Warning: mysqli::__construct(): (HY000/1203): User howardkn already has more than 'max_user_connections' active connections in D:\Inetpub\vhosts\howardknight.net\al.howardknight.net\includes\artfuncs.php on line 21
Failed to connect to MySQL: (1203) User howardkn already has more than 'max_user_connections' active connections
Warning: mysqli::query(): Couldn't fetch mysqli in D:\Inetpub\vhosts\howardknight.net\al.howardknight.net\index.php on line 66
Article <v39fll$2grvb$2@i2pn2.org>
Deutsch   English   Français   Italiano  
<v39fll$2grvb$2@i2pn2.org>

View for Bookmarking (what is this?)
Look up another Usenet article

Path: ...!weretis.net!feeder9.news.weretis.net!i2pn.org!i2pn2.org!.POSTED!not-for-mail
From: joes <noreply@example.com>
Newsgroups: comp.theory
Subject: Re: Two dozen people were simply wrong --- Try to prove otherwise
Date: Thu, 30 May 2024 09:08:05 -0000 (UTC)
Organization: i2pn2 (i2pn.org)
Message-ID: <v39fll$2grvb$2@i2pn2.org>
References: <v3501h$lpnh$1@dont-email.me> <v362eu$2d367$3@i2pn2.org>
	<v363js$vg63$2@dont-email.me> <v36803$2d368$3@i2pn2.org>
	<v368je$100kd$3@dont-email.me> <v373mr$2d367$5@i2pn2.org>
	<v37bpa$15n0b$1@dont-email.me> <v37i9p$lls$1@news.muc.de>
	<87y17smqnq.fsf@bsb.me.uk> <v37sap$18mfo$1@dont-email.me>
	<v38eq4$2foi0$1@i2pn2.org> <v38fe0$1bndb$1@dont-email.me>
	<v38g31$2foi0$11@i2pn2.org> <v38gi5$1bndb$3@dont-email.me>
	<v38ici$2fohv$2@i2pn2.org> <v38j17$1c8ir$2@dont-email.me>
	<v38jgo$2foi0$14@i2pn2.org> <v38jv9$1c8ir$4@dont-email.me>
	<v38kh7$2foi0$15@i2pn2.org> <v38lsl$1ggjs$1@dont-email.me>
	<v38o71$2foi0$17@i2pn2.org> <v38ogh$1grj4$1@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Injection-Date: Thu, 30 May 2024 09:08:05 -0000 (UTC)
Injection-Info: i2pn2.org;
	logging-data="2650091"; mail-complaints-to="usenet@i2pn2.org";
	posting-account="nS1KMHaUuWOnF/ukOJzx6Ssd8y16q9UPs1GZ+I3D0CM";
User-Agent: Pan/0.145 (Duplicitous mercenary valetism; d7e168a
 git.gnome.org/pan2)
X-Spam-Checker-Version: SpamAssassin 4.0.0
Bytes: 5007
Lines: 79

Am Wed, 29 May 2024 21:32:49 -0500 schrieb olcott:
> On 5/29/2024 9:27 PM, Richard Damon wrote:
>> On 5/29/24 9:48 PM, olcott wrote:
>>> On 5/29/2024 8:24 PM, Richard Damon wrote:
>>>> On 5/29/24 9:15 PM, olcott wrote:
>>>>> On 5/29/2024 8:07 PM, Richard Damon wrote:
>>>>>> On 5/29/24 8:59 PM, olcott wrote:
>>>>>>> On 5/29/2024 7:48 PM, Richard Damon wrote:
>>>>>>>> On 5/29/24 8:17 PM, olcott wrote:
>>>>>>>>> On 5/29/2024 7:09 PM, Richard Damon wrote:
>>>>>>>>>> On 5/29/24 7:57 PM, olcott wrote:
>>>>>>>>>>> On 5/29/2024 6:47 PM, Richard Damon wrote:
>>>>>>>>>>>> On 5/29/24 2:31 PM, olcott wrote:
>>>>>>>>>>>>> On 5/29/2024 1:14 PM, Ben Bacarisse wrote:
>>>>>>>>>>>>>> Alan Mackenzie <acm@muc.de> writes:
>>>>>>>>>>>>>>> How about a bit of respect?  Mike specifically asked you 
>>>>>>>>>>>>>>> not to cite his
>>>>>>>>>>>>>>> name as a back up for your points.  Why do you keep doing it?

>>>>>>>>>>>>> It turns out that two dozen people are easily proven wrong when
>>>>>>>>>>>>> they claimed that the correct simulation of the input to H(D,D)
>>>>>>>>>>>>> is the behavior of int main() { D(D); }
>>>>>>>>>>>>>
>>>>>>>>>>>> How is that?

>>>>>>>>>>>> Which isn't a "Correct Simulation" by the definition that 
>>>>>>>>>>>> allow the relating of a "Simulation" to the behavior of an 
>>>>>>>>>>>> input.
>>>>>>>>>>> Right the execution trace of D simulated by pure function H using
>>>>>>>>>>> an x86 emulator must show that D cannot possibly reach its own
>>>>>>>>>>> simulated final state and halt or the simulation of the machine
>>>>>>>>>>> language of D is incorrect or in the wrong order.
Or aborts prematurely.

>>>>>>>>>> So, you aren't going to resolve the question but just keep up 
>>>>>>>>>> with your contradiction that H is simulating a template (that 
>>>>>>>>>> doesn't HAVE any instrucitons of H in it) but also DOES 
>>>>>>>>>> simulate those non-existance instructions by LYING about what 
>>>>>>>>>> it does and simulating a SPECIFIC instance that it LIES behaves 
>>>>>>>>>> just like DIFFERENT specific instatces.

>>>>>>>> But the question ISN'T about the SIMULATED D, but about the 
>>>>>>>> behavior of the actual PROGRAM/MACHINE D
Which should be the same.
>>>>>>>> This seems to be your blind spot.

>>>>>>> ∃H  ∈ Turing_Machines
>>>>>>> ∀x  ∈ Turing_Machines_Descriptions
>>>>>>> ∀y  ∈ Finite_Strings
>>>>>>> such that H(x,y) = Halts(x,y)
>>>>>>>
>>>>>>> Not really the above formalization does not can cannot
>>>>>>> specify Turing Machines as the input to any decider H.
>>>>>>>
>>>>>> Then what is x representing?
>>>>>
>>>>> x <is> a finite string Turing machine description that SPECIFIES 
>>>>> behavior. The term: "representing" is inaccurate.
What’s the difference? 

>>>> No, it specifies the machine, and thus, though that, the behavior.
>>>>
>>> If we assume that a decider takes an actual Turing machine as its
>>> input that is correct otherwise that is one level of indirection
>>> away from what we are really looking at.
>>>
>>> The people have perpetuated this mistake for many decades never
>>> actually made it not a mistake.


>> You need to define what you mean by "Indirection", because you aren't 
>> using it in the normal manner.
> 
> I have conclusively proven that the behavior of the correct
> simulation of the x86 code of D by pure function H has
> different behavior than the direct execution of D(D).
Then H is not a correct simulator.

-- 
joes