Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Fri, Jan 25, 2008 at 05:02:09PM -0800, Darren New wrote:

"The Traveling Salesman Problem: A problem that has been vexing computer 
scientists for many decades, but which traveling salesmen solve on a daily 
basis."


Human traveling salesmen are good at coming up with a solution that's
reasonably good.  TSP itself involves finding the best problem.  Optimizers
try to also do the reasonably good aspect, which seems to be a lot harder
to implement than the intuitive approach we make as humans.

Dave

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

David Brown wrote:

Without it, I've not proven what I intended to prove.  Otherwise, we don't
know if there isn't some yet-undiscovered algorithm out there that solves
the problem more efficiently.  If we can prove that all algorithms must use
either O(2^2^n) space or time, the conclusion follows.


Yes, but that's a really hard problem, which is the point I'm making. :-)

Actually, proving this would probably lead to all kinds of good stuff 
coming out of it for compiler theory and such.



"The Traveling Salesman Problem: A problem that has been vexing computer 
scientists for many decades, but which traveling salesmen solve on a 
daily basis."



--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[SJS]

begin  quoting David Brown as of Fri, Jan 25, 2008 at 11:09:05AM -0800:
> On Fri, Jan 25, 2008 at 10:43:59AM -0800, SJS wrote:
> 
> >I think you're going to have a problem because you don't constrain what
> >"severely bounded" means. We're no longer talking turing machines; if we
> >bound our machine to run only primitive recursive programs, we're can
> >make confident judgements about the machine halting in fixed time.
> 
> Ok.  Let's make our only bound memory.  To simplify, let's have the program
> memory be separate, but finite.  The PC needs to be represented in our
> memory though (so basically the memory includes CPU registers).
> 
> There are no constraints on what the architecture is.

Better. :)

> The goal isn't to show it is possible to make a computer simple enough to
> solve halting, but to show that limiting memory still is impractical to
> solve halting.

Yes. I'm just trying to clarify the scope a bit.

> >>  1.  I define impractical as where the solution to any non-trivial input
> >>  size requires more storage than the size of the universe or more time
> >>  than the duration of the known universe.
> >
> >Duration is meaningless without a performance limit. More bits than the
> >current estimated size of the universe is a better yardstick, but even
> >then you're making assumptions as to how the target scheme will run.
> 
> Let me clarify.  Assume a clock tick of the planck time (~ 10e-44 s).
> Assuming 13.8e9 years for the age, there are about 4e61 possible clock
> ticks in the universe.  It is not "practical" to build a computer whose
> clock can run faster than the planck time.  I don't care if your computer
> is slower, since it only makes things worse.

We're limiting ourselves to von neumann architecture for our machines, I
presume?  (That seems like a fair assumption, given that we're talking
about practical computability theory.)

> >>  2.  Given a processing system with 'n' bits of memory.  Each processing
> >>  step modifies zero or more of these bits.  Halting is defined as a
> >>  processing step that modifies zero bits.  The processing system has no
> >>  other state.
> >
> >Don't we also need to define input?  I would think that our bounded machine
> >could be described with a DFA operating on a symbol set of size 1 and
> >an unbounded input string (the clock). We need to keep the input separate
> >so that our program is general, yes?
> 
> Any input is part of memory, that's the point of bounding it.

Okay, no I/O with the universe, all data is present at the start of
the program.

Or tested program/machine is bounded; is our testing program/machine?

> >>  3.  Because there are a finite number of states, it is sufficient to
> >>  detect a repeated state to detect a non-halt.  I know of two ways of
> >>  detecting loops (but I don't know how to prove there aren't more):
> >>
> >>a.  Track all visited states.  There are 2^(2^n) possible states.  
> >>This requires impractical storage for all non-trivial n.
> >
> >There may be clever ways around that. 
> 
> Not if you don't know anything about what possible states and transitions
> there are.

Just because I'm not clever enough to devise a way around it doesn't
mean that it can't be done.

> >>b.  Run two identical systems from the same starting system, but run
> >>one two steps for each step of the first.  If they ever reach the same
> >>state, there is a cycle, and the system will not halt.  This requires
> >>2*2^n storage.  But, the longest possible cycle would go through every
> >>state, requiring 2^(2^n) steps.  This requires impractical time.
> >
> >Take the trivial machine that halts.  The trailing machine will have the
> >same bit-pattern as the leading machine, and flag a cycle, despite no
> >state changing ever.
> 
> The proof is not that it is possible for you to construct a machine that
> can detect some halts, but that it is impossible to construct a machine
> that detects all halts.

And there's the degenerate case for approach (b). It needs to be
reformulated, or discounted entirely, no proof by induction needed as
the counter-example is by construction. :)

> >> It also necessary to
> >>prove that the repeat detection of 3 is necessary instead of merely
> >>sufficient.
> >
> >I can see why that would help, but why would that be true?
> 
> Without it, I've not proven what I intended to prove.  Otherwise, we don't
> know if there isn't some yet-undiscovered algorithm out there that solves
> the problem more efficiently.  If we can prove that all algorithms must use
> either O(2^2^n) space or time, the conclusion follows.

Yes, but "it has to be true for my argument to work" doesn't give me a
warm fuzzy about the assertion being true. 

Plus, doesn't sufficient trump necessary?

-- 
We will eventually test against the computer of Satan
In theory, of course, as the real one is out baitin'.
Stewart Stremler

-- 
KPLUG-LPSG@kernel-panic.

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Fri, Jan 25, 2008 at 10:43:59AM -0800, SJS wrote:


I think you're going to have a problem because you don't constrain what
"severely bounded" means. We're no longer talking turing machines; if we
bound our machine to run only primitive recursive programs, we're can
make confident judgements about the machine halting in fixed time.


Ok.  Let's make our only bound memory.  To simplify, let's have the program
memory be separate, but finite.  The PC needs to be represented in our
memory though (so basically the memory includes CPU registers).

There are no constraints on what the architecture is.

The goal isn't to show it is possible to make a computer simple enough to
solve halting, but to show that limiting memory still is impractical to
solve halting.


  1.  I define impractical as where the solution to any non-trivial input
  size requires more storage than the size of the universe or more time
  than the duration of the known universe.


Duration is meaningless without a performance limit. More bits than the
current estimated size of the universe is a better yardstick, but even
then you're making assumptions as to how the target scheme will run.


Let me clarify.  Assume a clock tick of the planck time (~ 10e-44 s).
Assuming 13.8e9 years for the age, there are about 4e61 possible clock
ticks in the universe.  It is not "practical" to build a computer whose
clock can run faster than the planck time.  I don't care if your computer
is slower, since it only makes things worse.


  2.  Given a processing system with 'n' bits of memory.  Each processing
  step modifies zero or more of these bits.  Halting is defined as a
  processing step that modifies zero bits.  The processing system has no
  other state.


Don't we also need to define input?  I would think that our bounded machine
could be described with a DFA operating on a symbol set of size 1 and
an unbounded input string (the clock). We need to keep the input separate
so that our program is general, yes?


Any input is part of memory, that's the point of bounding it.


  3.  Because there are a finite number of states, it is sufficient to
  detect a repeated state to detect a non-halt.  I know of two ways of
  detecting loops (but I don't know how to prove there aren't more):

a.  Track all visited states.  There are 2^(2^n) possible states.  This
requires impractical storage for all non-trivial n.


There may be clever ways around that. 


Not if you don't know anything about what possible states and transitions
there are.


b.  Run two identical systems from the same starting system, but run
one two steps for each step of the first.  If they ever reach the same
state, there is a cycle, and the system will not halt.  This requires
2*2^n storage.  But, the longest possible cycle would go through every
state, requiring 2^(2^n) steps.  This requires impractical time.


Take the trivial machine that halts.  The trailing machine will have the
same bit-pattern as the leading machine, and flag a cycle, despite no
state changing ever.


The proof is not that it is possible for you to construct a machine that
can detect some halts, but that it is impossible to construct a machine
that detects all halts.


 It also necessary to
prove that the repeat detection of 3 is necessary instead of merely
sufficient.


I can see why that would help, but why would that be true?


Without it, I've not proven what I intended to prove.  Otherwise, we don't
know if there isn't some yet-undiscovered algorithm out there that solves
the problem more efficiently.  If we can prove that all algorithms must use
either O(2^2^n) space or time, the conclusion follows.

Dave

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[SJS]

begin  quoting David Brown as of Fri, Jan 25, 2008 at 10:03:54AM -0800:
> On Fri, Jan 25, 2008 at 09:09:43AM -0800, Darren New wrote:
> >David Brown wrote:
> >>but even on a severely bounded machine,
> >>there is no practical solution.
> >
> >Prove it.
> 
> Well, here's part of a proof.

I think you're going to have a problem because you don't constrain what
"severely bounded" means. We're no longer talking turing machines; if we
bound our machine to run only primitive recursive programs, we're can
make confident judgements about the machine halting in fixed time.

Other ways of "severely bounding" the machine may also offer "practical"
solutions.

>   1.  I define impractical as where the solution to any non-trivial input
>   size requires more storage than the size of the universe or more time
>   than the duration of the known universe.

Duration is meaningless without a performance limit. More bits than the
current estimated size of the universe is a better yardstick, but even
then you're making assumptions as to how the target scheme will run.

>   2.  Given a processing system with 'n' bits of memory.  Each processing
>   step modifies zero or more of these bits.  Halting is defined as a
>   processing step that modifies zero bits.  The processing system has no
>   other state.

Don't we also need to define input?  I would think that our bounded machine
could be described with a DFA operating on a symbol set of size 1 and
an unbounded input string (the clock). We need to keep the input separate
so that our program is general, yes?

>   3.  Because there are a finite number of states, it is sufficient to
>   detect a repeated state to detect a non-halt.  I know of two ways of
>   detecting loops (but I don't know how to prove there aren't more):
> 
> a.  Track all visited states.  There are 2^(2^n) possible states.  This
> requires impractical storage for all non-trivial n.

There may be clever ways around that. 

> b.  Run two identical systems from the same starting system, but run
> one two steps for each step of the first.  If they ever reach the same
> state, there is a cycle, and the system will not halt.  This requires
> 2*2^n storage.  But, the longest possible cycle would go through every
> state, requiring 2^(2^n) steps.  This requires impractical time.

Take the trivial machine that halts.  The trailing machine will have the
same bit-pattern as the leading machine, and flag a cycle, despite no
state changing ever.

Special consideration needs to be taken for the initialization.

> I suspect 3a and 3b can be proven by contradiction.

"Assume that with storage of fixed size Q, we can detect a duplicate
state in 2^2^n states."?

Hm.  Perhaps.

>  It also necessary to
> prove that the repeat detection of 3 is necessary instead of merely
> sufficient.

I can see why that would help, but why would that be true?

-- 
No offense -- it's not that I disbelieve you,
But your argument has some convincing to do.
Stewart Stremler

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Tracy R Reed]

David Brown wrote:

  1.  I define impractical as where the solution to any non-trivial input
  size requires more storage than the size of the universe or more time
  than the duration of the known universe.


You are SO unreasonable.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Fri, Jan 25, 2008 at 09:09:43AM -0800, Darren New wrote:

David Brown wrote:

but even on a severely bounded machine,
there is no practical solution.


Prove it.


Well, here's part of a proof.

  1.  I define impractical as where the solution to any non-trivial input
  size requires more storage than the size of the universe or more time
  than the duration of the known universe.

  2.  Given a processing system with 'n' bits of memory.  Each processing
  step modifies zero or more of these bits.  Halting is defined as a
  processing step that modifies zero bits.  The processing system has no
  other state.

  3.  Because there are a finite number of states, it is sufficient to
  detect a repeated state to detect a non-halt.  I know of two ways of
  detecting loops (but I don't know how to prove there aren't more):

a.  Track all visited states.  There are 2^(2^n) possible states.  This
requires impractical storage for all non-trivial n.

b.  Run two identical systems from the same starting system, but run
one two steps for each step of the first.  If they ever reach the same
state, there is a cycle, and the system will not halt.  This requires
2*2^n storage.  But, the longest possible cycle would go through every
state, requiring 2^(2^n) steps.  This requires impractical time.

I suspect 3a and 3b can be proven by contradiction.  It also necessary to
prove that the repeat detection of 3 is necessary instead of merely
sufficient.

David

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

David Brown wrote:

but even on a severely bounded machine,
there is no practical solution.


Prove it.

--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Andrew Lentvorski]

[EMAIL PROTECTED] wrote:

Also, I don't know if this was mentioned yet, but the whole Star Trek
Transporter thing has a big hurdle...quantum mechanics.  The Heisenberg
Uncertainty Principle says you can't simultaneously know the position and
velocity of all the atoms in your body precisely.


That's not intrinsically a barrier.

As long as we put all of the atoms back in their spot with a "close 
enough" position and with a "close enough" velocity, the natural 
relaxation processes of chemistry and diffusion will put the system back 
together just fine.  The Heisenberg Uncertainty Principle holds for 
*all* observers, including the other subatomic particles in the local 
environment.


However, I'm with Leonard McCoy on this one.  I'll take the 
shuttlecraft, thanks.


-a

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[chris]

On Thu, Jan 24, 2008 at 07:40:43PM -0800, David Brown wrote:
> All I'm saying is that the halting problem states that problems on an
> unbounded machine isn't solvable, but even on a severely bounded machine,
> there is no practical solution.

Yes.  If I may throw in my 2 cents

Even apparently simple systems can exhibit enormously complex behvior.  Think
of fractals and the entire field of nonlinear dynamics.  Butterfly effect n'
all that.

Also, I don't know if this was mentioned yet, but the whole Star Trek
Transporter thing has a big hurdle...quantum mechanics.  The Heisenberg
Uncertainty Principle says you can't simultaneously know the position and
velocity of all the atoms in your body precisely.

Also, this whole philosophical discussion of computation theory applied to
the human brain has a nice pop science book about it called "The Emperor's New
Mind" by Penrose.

cs

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Thu, Jan 24, 2008 at 04:48:28PM -0800, Darren New wrote:

David Brown wrote:

I can definitively declare that this is unsolvable, just because of
magnitude.  In fact solving the halting problem this way only works for
trivially small computer system (a handfull of bits).


Not at all. Solve it on a Turing machine. That's kind of the point I'm 
making. Before you say "the brain can (or cannot) solve the halting 
problem," you have to understand what the halting problem is and what makes 
it unsolvable.


All I'm saying is that the halting problem states that problems on an
unbounded machine isn't solvable, but even on a severely bounded machine,
there is no practical solution.

David

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

Bob La Quey wrote:

So I would say that all programs running on real machines
will halt :)


Sure. But you stopped at the summary, rather than the "Formal Statement":

"""
The halting problem is a decision problem about properties of computer 
programs on a fixed Turing-complete model of computation.

"""

No real computer is Turing-complete, and Turing-complete models of 
computation are uninteresting as computers in the real world (having no 
I/O, for one).


Plus, there's no formal reason to believe biological systems are 
necessarily limited in time or space.  Finite, sure, but not necessarily 
bounded. Many fish, for example, will live and grow until you kill them.


This stuff is tricky. You really need to know what you're doing to make 
use of it. :-)


--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Bob La Quey]

On Jan 24, 2008 4:48 PM, Darren New <[EMAIL PROTECTED]> wrote:
> David Brown wrote:
> > I can definitively declare that this is unsolvable, just because of
> > magnitude.  In fact solving the halting problem this way only works for
> > trivially small computer system (a handfull of bits).
>
> Not at all. Solve it on a Turing machine. That's kind of the point I'm
> making. Before you say "the brain can (or cannot) solve the halting
> problem," you have to understand what the halting problem is and what
> makes it unsolvable.
>

Given a description of a program and a finite input, decide whether
the program finishes running or will run forever, given that input.

To the best of my knowledge no program will run forever on
a biological machine. Mostly as far as I know people consider
machines that are bounded in space, i.e. memory but many, indeed
most, real machines are bounded in time. Real machines cease to
exist after a while.

So I would say that all programs running on real machines
will halt :)

BobLQ

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

Darren New wrote:

The halting problem only is true for unbounded computers.


Or, to put it another way, the *proof* that the halting problem is 
uncomputable relies on the program running on an unbounded-storage 
machine. If you're going to argue that the halting problem is unsolvable 
for a Z80-based program, you can't argue that the program that solves 
the Z80-halting-problem on a TM wouldn't run on a real computer and 
therefore the Z80-halting problem is unsolvable. There may very well be 
a different, clever mechanism to solve the halting problem for all Z80 
computer programs, since there's already a non-clever mechanism for 
solving it.


I.e., it's entirely possible the human brain could solve the halting 
problem for all finite-storage programs. I haven't seen any proof that 
(for example) you *need* more storage to solve the 
finite-halting-problem than the finite-halting-program has available.


--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

SJS wrote:

Yah, but they violate that constraint,


Well, yeah. They're not even consistent about activating the 
communicators, which is technology we already have. :-)


I liked Larry Niven's approach 


I was thinking more John Scalzi's stuff, like 
http://books.google.com/books?id=_lO9ZroeXmgC


I won't say more, for fear of spoilers, but just from the title you can 
probably guess it would ruin the story to let soldiers be replicated. :-)


--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[SJS]

begin  quoting Darren New as of Thu, Jan 24, 2008 at 04:41:28PM -0800:
> David Brown wrote:
> >It's done occasionally, one I can specifically think of is with Riker's
> >transporter accident that created an identical copy.
> 
> They had Kirk be an evil/good Kirk once too, as well as using it to 
> (IIRC) replace someone infected with a disease with the body that was 
> sent to the diseased world instead of the body that came back. Or something.
> 
> The usual excuse is that there's too much data to actually store a copy 
> - you have to send it and reconstruct it as you create the data.

Yah, but they violate that constraint, so you can't even retcon a decent
explanation. That's one of the reasons I went from being an avid fan to
actively avoiding the show. 

(I always figured Dr. McCoy had a better grasp of the actual risk...)

I liked Larry Niven's approach -- you turn the thing you're sending
into an FTL particle with a rest-mass equivalent to that of which you're
sending (space-ship + people).  Analog, no copies, no messy digitization.

-- 
Everyone can have a personal force-shield suit that lets them walk in space
But next week everyone is freezing to death as they crash into a star's face
Stewart Stremler

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

David Brown wrote:

I can definitively declare that this is unsolvable, just because of
magnitude.  In fact solving the halting problem this way only works for
trivially small computer system (a handfull of bits).


Not at all. Solve it on a Turing machine. That's kind of the point I'm 
making. Before you say "the brain can (or cannot) solve the halting 
problem," you have to understand what the halting problem is and what 
makes it unsolvable.



First, as others have mentioned, it's 2^(8*2^16) different states.  8*2^16
is the number of bits of memory, each of which can be in a different state.


OK. No biggie. You're either unbounded, or you're not.


This is approximately  2.6 x 10^157826  different states that you would
have to store.  The magnitude of this number is just numbing.


That doesn't make it any less mathematically intractable. People work 
with the set of real numbers all the time. Between any two real numbers 
are more real numbers than all the integers put together. Let alone a 
paultry 10^157826.


The halting problem only is true for unbounded computers. Otherwise, I 
can write a Turing machine program for any program on a finite size 
computer that will check in finite time whether that program halts or 
not. That's the definition of the halting problem.


That you can't actually *run* the Turing machine program I provide on a 
machine you can actually build doesn't really say much about whether the 
brain can solve the halting problem. Any more than if I give you a 
physical machine which always breaks down after 5000 instructions, you 
can claim you can solve the halting problem for that machine.


--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

David Brown wrote:

It's done occasionally, one I can specifically think of is with Riker's
transporter accident that created an identical copy.


They had Kirk be an evil/good Kirk once too, as well as using it to 
(IIRC) replace someone infected with a disease with the body that was 
sent to the diseased world instead of the body that came back. Or something.


The usual excuse is that there's too much data to actually store a copy 
- you have to send it and reconstruct it as you create the data.


--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Thu, Jan 24, 2008 at 12:52:26PM -0800, Ralph Shumaker wrote:

The biggest problem I had with the various series was that they somehow 
only were able to do mv and not cp.  Do they expect us to believe that we 
lose the computational ability to do cp in the next 400 years?  Apparently, 
they are completely lost on the concept of working with a backup, except 
when it's necessary to pull their ass out of the fire.  Then they always 
manage to find the backups that they never seem to make otherwise.


It's done occasionally, one I can specifically think of is with Riker's
transporter accident that created an identical copy.

I suspect it isn't done much because it makes people think about the
concept of self, and how the transporter accident doesn't fit with how most
people think of their own identity.

Dave

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

RE: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Gabriel Sechan]

> From: [EMAIL PROTECTED]
> James G. Sack (jim) wrote:
> > Surely you are aware of mind melds?
> >
> >   :-)
> >
> > Regards, LL&P,
> > ..jim (IIRC, there were offload/upload/reboot episodes, too)
> >   
> 
> The biggest problem I had with the various series was that they somehow 
> only were able to do mv and not cp.  Do they expect us to believe that 
> we lose the computational ability to do cp in the next 400 years?  
> Apparently, they are completely lost on the concept of working with a 
> backup, except when it's necessary to pull their ass out of the fire.  
> Then they always manage to find the backups that they never seem to make 
> otherwise.
> 
cp was found to be an illegal device allowing one to circumvent a content 
protection mechanism, thus violating the DMCA-  RIAA v Common Sense, 2010.


Gabe
_
Helping your favorite cause is as easy as instant messaging. You IM, we give.
http://im.live.com/Messenger/IM/Home/?source=text_hotmail_join--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Ralph Shumaker]

James G. Sack (jim) wrote:

Ralph Shumaker wrote:
  

Andrew Lentvorski wrote:


David Brown wrote:
  

Computers we make have to be laboriously programmed.


Um, excuse me?  Literacy and education don't come for free, you know.
  

That, and the programming of the silicon is generally done once, on one
machine, and then just loaded by others like it.  I don't know of anyone
who has been able to download any of his schooling into the analog
computer of anyone else.  Compile time applies to each person's analog
gray matter computer.  No RPMs there.




Surely you are aware of mind melds?

  :-)

Regards, LL&P,
..jim (IIRC, there were offload/upload/reboot episodes, too)
  


The biggest problem I had with the various series was that they somehow 
only were able to do mv and not cp.  Do they expect us to believe that 
we lose the computational ability to do cp in the next 400 years?  
Apparently, they are completely lost on the concept of working with a 
backup, except when it's necessary to pull their ass out of the fire.  
Then they always manage to find the backups that they never seem to make 
otherwise.




--
Ralph


One day I stumbled across a case of Scotch.

As I recall, I stumbled several days thereafter.
--W.C. Fields

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Andrew Lentvorski]

Gus Wirth wrote:
And what would you choose to get rid of to make way for that new 
language?  Although the capacity of the human brain is large, it is not

infinite.


No, but it sure does store a lot of extra junk and forgets a lot, as well.

In *spite* of my active avoidance of TV and tabloids, I seem to have a 
remarkable amount of cells that have been bombarded with the latest 
exploits of Britney Spears, Paris Hilton, et al.


I would *happily* repurpose those brain cells for something useful like 
a new language.


-a

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[James G. Sack (jim)]

Ralph Shumaker wrote:
> Andrew Lentvorski wrote:
>> David Brown wrote:
>>> Computers we make have to be laboriously programmed.
>>
>> Um, excuse me?  Literacy and education don't come for free, you know.
> 
> That, and the programming of the silicon is generally done once, on one
> machine, and then just loaded by others like it.  I don't know of anyone
> who has been able to download any of his schooling into the analog
> computer of anyone else.  Compile time applies to each person's analog
> gray matter computer.  No RPMs there.
> 

Surely you are aware of mind melds?

  :-)

Regards, LL&P,
..jim (IIRC, there were offload/upload/reboot episodes, too)

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Tom Gal]

They are the same thing. Really what you call digital is the idealization of
analog signals by abstracting certain things away so you can use an easier
framework to solve a problem.

It's also true that in this day and age we're getting to the point where
many functions like comparators and filters that can easily be implemented
as analog circuits are limited by component accuracy to the point that it's
easier to just digitize a signal and execute the function digitally.

Basically this guy gave you a demo that in my mind meant nothing in my
opinion. You don't have to understand your engine to use 2-3 pedals, the
steering wheel, and maybe the clutch on your car right? Sometimes we still
get screwed by mechanics though :)

-Tom

On Jan 23, 2008 10:54 AM, Ralph Shumaker <[EMAIL PROTECTED]> wrote:

> Gus Wirth wrote:
> > SJS wrote:
> > [snip]
> >> Of course, you could solve mathematical equations directly with an
> >> analog
> >> computer. . . look, ma, no registers!
> >
> > Which I have done. It is particularly useful for solving things like
> > partial differential equations. The other nice thing about analog
> > computers is how easy they are to program. Just a few dials and patch
> > cords, set a few initial conditions, then let it run. Of course, it
> > helps to know advanced calculus and things like Laplace transforms. On
> > my bookshelf I have "Introduction to Analog Computation" by Joseph J.
> > Blum, Harcort, Brace & World, Inc. Maybe I should post some problems
> > and see how people would solve them in the digital realm.
> >
> > Gus
>
> Are there general purpose analog computers?  Can they run programs in
> the sense that digital computers can?
>
>  From what I understand, analog circuits are not as susceptible to EMP
> or short circuits.  I once saw a show where a guy had built two
> mechanical spider like robots, one with digital circuits, the other
> analog.  They were both crawling along when he took his soldering iron
> and ran it along the exposed circuit board of the one with digital
> circuits and it instantly stopped cold.  It was toast.  Then he did the
> same with the other one, raking the soldering iron back and forth
> repeatedly.  Aside from the sparks, it was almost like he was merely
> petting a crawling spider.  It was undaunted and kept on going, like it
> didn't even notice, or just didn't care.  But analog circuits don't
> necessarily mean that it's an analog computer, or does it?
>
> I think I may have asked this before, but it's been a while and I don't
> recall the answers.
>
>
>
> --
> Ralph
>
> 
> I hope I never meet a man so narrow minded as to spell a word in only one
> way.
> --Thomas Jefferson
>
> --
> KPLUG-LPSG@kernel-panic.org
> http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
>



-- 
-Thomas Gal
-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Thu, Jan 24, 2008 at 08:51:07AM -0800, Gus Wirth wrote:

And what would you choose to get rid of to make way for that new language? 
Although the capacity of the human brain is large, it is not infinite.


I wonder how we can measure when we are near the limit of capacity?


"Teacher, My Brain is Full" -- Gary Larson, The Far Side

David

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Wed, Jan 23, 2008 at 11:38:43PM -0800, Bob La Quey wrote:


As a consequence I have thought a lot about this
problem. Why, for instance, would evolution tolerate
such an "obviously dysfunctional" disease.




for example.  The current theory is that it is a consequence/tradeoff
of what makes creativity possible.  Creativity definitely definitively has
survival benefits.

Also look at , a term used
to describe something that evolves in correlation to something else being
selected.

David

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Thu, Jan 24, 2008 at 08:13:30AM -0800, Darren New wrote:

The Z80 has 2^16 bytes of address space (add the appropriate amount of 
space for registers). That means the processor can only be in one of 8*2^16 
different states (again, modulo register content, which is just more 
memory, and ignoring external stuff like removable media, interrupts, 
keyboards, etc).


I can definitively declare that this is unsolvable, just because of
magnitude.  In fact solving the halting problem this way only works for
trivially small computer system (a handfull of bits).

First, as others have mentioned, it's 2^(8*2^16) different states.  8*2^16
is the number of bits of memory, each of which can be in a different state.

This is approximately  2.6 x 10^157826  different states that you would
have to store.  The magnitude of this number is just numbing.

Let's look at it another way.  John Allen Pauls estimates that if the known
universe were filled with just quarks and nothing else, there would be
about 10^143 of them.  A computer with 60 bytes of memory can exist in more
states than this.  So it isn't even possible to store a bitmap of what
states you've visited.

It doesn't help to not store the states, and run two computers at different
speeds to detect the loop.  It just moves the problem from absurd memory
requirements to absurd time requirements, things that make the universe
seem quite young.

Dave

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Gus Wirth]

Andrew Lentvorski wrote:

Ralph Shumaker wrote:

Andrew Lentvorski wrote:

David Brown wrote:

Computers we make have to be laboriously programmed.


Um, excuse me?  Literacy and education don't come for free, you know.


That, and the programming of the silicon is generally done once, on 
one machine, and then just loaded by others like it.  I don't know of 
anyone who has been able to download any of his schooling into the 
analog computer of anyone else.  Compile time applies to each person's 
analog gray matter computer.  No RPMs there.


And, boy, would I welcome the ability to download a foreign language 
directly into my brain.


And what would you choose to get rid of to make way for that new 
language? Although the capacity of the human brain is large, it is not 
infinite.


I wonder how we can measure when we are near the limit of capacity?

Gus

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

Darren New wrote:
That means the processor can only be in one of 8*2^16 different states 


Of course, that would be 2^8*2^16. You get the idea.

--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

Tracy R Reed wrote:

Darren New wrote:
Depends on which analog computer. Most are rather a lot smaller than a 
Turing machine.  Remember the halting problem only exists for 
infinite-storage computers.  Anything with a fixed number of states is 
"trivial" to check whether it halts.


A modern digital computer has a fixed number of states but given the 
definition of a program and a finite input is unable to tell you whether 
that program will run forever or not.


I can tell you whether a program on (for example) a Z80 computer will 
run forever fairly easily, assuming it's a self-contained system.


The Z80 has 2^16 bytes of address space (add the appropriate amount of 
space for registers). That means the processor can only be in one of 
8*2^16 different states (again, modulo register content, which is just 
more memory, and ignoring external stuff like removable media, 
interrupts, keyboards, etc).


Run the processor, making a complete copy of memory after every 
instruction. Continue until you execute a halt instruction or until the 
state of memory matches what you have already seen. In the former case, 
the program halts. In the latter, it is in an infinite loop. Both cases 
will be determined in finite time.


Trivial, except for the cost of storing (8*2^16)^(8*2^16) copies of the 
memory, or so.


The reason you can't solve the halting problem on a Turing machine is 
you don't have an upper limit on the number of states you have to 
examine, since memory is unbounded.


> The last two sentences of the
above paragraph are false. Were they true it would imply that my 
computer with a mere 4G of RAM is not an infinite-storage computer and 
can therefore decide the halting problem.


No, it can't decide the halting problem due to being finite. It can be 
the subject of the decision due to being finite.


It also has a fixed number of 
states but checking whether a program halts is far from trivial.


It's quite trivial if the machine doing the checking has unbounded 
amounts of memory.



Not sure where the halting problem comes in at all, actually.


The human brain can analyze a program and determine whether or not it 
will ever halt.


The halting problem isn't that "there are no programs for which you can 
decide whether it halts." The halting problem is "there are some 
programs for which you cannot decide whether it halts." There are many, 
many programs for which it is easy to determine whether the program 
halts - all programs lacking indefinite loops, for example. The problem 
is whether you can do this for *every* program, and no, the human brain 
cannot look at *every* program and figure out whether it halts.


--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Andrew Lentvorski]

Andrew Lentvorski wrote:

Bob La Quey wrote:


So I have a lot of direct experience with good
people who suffer from the modes of failure. This
experience has caused me to wonder, "Why would
such a gene survive the evolutionary process?


Certainly one of the big questions.

However, there are other good examples.  Sickle-cell anemia, for 
example, exists because if you only have one of the sickle-cell genes, 
you have protection against malaria.  If you have both, well, you didn't 
live long in the tropics.


Also, there is some increasing evidence that "toxoplasma gondii" (a 
fairly common parasite from cats) combines with a genetic predisposition 
to be one of the causes of schizophrenia.



In this respect the mentally ill have a lot in common
with geniuses and other extremely creative individuals
who are driven by internal as opposed to external social
realities.


I would argue that many of the same genes that make us "unusually smart" 
are on the borderline of "diseased".


I'm also of the opinion that we need to start treating things like ADHD 
which are now >10% of the population as a "disease" and more as "human 


Urk ... that should read:

which are now >10% of the population as less a "disease" and more as "human



is simply wired differently".

Not everyone is the same.  We shouldn't treat them like they are.

-a



--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Andrew Lentvorski]

Ralph Shumaker wrote:

Andrew Lentvorski wrote:

David Brown wrote:

Computers we make have to be laboriously programmed.


Um, excuse me?  Literacy and education don't come for free, you know.


That, and the programming of the silicon is generally done once, on one 
machine, and then just loaded by others like it.  I don't know of anyone 
who has been able to download any of his schooling into the analog 
computer of anyone else.  Compile time applies to each person's analog 
gray matter computer.  No RPMs there.


And, boy, would I welcome the ability to download a foreign language 
directly into my brain.


-a

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Andrew Lentvorski]

Bob La Quey wrote:


So I have a lot of direct experience with good
people who suffer from the modes of failure. This
experience has caused me to wonder, "Why would
such a gene survive the evolutionary process?


Certainly one of the big questions.

However, there are other good examples.  Sickle-cell anemia, for 
example, exists because if you only have one of the sickle-cell genes, 
you have protection against malaria.  If you have both, well, you didn't 
live long in the tropics.


Also, there is some increasing evidence that "toxoplasma gondii" (a 
fairly common parasite from cats) combines with a genetic predisposition 
to be one of the causes of schizophrenia.



In this respect the mentally ill have a lot in common
with geniuses and other extremely creative individuals
who are driven by internal as opposed to external social
realities.


I would argue that many of the same genes that make us "unusually smart" 
are on the borderline of "diseased".


I'm also of the opinion that we need to start treating things like ADHD 
which are now >10% of the population as a "disease" and more as "human 
is simply wired differently".


Not everyone is the same.  We shouldn't treat them like they are.

-a

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: [Kooler] Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Paul G. Allen]

Ralph Shumaker wrote:



1)  God created the cosmos perfect but gave Man free will, and choice.
2)  Man chose to sin.
3)  God pronounced judgment which basically amounted to the decay of the 
universe (entropy), specifically including the death of all living 
systems, particularly animals and man.
4)  The reason for death to include animals (and the universe) has to do 
with His plan for redemption of Man (and the universe).
5)  Defective genes, when defective enough, are designed to become 
closer and closer to sterility.  For example, cross-bred species (plants 
*or* animals) are usually sterile (e.g. the mule).
6)  God commanded us to fill the Earth and subdue it.  Despite some 
abusing the meaning of this, He makes it clear that we are merely 
stewards of the Earth and He will hold us accountable for how we take 
care of it.  He gave us all the plants for nourishment which includes 
treatments for ailments.  I wouldn't doubt that our use of the plants 
for finding cures will be one of the things included in the command to 
"subdue" the Earth.




Well, I'm not on Kooler, so I will just say that I agree with Ralph and 
his summary above before the thread moves to Kooler and is lost to me.


I will also add, that to me, to believe that everything I see in nature, 
and everything we know about the universe (and what we are yet to 
learn), came about accidentally, with no intelligent entity creating it, 
is simply beyond scientific reason.


PGA
--
Paul G. Allen, BSIT/SE
Owner, Sr. Engineer
Random Logic Consulting Services
www.randomlogic.com

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

[Kooler] Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Ralph Shumaker]

Bob La Quey wrote:

The syndrome has some common consequences. One of
them is that ones relationships with other humans
almost always deteriorates. Very few people seek
out schizophrenics as friends for obvious reasons.
Especially when the schizophrenic is, as is often
the case, paranoid and potentially (actually rare)
violent.

So I have a lot of direct experience with good
people who suffer from the modes of failure. This
experience has caused me to wonder, "Why would
such a gene survive the evolutionary process? I
would ask Ralph, 'What would this be a part of
an intelligent design?" I must note though that
I am no believer in ID. I am a believer in classical
evolution.
  


Well, I know that no one here wants a long drawn out response, so I will 
attempt to keep it concise.


1)  God created the cosmos perfect but gave Man free will, and choice.
2)  Man chose to sin.
3)  God pronounced judgment which basically amounted to the decay of the 
universe (entropy), specifically including the death of all living 
systems, particularly animals and man.
4)  The reason for death to include animals (and the universe) has to do 
with His plan for redemption of Man (and the universe).
5)  Defective genes, when defective enough, are designed to become 
closer and closer to sterility.  For example, cross-bred species (plants 
*or* animals) are usually sterile (e.g. the mule).
6)  God commanded us to fill the Earth and subdue it.  Despite some 
abusing the meaning of this, He makes it clear that we are merely 
stewards of the Earth and He will hold us accountable for how we take 
care of it.  He gave us all the plants for nourishment which includes 
treatments for ailments.  I wouldn't doubt that our use of the plants 
for finding cures will be one of the things included in the command to 
"subdue" the Earth.


Mind you, this is a very in-a-nutshell version.  It would be easy to 
pick these things to pieces as presented.  They are merely meant as 
generalizations.  I hope they provide a decent example of the way I see 
it.  And hopefully, I answered your question Bob.


And Bob, I realize that there are perhaps 3 members here that may 
believe in ID.  And I'm not even sure about the third one.  If there are 
others besides the 3, I am not aware of them.  So I recognize that the 
ratio of evolutionists to creationists here is perhaps 100:1, more or 
less.  And the other 2 creationists (if they are) don't often get much 
involved in the debates, so it often *feels* like I am alone in my 
stance.  But that's ok.  I do _try_ not to let it get personal.


(And upon proofreading this, I don't see any strong connection to the 
subject line *or* to the context of the -lpsg list in general, so I'm 
adding the [Kooler] tag to the subject line.  Please direct any 
responses to this directly to the -kooler list, or to me personally if 
you prefer.  Thank you.)




--
Ralph


One day I stumbled across a case of Scotch.

As I recall, I stumbled several days thereafter.
--W.C. Fields

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Ralph Shumaker]

Andrew Lentvorski wrote:

David Brown wrote:

Computers we make have to be laboriously programmed.


Um, excuse me?  Literacy and education don't come for free, you know.


That, and the programming of the silicon is generally done once, on one 
machine, and then just loaded by others like it.  I don't know of anyone 
who has been able to download any of his schooling into the analog 
computer of anyone else.  Compile time applies to each person's analog 
gray matter computer.  No RPMs there.




--
Ralph


One day I stumbled across a case of Scotch.

As I recall, I stumbled several days thereafter.
--W.C. Fields

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Ralph Shumaker]

Darren New wrote:

Ralph Shumaker wrote:
Are there general purpose analog computers?  Can they run programs in 
the sense that digital computers can?


You're thinking with one. ;-)


Anyone have the hack to allow me to load gnumeric?  ;)



--
Ralph


One day I stumbled across a case of Scotch.

As I recall, I stumbled several days thereafter.
--W.C. Fields

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Bob La Quey]

On Jan 23, 2008 9:52 PM, Andrew Lentvorski <[EMAIL PROTECTED]> wrote:
> David Brown wrote:
> > I would say that the human brain is more general purpose than any computer
> > that we've ever built.
>
> I might agree.  Except that its "general purpose" is actually "pattern
> matching for survival".
>
> > Computers we make have to be laboriously programmed.
>
> Um, excuse me?  Literacy and education don't come for free, you know.
>
> > The brain learns on its own.  It can't be directly programmed,
> > and I don't think we're even close to understanding how it works.  But, it
> > can be indirectly be programmed.  Words like "study", and "practice" come
> > to mind.
>
> And it has some absolutely spectacular failure modes: superstition,
> schizophrenia, paranoia, depression, etc.

Useful observations. I come from a family where
these "failure modes" are manifest. My father and
one (of two) brothers were paranoid schizophrenic,
clinically diagnosed and often hospitalized for this
illness, a mental illness with a large genetic
component, i.e. we know as a fact that it runs
in families. My family is a good example.

As a consequence I have thought a lot about this
problem. Why, for instance, would evolution tolerate
such an "obviously dysfunctional" disease.

I use the term disease on purpose because I would
note that schizophrenia is a syndrome (perhaps many
diseases) that is in many, if not most cases, a
result of a physical disease of the brain. This
disease(s) causes serious cognitive differences in
brain functioning compared to "normal" people, i.e.
schizophrenics see and hear things that the vast
majority of people around them are _not_ seeing
and hearing. Simply put, they hallucinate.

The syndrome has some common consequences. One of
them is that ones relationships with other humans
almost always deteriorates. Very few people seek
out schizophrenics as friends for obvious reasons.
Especially when the schizophrenic is, as is often
the case, paranoid and potentially (actually rare)
violent.

So I have a lot of direct experience with good
people who suffer from the modes of failure. This
experience has caused me to wonder, "Why would
such a gene survive the evolutionary process? I
would ask Ralph, 'What would this be a part of
an intelligent design?" I must note though that
I am no believer in ID. I am a believer in classical
evolution.

I come up with a couple of speculations for why
schizophrenia survives, i.e. why it has evolutionary
value. However it is late tonight. I am very tired.
I will address this tomorrow in another post.

Meanwhile I would solicit comments and experience
of others.

> And it has almost no debugging facilities.

Not true. The debugging is done by interaction with
other humans. Call it culture if you will. This
environmental debugging works externally and is
very powerful. Most people fall in line with the
external feedback.

One of the facts about the mentally ill, schizophrenia
being only one of several examples, is that they do
_not_ pay nearly as much attention to this external
feedback. They do _not_ debug like you and I.

In this respect the mentally ill have a lot in common
with geniuses and other extremely creative individuals
who are driven by internal as opposed to external social
realities.

BobLQ

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Tracy R Reed]

Tracy R Reed wrote:
The human brain can analyze a program and determine whether or not it 
will ever halt.


I take this part back as it turns out to be untrue. A human cannot 
necessary tell if a program halts or not. It can sometimes tell but not 
always. The halting problem specifies that to be decidable a general 
algorithm which always works must exist and humans can not always 
determine if a program will halt.



--
Tracy R Reed  Read my blog at http://ultraviolet.org
Key fingerprint = D4A8 4860 535C ABF8 BA97  25A6 F4F2 1829 9615 02AD
Non-GPG signed mail gets read only if I can find it among the spam.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Tracy R Reed]

Darren New wrote:
Depends on which analog computer. Most are rather a lot smaller than a 
Turing machine.  Remember the halting problem only exists for 
infinite-storage computers.  Anything with a fixed number of states is 
"trivial" to check whether it halts.


A modern digital computer has a fixed number of states but given the 
definition of a program and a finite input is unable to tell you whether 
that program will run forever or not. The last two sentences of the 
above paragraph are false. Were they true it would imply that my 
computer with a mere 4G of RAM is not an infinite-storage computer and 
can therefore decide the halting problem. It also has a fixed number of 
states but checking whether a program halts is far from trivial.



Not sure where the halting problem comes in at all, actually.


The human brain can analyze a program and determine whether or not it 
will ever halt.


--
Tracy R Reed  Read my blog at http://ultraviolet.org
Key fingerprint = D4A8 4860 535C ABF8 BA97  25A6 F4F2 1829 9615 02AD
Non-GPG signed mail gets read only if I can find it among the spam.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[James G. Sack (jim)]

Andrew Lentvorski wrote:
> David Brown wrote:
>> I would say that the human brain is more general purpose than any
>> computer
>> that we've ever built.
> 
> I might agree.  Except that its "general purpose" is actually "pattern
> matching for survival".
> 
>> Computers we make have to be laboriously programmed.
> 
> Um, excuse me?  Literacy and education don't come for free, you know.
> 
>> The brain learns on its own.  It can't be directly programmed,
>> and I don't think we're even close to understanding how it works. 
>> But, it
>> can be indirectly be programmed.  Words like "study", and "practice" come
>> to mind.
> 
> And it has some absolutely spectacular failure modes: superstition,
> schizophrenia, paranoia, depression, etc.
> 
> And it has almost no debugging facilities.

Har! :-)

Regards,
..jim

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

Tracy R Reed wrote:

Darren New wrote:
Sounds like a "yes" to me. You are running a general purpose analog 
computer, as a brain. It can run programs in the sense that a digital 
computer can, or you'd have a really hard time writing programs, 
compilers, and figuring out where your bugs are.  You would be unable 
to simulate a turing machine with pencil and paper. :-)


This implies that analog computers are not subject to the halting 
problem doesn't it?


Depends on which analog computer. Most are rather a lot smaller than a 
Turing machine.  Remember the halting problem only exists for 
infinite-storage computers.  Anything with a fixed number of states is 
"trivial" to check whether it halts.


What I meant was that if you couldn't simulate a computer program in 
your brain, you couldn't write, debug, or compile code manually.


Not sure where the halting problem comes in at all, actually.

--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

Andrew Lentvorski wrote:

David Brown wrote:
I would say that the human brain is more general purpose than any 
computer

that we've ever built.


I might agree.  Except that its "general purpose" is actually "pattern 
matching for survival".


I don't know of any other computational device that creates other 
computational devices without outside assistance. That seems like a 
pretty powerful device to me.


And it has some absolutely spectacular failure modes: superstition, 
schizophrenia, paranoia, depression, etc.


And it has almost no debugging facilities.


Except science. But that's another discussion. ;-)

--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Andrew Lentvorski]

David Brown wrote:

I would say that the human brain is more general purpose than any computer
that we've ever built.


I might agree.  Except that its "general purpose" is actually "pattern 
matching for survival".



Computers we make have to be laboriously programmed.


Um, excuse me?  Literacy and education don't come for free, you know.


The brain learns on its own.  It can't be directly programmed,
and I don't think we're even close to understanding how it works.  But, it
can be indirectly be programmed.  Words like "study", and "practice" come
to mind.


And it has some absolutely spectacular failure modes: superstition, 
schizophrenia, paranoia, depression, etc.


And it has almost no debugging facilities.

-a

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Tracy R Reed]

Darren New wrote:
Sounds like a "yes" to me. You are running a general purpose analog 
computer, as a brain. It can run programs in the sense that a digital 
computer can, or you'd have a really hard time writing programs, 
compilers, and figuring out where your bugs are.  You would be unable to 
simulate a turing machine with pencil and paper. :-)


This implies that analog computers are not subject to the halting 
problem doesn't it?


--
Tracy R Reed  Read my blog at http://ultraviolet.org
Key fingerprint = D4A8 4860 535C ABF8 BA97  25A6 F4F2 1829 9615 02AD
Non-GPG signed mail gets read only if I can find it among the spam.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Wed, Jan 23, 2008 at 08:24:02PM -0800, SJS wrote:

begin  quoting Darren New as of Wed, Jan 23, 2008 at 07:11:50PM -0800:

SJS wrote:
>begin  quoting Darren New as of Wed, Jan 23, 2008 at 05:40:50PM -0800:
>>Ralph Shumaker wrote:
>>>Are there general purpose analog computers?  Can they run programs in 
>>>the sense that digital computers can?

>>You're thinking with one. ;-)
>
>That would be a big "no", then. 


Sounds like a "yes" to me.


I'm not interpreting "in the sense" in the sense that you seem to be.

   You are running a general purpose analog 
computer, as a brain.


I thought it was pretty specialized, and not general purpose.


I would say that the human brain is more general purpose than any computer
that we've ever built.  Computers we make have to be laboriously
programmed.  The brain learns on its own.  It can't be directly programmed,
and I don't think we're even close to understanding how it works.  But, it
can be indirectly be programmed.  Words like "study", and "practice" come
to mind.

I remember a documentary about these kids in a Japanese school that learned
arithmetic on an abacus.  Many of them would practice to the point where
they didn't actually need the abacus.  They had learned to visualize the
entire abacus and move the beads in their mind.  Interestingly, most of
them still move their fingers in the motions of moving the beads.  These
kids could perform large multiplications entirely in their heads at amazing
speed, on the order of seconds for 5-6 digit number being multiplied.

Dave

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[SJS]

begin  quoting Darren New as of Wed, Jan 23, 2008 at 07:11:50PM -0800:
> SJS wrote:
> >begin  quoting Darren New as of Wed, Jan 23, 2008 at 05:40:50PM -0800:
> >>Ralph Shumaker wrote:
> >>>Are there general purpose analog computers?  Can they run programs in 
> >>>the sense that digital computers can?
> >>You're thinking with one. ;-)
> >
> >That would be a big "no", then. 
> 
> Sounds like a "yes" to me.

I'm not interpreting "in the sense" in the sense that you seem to be.

>You are running a general purpose analog 
> computer, as a brain.

I thought it was pretty specialized, and not general purpose.

>   It can run programs in the sense that a digital 
> computer can, or you'd have a really hard time writing programs, 
> compilers, and figuring out where your bugs are.  You would be unable to 
> simulate a turing machine with pencil and paper. :-)

Um, I don't do stored programs very well.

Although, it would be way cool.

Whoa, I know kung-fu.

-- 
Once again, the horizon does beckon
To the topic, for thread wreckin'
Stewart Stremler

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

SJS wrote:

begin  quoting Darren New as of Wed, Jan 23, 2008 at 05:40:50PM -0800:

Ralph Shumaker wrote:
Are there general purpose analog computers?  Can they run programs in 
the sense that digital computers can?

You're thinking with one. ;-)


That would be a big "no", then. 


Sounds like a "yes" to me. You are running a general purpose analog 
computer, as a brain. It can run programs in the sense that a digital 
computer can, or you'd have a really hard time writing programs, 
compilers, and figuring out where your bugs are.  You would be unable to 
simulate a turing machine with pencil and paper. :-)


--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[SJS]

begin  quoting Darren New as of Wed, Jan 23, 2008 at 05:40:50PM -0800:
> Ralph Shumaker wrote:
> >Are there general purpose analog computers?  Can they run programs in 
> >the sense that digital computers can?
> 
> You're thinking with one. ;-)

That would be a big "no", then. 

:-P

-- 
It's the Matrix:
Too big to fix!
Stewart Stremler

-- 
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[Darren New]

Ralph Shumaker wrote:
Are there general purpose analog computers?  Can they run programs in 
the sense that digital computers can?


You're thinking with one. ;-)

--
  Darren New / San Diego, CA, USA (PST)
It's not feature creep if you put it
at the end and adjust the release date.

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg

Re: Analog vs Digital (was Re: Is Scheme/Lisp somehow more "fundamental" than other languages?)

[David Brown]

On Wed, Jan 23, 2008 at 10:54:04AM -0800, Ralph Shumaker wrote:

Are there general purpose analog computers?  Can they run programs in the 
sense that digital computers can?


I'm not aware of any.  The idea of a stored program is more of a digital
concept.  Analog computers are typically programmed by changing their
configuration.

I think the most familiar analog computer to most people would be a
pre-1990's automatic transmission valve body.  This functionality has
largely been passed off to a computer that can get a lot more inputs from
the engine and other places.  Still, the old transmissions worked amazingly
well.  The mechanic who rebuilt my 300TD transmission was an amazing
artisan.  He understood the mechanisms well enough to even make changes in
the performance.

Dave

--
KPLUG-LPSG@kernel-panic.org
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg