On 2025-10-19 09:31:21 +0000, Richard Heathfield said:
On 19/10/2025 09:39, Mikko wrote:
On 2025-10-18 10:47:45 +0000, olcott said:
<snip>
This is the complete proof to anyone that knows the
semantics of C. Every LLM of five has validated it.
By the C rules there is no valid way to use programs as data.
Not quite true. They are stored on the filesystem as data, and
therefore can be read as data.
You are right, the C rules don't prohibit that (though don't require
that it be possible, either).
What he /can't/ by C rules is to claim that a function pointer
is a program. It isn't. It's a pointer.
And the C rules don't allow reading the pointed machine code
as data.
On 2025-10-18 10:47:45 +0000, olcott said:
On 10/18/2025 3:58 AM, Mikko wrote:
Of course you can. There are Turing machine simulators on the Web,
and making one's own is much easier than making one's own C
simulator.
Likewise you can mow your one acre lawn with
fingernail clippers.
Doable but even less sensible than writing programs in C.
You have not succeeded in presenting your point fully with C.
Instead you have needed to discuss the machine code translation
of the C code. The core part of your "deciders" is a machine
code simulator, not a C code simulator. When you have presented
execution traces you have presented machine code traces, not
C traces.
This is the complete proof to anyone that knows the
semantics of C. Every LLM of five has validated it.
By the C rules there is no valid way to use programs as data.
Therefore
there is no way to implement your HHH without implementation defined extensions that are not known to everyone that knows the semantics of
C. Also note that those who use C often don't know all about the
semantics specified by the latest C standard as the semantics of the implementation they use is more important to them, and even that only
to the extent it is relevant to the purpose of whatever they are doing.
On 2025-10-18 10:47:45 +0000, olcott said:
On 10/18/2025 3:58 AM, Mikko wrote:
On 2025-10-17 15:11:57 +0000, olcott said:
On 10/17/2025 3:45 AM, Mikko wrote:
On 2025-10-16 13:27:15 +0000, olcott said:
On 10/16/2025 4:34 AM, Mikko wrote:
On 2025-10-15 15:10:47 +0000, olcott said:
On 10/15/2025 9:50 AM, tTh wrote:
On 10/15/25 14:32, olcott wrote:
Here is that full proof.Can you take those insanity out of comp.lang.c ?
https://chatgpt.com/share/68eef2df-0f10-8011-8e92-264651cc518c >>>>>>>>>
I just needed a couple of cross posts.
Here is how it is related to C/C++
There is nothing language specific in the halting problem.
It is just that the C code is fully operational
code that succinctly makes my point.
That it is C code is not essential to its purpose.
It makes my point 100% concrete and empirically testable.
One can't do that with the abstractions of Turing machines.
Of course you can. There are Turing machine simulators on the Web,
and making one's own is much easier than making one's own C
simulator.
Likewise you can mow your one acre lawn with
fingernail clippers.
Doable but even less sensible than writing programs in C.
You have not succeeded in presenting your point fully with C.
Instead you have needed to discuss the machine code translation
of the C code. The core part of your "deciders" is a machine
code simulator, not a C code simulator. When you have presented
execution traces you have presented machine code traces, not
C traces.
This is the complete proof to anyone that knows the
semantics of C. Every LLM of five has validated it.
By the C rules there is no valid way to use programs as data.
Therefore
there is no way to implement your HHH without implementation defined extensions that are not known to everyone that knows the semantics of
C. Also note that those who use C often don't know all about the
semantics specified by the latest C standard as the semantics of the implementation they use is more important to them, and even that only
to the extent it is relevant to the purpose of whatever they are doing.
On 10/19/2025 1:39 AM, Mikko wrote:
On 2025-10-18 10:47:45 +0000, olcott said:
On 10/18/2025 3:58 AM, Mikko wrote:
On 2025-10-17 15:11:57 +0000, olcott said:
On 10/17/2025 3:45 AM, Mikko wrote:
On 2025-10-16 13:27:15 +0000, olcott said:
On 10/16/2025 4:34 AM, Mikko wrote:
On 2025-10-15 15:10:47 +0000, olcott said:
On 10/15/2025 9:50 AM, tTh wrote:
On 10/15/25 14:32, olcott wrote:
Here is that full proof.Can you take those insanity out of comp.lang.c ?
https://chatgpt.com/share/68eef2df-0f10-8011-8e92-264651cc518c >>>>>>>>>>
I just needed a couple of cross posts.
Here is how it is related to C/C++
There is nothing language specific in the halting problem.
It is just that the C code is fully operational
code that succinctly makes my point.
That it is C code is not essential to its purpose.
It makes my point 100% concrete and empirically testable.
One can't do that with the abstractions of Turing machines.
Of course you can. There are Turing machine simulators on the Web,
and making one's own is much easier than making one's own C
simulator.
Likewise you can mow your one acre lawn with
fingernail clippers.
Doable but even less sensible than writing programs in C.
You have not succeeded in presenting your point fully with C.
Instead you have needed to discuss the machine code translation
of the C code. The core part of your "deciders" is a machine
code simulator, not a C code simulator. When you have presented
execution traces you have presented machine code traces, not
C traces.
This is the complete proof to anyone that knows the
semantics of C. Every LLM of five has validated it.
By the C rules there is no valid way to use programs as data.
A c function that takes in source code? Has a built in compiler to start
the process of compiling it?
Therefore
there is no way to implement your HHH without implementation defined
extensions that are not known to everyone that knows the semantics of
C. Also note that those who use C often don't know all about the
semantics specified by the latest C standard as the semantics of the
implementation they use is more important to them, and even that only
to the extent it is relevant to the purpose of whatever they are doing.
| Sysop: | DaiTengu |
|---|---|
| Location: | Appleton, WI |
| Users: | 1,073 |
| Nodes: | 10 (0 / 10) |
| Uptime: | 215:50:47 |
| Calls: | 13,783 |
| Calls today: | 1 |
| Files: | 186,987 |
| D/L today: |
341 files (78,091K bytes) |
| Messages: | 2,434,663 |