Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Re: Does everyone agree with this halt status decision? [ Richard cannot even compute 8th grade arithmetic ]

8 views
Skip to first unread message

olcott

unread,
Sep 4, 2022, 9:42:12 AM9/4/22
to
On 9/4/2022 5:56 AM, Richard Damon wrote:
> On 9/4/22 3:08 AM, Skep Dick wrote:
>> On Saturday, 3 September 2022 at 19:24:40 UTC+2, richar...@gmail.com
>> wrote:
>>> On 9/3/22 1:18 PM, olcott wrote:
>>>> On 9/3/2022 12:08 PM, Richard Damon wrote:
>>>>> On 9/3/22 12:56 PM, olcott wrote:
>>>>>> On 9/3/2022 11:47 AM, Richard Damon wrote:
>>>>>>> On 9/3/22 12:32 PM, olcott wrote:
>>>>>>>> On 9/3/2022 11:14 AM, Richard Damon wrote:
>>>>>>>>>
>>>>>>>>> On 9/3/22 12:05 PM, olcott wrote:
>>>>>>>>>> On 9/3/2022 11:00 AM, Richard Damon wrote:
>>>>>>>>>>> On 9/3/22 11:53 AM, olcott wrote:
>>>>>>>>>>>> On 9/3/2022 10:31 AM, Richard Damon wrote:
>>>>>>>>>>>>> On 9/3/22 11:20 AM, olcott wrote:
>>>>>>>>>>>>>> On 9/3/2022 9:33 AM, Richard Damon wrote:
>>>>>>>>>>>>>>> On 9/3/22 10:28 AM, olcott wrote:
>>>>>>>>>>>>>>>> On 9/3/2022 9:26 AM, Richard Damon wrote:
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> On 9/3/22 10:14 AM, olcott wrote:
>>>>>>>>>>>>>>>>>> On 9/3/2022 9:06 AM, Paul N wrote:
>>>>>>>>>>>>>>>>>>> On Saturday, September 3, 2022 at 2:03:48 PM UTC+1,
>>>>>>>>>>>>>>>>>>> olcott wrote:
>>>>>>>>>>>>>>>>>>>> On 9/3/2022 7:50 AM, Paul N wrote:
>>>>>>>>>>>>>>>>>>>>> On Saturday, September 3, 2022 at 1:43:43 PM UTC+1,
>>>>>>>>>>>>>>>>>>>>> olcott wrote:
>>>>>>>>>>>>>>>>>>>>>> On 9/3/2022 7:00 AM, Paul N wrote:
>>>>>>>>>>>>>>>>>>>>>>> On Saturday, September 3, 2022 at 2:54:54 AM UTC+1,
>>>>>>>>>>>>>>>>>>>>>>> olcott wrote:
>>>>>>>>>>>>>>>>>>>>>>>> On 9/2/2022 7:46 PM, Richard Damon wrote:
>>>>>>>>>>>>>>>>>>>>>>>>> If you won't go by the oficial definitions, your H
>>>>>>>>>>>>>>>>>>>>>>>>> just isn't a Halt
>>>>>>>>>>>>>>>>>>>>>>>>> Decider.
>>>>>>>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>>>>>> If the official definitions are self-contradictory
>>>>>>>>>>>>>>>>>>>>>>>> then at least one of
>>>>>>>>>>>>>>>>>>>>>>>> them must be rejected as incorrect.
>>>>>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>>>>> No, you are wrong here. The official definition of a
>>>>>>>>>>>>>>>>>>>>>>> halt decider is self-contradictory, in that it is
>>>>>>>>>>>>>>>>>>>>>>> impossible to build anything which meets the
>>>>>>>>>>>>>>>>>>>>>>> definition. This has been proved.
>>>>>>>>>>>>>>>>>>>>>> That is not the way truth works mate. If the official
>>>>>>>>>>>>>>>>>>>>>> definitions are
>>>>>>>>>>>>>>>>>>>>>> self-contradictory then I have also had new insight
>>>>>>>>>>>>>>>>>>>>>> on that another
>>>>>>>>>>>>>>>>>>>>>> aspect of computer science is incorrect.
>>>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>>> So you are not disputing that official definition of a
>>>>>>>>>>>>>>>>>>>>> halt decider is self-contradictory?
>>>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>> The correct and complete simulation of an input is
>>>>>>>>>>>>>>>>>>>> guaranteed to derive
>>>>>>>>>>>>>>>>>>>> the actual behavior of this input.
>>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>> When-so-ever a simulating halt decider (SHD) must abort
>>>>>>>>>>>>>>>>>>>> the simulation
>>>>>>>>>>>>>>>>>>>> of its input to prevent the infinite execution of this
>>>>>>>>>>>>>>>>>>>> input is merely
>>>>>>>>>>>>>>>>>>>> another way of saying the the correct and complete
>>>>>>>>>>>>>>>>>>>> simulation of the
>>>>>>>>>>>>>>>>>>>> input by this SHD would never stop running.
>>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>> When computer science textbooks say that the behavior
>>>>>>>>>>>>>>>>>>>> that a halt
>>>>>>>>>>>>>>>>>>>> decider must report on is the behavior of the direct
>>>>>>>>>>>>>>>>>>>> execution of the
>>>>>>>>>>>>>>>>>>>> machine represented by this input and this behavior is
>>>>>>>>>>>>>>>>>>>> not the same as
>>>>>>>>>>>>>>>>>>>> the correct and complete simulation of this input then
>>>>>>>>>>>>>>>>>>>> the computer
>>>>>>>>>>>>>>>>>>>> science textbooks are wrong because they reject the
>>>>>>>>>>>>>>>>>>>> definition of a UTM.
>>>>>>>>>>>>>>>>>>>> (simulator).
>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>> I was hoping for more of a yes-or-no type answer.
>>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>> When I provide a complete explanation of the reasoning
>>>>>>>>>>>>>>>>>> behind the answer all those not understanding these
>>>>>>>>>>>>>>>>>> things cannot simply baselessly disagree. All
>>>>>>>>>>>>>>>>>> disagreement is thus required to have a basis. This way
>>>>>>>>>>>>>>>>>> people are not lead astray by a bunch of baseless
>>>>>>>>>>>>>>>>>> disagreement, they see that the disagreement has no basis
>>>>>>>>>>>>>>>>>> and reject it.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> So, you need to try to actually provide VALID reasoning,
>>>>>>>>>>>>>>>>> starting from ACCEPTED definition.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> You just make wild claims based on your own understanding
>>>>>>>>>>>>>>>>> of the words.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Since you don't even attempt to actually DEFINE the words,
>>>>>>>>>>>>>>>>> no one can help you with your logic.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Note, Most of the things you are talking about HAVE well
>>>>>>>>>>>>>>>>> defined meanings, so you either need to use that meaning,
>>>>>>>>>>>>>>>>> or you really need to give your idea a new name (maybe
>>>>>>>>>>>>>>>>> basd on the common name with a modifier).
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Thus, if you definition of "Halting" isn't the same as the
>>>>>>>>>>>>>>>>> accepted definition, you can call it PO-Halting.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> Of course the problem with this is it makes it clear that
>>>>>>>>>>>>>>>>> you are talking about the PO-Halting Problem, with a
>>>>>>>>>>>>>>>>> PO-Halting Decider, so you counter example doesn't affect
>>>>>>>>>>>>>>>>> the actual Halting Problem, so it is clear you haven't
>>>>>>>>>>>>>>>>> proven what you claim to have proven.
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>>>>> It's not immediately obvious, but it can be proved
>>>>>>>>>>>>>>>>>>>>> that the definition is self-contradictory. You're not
>>>>>>>>>>>>>>>>>>>>> disputing that proof?
>>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>>> You haven't actually given a proof, since you are clearly
>>>>>>>>>>>>>>>>> using altered definition that you aren't providing.
>>>>>>>>>>>>>>>> You reject the notion of a UTM so you are starting with an
>>>>>>>>>>>>>>>> incorrect basis.
>>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>>
>>>>>>>>>>>>>>> I have done no such thing. UTM's exist, and UTM(P,d) by
>>>>>>>>>>>>>>> DEFINITION gives the exact same result as P(d).
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> When you get a glass of water and dump it on the floor you
>>>>>>>>>>>>>> cannot truthfully deny that the floor is wet.
>>>>>>>>>>>>>
>>>>>>>>>>>>> And I haven't
>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> When the correct and complete simulation of the input by
>>>>>>>>>>>>>> H(P,P) would have different behavior than the direct
>>>>>>>>>>>>>> execution of P(P) this cannot simply be ignored.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> And the correct and complete simulation of the input to H(P,P)
>>>>>>>>>>>>> Halts.
>>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> The fact that the simulation by H(P,P) of its input never stops
>>>>>>>>>>>> running unless H aborts the simulation of this input is merely
>>>>>>>>>>>> another way of saying that the correct and complete simulation
>>>>>>>>>>>> of this input by H never halts.
>>>>>>>>>>>
>>>>>>>>>>> You have bad logic there.
>>>>>>>>>>
>>>>>>>>>> It is a tautology:
>>>>>>>>>>
>>>>>>>>>> void Infinite_Loop()
>>>>>>>>>> {
>>>>>>>>>>     HERE: goto HERE;
>>>>>>>>>> }
>>>>>>>>>>
>>>>>>>>>> int main()
>>>>>>>>>> {
>>>>>>>>>>     Output("Input_Halts = ", H0((u32)Infinite_Loop));
>>>>>>>>>> }
>>>>>>>>>>
>>>>>>>>>> _Infinite_Loop()
>>>>>>>>>> [00001102](01)  55         push ebp
>>>>>>>>>> [00001103](02)  8bec       mov ebp,esp
>>>>>>>>>> [00001105](02)  ebfe       jmp 00001105
>>>>>>>>>> [00001107](01)  5d         pop ebp
>>>>>>>>>> [00001108](01)  c3         ret
>>>>>>>>>> Size in bytes:(0007) [00001108]
>>>>>>>>>>
>>>>>>>>>> The fact that the simulation by H0(Infinite_Loop) of its input
>>>>>>>>>> would never stop running unless H0 aborts the simulation of this
>>>>>>>>>> input is merely another way of saying that the correct and
>>>>>>>>>> complete simulation of this input by H0 never halts.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Strawman.
>>>>>>>>>
>>>>>>>>> Fallacy of Proof by Example.
>>>>>>>>
>>>>>>>> It is only a fallacy by proof of example when a single example is
>>>>>>>> not representative of the entire class of all such examples.
>>>>>>>
>>>>>>> Which it isn't, so FAIL.
>>>>>>>
>>>>>>>>
>>>>>>>> You are confusing existential quantification with universal
>>>>>>>> quantification.
>>>>>>>
>>>>>>> No you are, since you are claiming the rule works for ALL inputs (or
>>>>>>> at least P, which isn't the input you showed it worked for).
>>>>>>>
>>>>>>>>
>>>>>>>> ∀ simulating halt decider H
>>>>>>>> ∀ machine description P
>>>>>>>
>>>>>>> Right. ∀, so UNIVERSAL, not EXISTENTIAL, so FAIL.
>>>>>>>
>>>>>>> You don't seem to know what those words mean.
>>>>>>>
>>>>>>>>
>>>>>>>> The fact that the correct simulation of any input P to any
>>>>>>>> simulating halt decider (SHD) H would never stop running unless
>>>>>>>> this simulation was aborted by this SHD is merely another way of
>>>>>>>> saying that the correct and complete simulation of this input by
>>>>>>>> this SHD would never stop running.
>>>>>>>>
>>>>>>>
>>>>>>> Nope, not when the input depends on the behavior of the decider in
>>>>>>> the way you are doing it.
>>>>>>
>>>>>> P was intentionally defined to depend on the behavior of its decider
>>>>>> *NITWIT*
>>>>>
>>>>> Right, but your "example" didn't, and this dependency is the case that
>>>>> breaks your "proof" and definitions.
>>>> The correct simulation of the input by H(P,P) is the simulation of this
>>>> input at the same point in the execution trace where H is invoked.
>>>>
>>>> (2 + 3) * 5   !=   2 + (3 * 5) // order matters.
>>>>
>>>>
>>> Red Herring. Simulate(P,P) is ALWAYS the same (for a given P)
>>> irrespective of "where" that simulation is done, at least if H is a
>>> computation, and if it isn't it can't be a decider.
>>>
>>> The CORRECT equivalent to the above would be:
>>>
>>> Add(2, 3) * 5 == 5 * Add(2, 3)
>> You shit for brains fucking idiot! That is not true.
>>
>> https://en.wikipedia.org/wiki/Noncommutative_ring
>>
>
> Except both are 5 * 5 so the commutative property isn't needed.
>
> You are just being a DICK.

(2 + 3) * 5 == 25 // 2+3== 5 then 5* 5==25
2 + (3 * 5) == 17 // 3*5==15 then 2+15==17

If you can't even correctly compute 8th grade arithmetic how can you
possibly have sufficient knowledge of more complex computations?


--
Copyright 2022 Pete Olcott "Talent hits a target no one else can hit;
Genius hits a target no one else can see." Arthur Schopenhauer


Mr Flibble

unread,
Sep 4, 2022, 11:08:10 AM9/4/22
to
Except Richard didn't imply that, Richard implied that multiplication
is commutative for real numbers, which it is, so this is just an
obvious strawman you lying cunt.

/Flibble


0 new messages