>>> consensus theory.

    u fucking retard.

    u keep asserting SOMETHING THAT DOESN'T EVEN EXIST is somehow >>>>>>>>>> more correct than SOMETHING THAT CAN EXIST

    The DEFINITION of a Halting Decider exists.

    The fact that the definiton is not realizable, proves the theory. >>>>>>>>>
    To call something a Halt Decider that doesn't meet the
    definition, is just a lie, making the person making the claim >>>>>>>>> just a liar.

    just like there's more than one way to ask the same question, >>>>>>>> there's more than one way to define a halting decider, which is >>>>>>>> the interface into an algorithm, not the algorithm itself.

    Right, but if they are semantically different, it isn't the same >>>>>>> question.


    the one i'm proposing can actually generally exist, unlike the >>>>>>>> classical definition.

    But only is a Computation System that is broken, because programs >>>>>>> don't reliable produce the same answer when run.

    Once you allow things to lie, you have destroyed the concept of >>>>>>> truth.


    it's really that simple.

    Yes, allowing lying seems simple until you need to count on
    something.



    We can easily DEFINE things that don't exist.>


    which is the point of the context-dependent decider. halts() >>>>>>>>>>>> is required to return /true/ only whereever both /possible/ >>>>>>>>>>>> and / truthful/ to do so.

    So, if you want to try to talk about this, you first need to >>>>>>>>>>> try to develop a basic theory about context dependent
    Computations, to see if it can actually be useful for anything. >>>>>>>>>>
    i fucking already did. u just aren't reading cause ur too busy >>>>>>>>>> trying to show off how big a fucking retard u can be.

    No you haven't, unless you consider your claim that is the
    equivalent of defining that 1 == 2 is a definition.

    DESCRIBING what you want to do is NOT the same as DEFINING how >>>>>>>>> to do it.

    ur special pleading is just so mindnumbingly boring. i DEFINED >>>>>>>> my mechanics in MORE detail that turing did his, and yet u
    arbitrarily reject it as "describing" which seems like a made up >>>>>>>> distinction here.

    NO YOU HAVEN'T.

    You have PRESUMED an ability on the machine that can not actually >>>>>>> be implemented on it without breaking it.

    i'm intelligent enough to write down the machine description to
    the tape when i see the operation REFLECT, and therefore accept
    the defined ability

    But is it the RIGHT machine description for the program in question? >>>>>
    And where does it get the contents of the original tape?


    i get that ur apparently too fucking retarded to do that, and
    therefore don't accept it.

    Nope, it seems you are too retarded to actually think about what
    other say about your system. You suffer from the fallacy that you
    assume you are right and thus ignore the definitions of what you
    are talking about.



    Your "Definition" is the equivalent of defining that for this
    work, I will consider 1 to be equal to 2.


    sometimes i really do feel like just buying a gun a shooting
    myself, cause i really am stuck on a planet of fucking idiots. >>>>>>>>
    y must u endlessly torture me so with ur fucking idiocracy?


    I could say the same to you. Why are you so stuck on insisting on >>>>>>> things that are not.

    Computations, as defined, can't change their answers for the same >>>>>>> input. PERIOD.

    To insist they can is just to admit you are just a lunatic.

    right, printing stack traces actually isn't part of computing.
    every time we step thru debugging we're doing theoretically
    indescribable magic, according to richard


    And what does that have to do with what I said?

    And yes, a routine looking at who called it is NOT part of
    "Computing" per computation theory, as it can't matter.

    It seems part of your problem is you don't actually know what you
    are talking about.

    Remember, Computing isn't directly about "Computers" as we
    currently know them, but is more fundamental.

    It seems you have forgotten that "Computing" as a term existed
    BEFORE the computers you want to talk about.


    blows my mind that u think a computing function that prints a stack
    trace does something which cannot be done on and explained by a
    turing machine

    Well, try to define a Turing Machinne that does that.

    Your first problem will be that Turing Machines don't have a stack to
    dump.


    i'm not sure y u think this mindnumbing gishgallop u shit out is
    worth a continued response to

    can u explain what my argument even is?

    or are you basically just a fucking chatbot keying off one sentence
    at a time?


    No,  but it seems you are just a idiot that refuses to learn the
    meaning of the words you are using.

    As I have pointed out several times, trying to relate a "Computation"
    per COmputation Theory to behavior of modern computers is just an error.

    well, that's just fucking retarded.

    do you honestly think the behavior exhibiting within modern computing
    somehow exists outside the theory of computing???


    Sure does, at least the way you are talking about it.

    Not that surprising when you remember that they didn't exist when the
    theory was developed.


    great! we should probably look into updating our theory according, eh???
    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Kaz Kylheku@[email protected] to comp.theory on Mon Nov 3 05:35:27 2025
    From Newsgroup: comp.theory

    On 2025-11-03, dart200 <[email protected]d> wrote:
    On 11/1/25 1:26 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 11/1/25 9:37 AM, Kaz Kylheku wrote:
    The result of combining partial deciders will always be a decider
    that can be shown not to be total.

    The can is indeed always kicked down the road.

    great! so ur own answered question of why i'm not particularly excited
    by partial deciders, and want a way to /effectively compute/ whether an input machine halts not

    That's mainly just to lay to rest the stuff about paradoxes.

    There is a real can, a real road, and you can kick one farther
    down the other.

    do we need to have this discussion again kaz, or???

    Not unless you're a retired coding technician with anterograde amnesia,
    and signs of dementia.
    --
    TXR Programming Language: http://nongnu.org/txr
    Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
    Mastodon: @[email protected]
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Sun Nov 2 22:01:40 2025
    From Newsgroup: comp.theory

    On 11/2/25 9:35 PM, Kaz Kylheku wrote:
    On 2025-11-03, dart200 <[email protected]d> wrote:
    On 11/1/25 1:26 PM, Kaz Kylheku wrote:
    On 2025-11-01, dart200 <[email protected]d> wrote:
    On 11/1/25 9:37 AM, Kaz Kylheku wrote:
    The result of combining partial deciders will always be a decider
    that can be shown not to be total.

    The can is indeed always kicked down the road.

    great! so ur own answered question of why i'm not particularly excited
    by partial deciders, and want a way to /effectively compute/ whether an
    input machine halts not

    That's mainly just to lay to rest the stuff about paradoxes.

    There is a real can, a real road, and you can kick one farther
    down the other.

    my proposal is not kicking the can down the road. it's not attempting to resolve the problem with TMs, that problem will simply remain.

    it's solving the problem within computing theory more generally by
    proposing an update to TMs adding full reflection, such it can robustly
    handle what previously would have been a semantic paradox.


    do we need to have this discussion again kaz, or???

    Not unless you're a retired coding technician with anterograde amnesia,
    and signs of dementia.


    awesome, and since ur not retarded richard, i won't hear any more of
    this "partial deciders are fine" non-sense from you, eh???
    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Tristan Wibberley@[email protected] to comp.theory on Mon Nov 3 17:59:30 2025
    From Newsgroup: comp.theory

    On 03/11/2025 04:22, Richard Damon wrote:
    On 11/2/25 10:29 PM, dart200 wrote:

    [I seem to be doing more than my fair share of snipping]

    ... RTMs themselves don't just then solve the halting
    problem.
    ^^^^^^^^^^

    "Don't just" is normally reserved here to trigger an inference in the
    message interpreter that the message emitter thinks they do at least as
    much. O.o


    ... RTMs allow for context-aware computing, which we can use to
    propose a context-aware decider, that can proactively avoid succumbing
    to semantic paradoxes

    But they CAN'T solve the Halting Problem as defined, as the input for
    the problem is to be a computation, and the decider needs to be a
    computation that always gives the right answer.

    It's still a computation by the word's meaning in computation theory but
    indeed they can't be programmed with a universal halt decider as far as
    I understand the theory. I think that's conventional and undisputed by
    Olcott, at least.

    It looks like dart200 thinks they /can/ be so programmed!

    [I seem to be doing more than my fair share of snipping]

    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From olcott@[email protected] to comp.theory on Mon Nov 3 12:05:52 2025
    From Newsgroup: comp.theory

    On 11/3/2025 11:59 AM, Tristan Wibberley wrote:
    On 03/11/2025 04:22, Richard Damon wrote:
    On 11/2/25 10:29 PM, dart200 wrote:

    [I seem to be doing more than my fair share of snipping]

    ... RTMs themselves don't just then solve the halting
    problem.
    ^^^^^^^^^^

    "Don't just" is normally reserved here to trigger an inference in the
    message interpreter that the message emitter thinks they do at least as
    much. O.o


    ... RTMs allow for context-aware computing, which we can use to
    propose a context-aware decider, that can proactively avoid succumbing
    to semantic paradoxes

    But they CAN'T solve the Halting Problem as defined, as the input for
    the problem is to be a computation, and the decider needs to be a
    computation that always gives the right answer.

    It's still a computation by the word's meaning in computation theory but indeed they can't be programmed with a universal halt decider as far as
    I understand the theory. I think that's conventional and undisputed by Olcott, at least.

    It looks like dart200 thinks they /can/ be so programmed!

    [I seem to be doing more than my fair share of snipping]



    The halting problem proofs utterly fail when their basis
    is restrained to remain inside the scope of the definition
    Turing machines that only compute the mapping from their
    input finite strings on the basis of what these actual
    input finite strings actually specify.

    LLMs have no trouble proving that they understand single
    sentences that are a whole paragraph, such as above.

    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --
    Copyright 2025 Olcott "Talent hits a target no one else can hit; Genius
    hits a target no one else can see." Arthur Schopenhauer
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Tristan Wibberley@[email protected] to comp.theory on Mon Nov 3 18:17:21 2025
    From Newsgroup: comp.theory

    On 03/11/2025 01:05, dart200 wrote:
    what i'm proposing obviously will look very much like TMs in many if not
    most computatation. in fact: any computation can run on a TM will in
    fact be run the same on an RTM.

    An RTM is an example of a TM.

    ITYM "any computation that can run on a TM will in fact be run the same
    on its corresponding RTMs." But that's more of a constraint on the
    meaning of "RTM" and "corresponding" so I would check the definition of
    RTM carefully before using that alternative wording. In fact I would
    expect it to be found overly specific. Maybe "any computation that can
    run on a TM will in fact be run the same on those RTMs that extend it."
    but that really just defines a concept I might ascribe as a meaning for "extend".


    the differences in RTMs only come up in computations that aren't
    directly expressible on TMs

    ITYM "aren't directly expressible for other TMs".


    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Tristan Wibberley@[email protected] to comp.theory on Mon Nov 3 18:32:58 2025
    From Newsgroup: comp.theory

    On 11/1/25 1:26 PM, Kaz Kylheku wrote:
    The result of combining partial deciders will always be a decider
    that can be shown not to be total.

    The can is indeed always kicked down the road.

    ...

    On 03/11/2025 06:01, dart200 wrote, after some awkward gaming between the
    two that made for yet more uncomfortable reading:

    my proposal is not kicking the can down the road. it's not attempting to resolve the problem with TMs, that problem will simply remain.

    it's solving the problem within computing theory more generally by
    proposing an update to TMs adding full reflection, such it can robustly handle what previously would have been a semantic paradox.

    I think its uncontroversial that doing so simply defines a mapping from
    an RTM and its programs to a TM and /its/ programs. That is, every
    program for an RTM can be translated to a program for a non-reflecting
    TM. The new program is an emulator (in modern industry parlance) of the
    RTM that runs on the non-reflecting TM and which is pre-paired with the
    program to be executed under simulation.


    do we need to have this discussion again kaz, or???

    Not unless you're a retired coding technician with anterograde amnesia,
    and signs of dementia.

    Hey there!


    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Mon Nov 3 11:10:11 2025
    From Newsgroup: comp.theory

    On 11/3/25 9:59 AM, Tristan Wibberley wrote:
    On 03/11/2025 04:22, Richard Damon wrote:
    On 11/2/25 10:29 PM, dart200 wrote:

    [I seem to be doing more than my fair share of snipping]

    ... RTMs themselves don't just then solve the halting
    problem.
    ^^^^^^^^^^

    "Don't just" is normally reserved here to trigger an inference in the
    message interpreter that the message emitter thinks they do at least as
    much. O.o

    well, at a fundamental level they allow access to the information
    required, but if u take the original halting spec and just run it on
    RTMs, that's not enough.

    you need to then use this new information access to form a new spec for
    the halting decider ... which then can transcend the halting problem to
    form a universal halting decider.


    ... RTMs allow for context-aware computing, which we can use to
    propose a context-aware decider, that can proactively avoid succumbing
    to semantic paradoxes

    But they CAN'T solve the Halting Problem as defined, as the input for
    the problem is to be a computation, and the decider needs to be a
    computation that always gives the right answer.

    It's still a computation by the word's meaning in computation theory but indeed they can't be programmed with a universal halt decider as far as
    I understand the theory. I think that's conventional and undisputed by Olcott, at least.

    It looks like dart200 thinks they /can/ be so programmed!

    correct, me and polcott have differing proposals,

    and from my perspective polcott has a half solution.


    [I seem to be doing more than my fair share of snipping]

    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Mon Nov 3 11:38:23 2025
    From Newsgroup: comp.theory

    On 11/3/25 10:17 AM, Tristan Wibberley wrote:
    On 03/11/2025 01:05, dart200 wrote:
    what i'm proposing obviously will look very much like TMs in many if not
    most computatation. in fact: any computation can run on a TM will in
    fact be run the same on an RTM.

    An RTM is an example of a TM.

    *extension* of

    the fact we're baking TM computable mechanics into the machine itself vs constructing it in the machine description being run by the machine, has theoretical ramifications that we: (a) cannot ignore and (b) have yet to really acknowledge


    ITYM "any computation that can run on a TM will in fact be run the same
    on its corresponding RTMs." But that's more of a constraint on the
    meaning of "RTM" and "corresponding" so I would check the definition of
    RTM carefully before using that alternative wording. In fact I would
    expect it to be found overly specific. Maybe "any computation that can
    run on a TM will in fact be run the same on those RTMs that extend it."
    but that really just defines a concept I might ascribe as a meaning for "extend".

    let put me try to make this clear:

    if you have a machine description that /only/ uses the subset of
    operations found on a TM (HEAD_LEFT, HEAD_RIGHT, WRITE_SYMBOL), then it
    will run *exactly* the same on an RTM as a TM.

    the novel capabilities of RTMs only arise when we start using RELFECT in
    our computations, allowing us the express semantic deciders that are
    robust against semantic paradoxes.

    what this ultimately means is these semantic paradoxes become
    un-expressable within the framework of RTMs, despite the fact semantic deciders exist and can be used in a self-referential manner.

    you can try to express them, sure, but they don't actually create an undecidable paradox.



    the differences in RTMs only come up in computations that aren't
    directly expressible on TMs

    ITYM "aren't directly expressible for other TMs".

    computations involving REFLECT cannot be directly run on TMs

    TMs can simulate RTMs but they don't gain theoretical power, as without REFLECT built into the machine itself, we can cannot theoretically block
    the expression of semantic paradoxes with semantic deciders.
    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Tristan Wibberley@[email protected] to comp.theory on Mon Nov 3 20:04:18 2025
    From Newsgroup: comp.theory

    On 03/11/2025 19:38, dart200 wrote:
    On 11/3/25 10:17 AM, Tristan Wibberley wrote:
    An RTM is an example of a TM.

    *extension* of

    both, isn't it? Or is there some restriction on TMs that precludes
    reflection? Each state of an RTM can have a symbol to be written at the
    head, I'm sure you can encode a TM description that has that reflection.

    Is it than the class "RTM" refers--in part--to TMs that complete a
    sequence of writes and moves (a stored macro?) upon reading a special
    symbol and that the presence of such a reduction rule excludes the
    machine from the TMs?


    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Tristan Wibberley@[email protected] to comp.theory on Mon Nov 3 20:07:00 2025
    From Newsgroup: comp.theory

    On 03/11/2025 19:38, dart200 wrote:
    if you have a machine description that /only/ uses the subset of
    operations found on a TM (HEAD_LEFT, HEAD_RIGHT, WRITE_SYMBOL), then it
    will run *exactly* the same on an RTM as a TM.

    My understanding of the term "TM" is that it's short for "Turing
    Machine" which refers to a class of machines including those with other operations in addition to those you mention. Is that an inaccurate understanding?

    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From dart200@[email protected] to comp.theory on Mon Nov 3 12:26:57 2025
    From Newsgroup: comp.theory

    On 11/3/25 12:07 PM, Tristan Wibberley wrote:
    On 03/11/2025 19:38, dart200 wrote:
    if you have a machine description that /only/ uses the subset of
    operations found on a TM (HEAD_LEFT, HEAD_RIGHT, WRITE_SYMBOL), then it
    will run *exactly* the same on an RTM as a TM.

    My understanding of the term "TM" is that it's short for "Turing
    Machine" which refers to a class of machines including those with other operations in addition to those you mention. Is that an inaccurate understanding?

    turing machines strictly only have moving the head and writing one
    symbol to the tape as available operations.

    RTMs dont solve TM halting problems because RTMs cant then ensure
    reflection is always available and accurate in TM computations.

    TMs can simulate RTMs just fine, but because that simulation
    infrastructure isn't built into TMs directly, the results of this
    simulation do not improve TM computations generally.

    what i'm proposing is that RTMs trivialize TM halting problems because
    RTMs can express everything TMs can, plus they can express novel
    computations necessarily involving reflection like general semantic
    analysis (including halting analysis)


    --
    Tristan Wibberley

    The message body is Copyright (C) 2025 Tristan Wibberley except
    citations and quotations noted. All Rights Reserved except that you may,
    of course, cite it academically giving credit to me, distribute it
    verbatim as part of a usenet system or its archives, and use it to
    promote my greatness and general superiority without misrepresentation
    of my opinions other than my opinion of my greatness and general
    superiority which you _may_ misrepresent. You definitely MAY NOT train
    any production AI system with it but you may train experimental AI that
    will only be used for evaluation of the AI methods it implements.

    --
    a burnt out swe investigating into why our tooling doesn't involve
    basic semantic proofs like halting analysis

    please excuse my pseudo-pyscript,

    ~ nick
    --- Synchronet 3.21a-Linux NewsLink 1.2