Dear All,
 I saw the phrase "Information really is matter/energy" somewhere. Is
there really a connection between the Information Theory and
matter/energy?

 How will things change if we find something new in Information
Theory?
Will it also affect the rules of physics or chemistry? Are shannon's
theorems tightly connected to the real world?

Thanks

Generally unrelated.  Energy per bit is used to compare efficiencies of
modulation schemes. Just a dimensional factor.
Things have changed, checkout Turbo Coding.
Physics and Chem are basic.
Shannons theorem on intersymbol interference is based upon assumptions that
were surpassed about 20 years ago.

"permutation" <permutationcombination7@hotmail.com> wrote in message
news:4e2be3e5.0404191021.1e2894aa@posting.google.com...
> Dear All,
>  I saw the phrase "Information really is matter/energy" somewhere. Is
> there really a connection between the Information Theory and
> matter/energy?
>
>  How will things change if we find something new in Information
> Theory?
> Will it also affect the rules of physics or chemistry? Are shannon's
> theorems tightly connected to the real world?
>
> Thanks

permutation wrote:
> 
> Dear All,
>  I saw the phrase "Information really is matter/energy" somewhere. Is
> there really a connection between the Information Theory and
> matter/energy?

Nice reference you've got there.  Scholarly.  It demonstrates the
difference between groundlings and their betters.  You illustrate why
"everybody goes to college" is bullshit.  One doubts more than 10% of
the population could benefit from higher education.  Start by
eliminating more than 300,000 BS and BA degrees in Liberal Arts/year. 
Don't produce what you cannot sell.  Objective standards and quality
assurance make a difference.

Read Claude Shannon,
<http://www.essrl.wustl.edu/~jao/itrg/shannon.pdf>
<http://cm.bell-labs.com/cm/ms/what/shannonday/shannon1948.pdf>
 (Bell Labs is now Lucent Techologies, being expertly managed.  Bell
Labs now sucks dead donkey dicks.)

draw your own conclusions. If you have no conclusions, think about
something you *can* think about.  Are you confused about the energy
needed to compute?  Only erasure consumes net energy - in theory.

[snip crap]

Uncle Al says, "Yesterday's Nobel Prize is tomorrow's homework
assignment."

-- 
Uncle Al
http://www.mazepath.com/uncleal/
 (Toxic URL! Unsafe for children and most mammals)
"Quis custodiet ipsos custodes?"  The Net!

On 19 Apr 2004 11:21:32 -0700, permutationcombination7@hotmail.com
(permutation) wrote:

>Dear All,
> I saw the phrase "Information really is matter/energy" somewhere. Is
>there really a connection between the Information Theory and
>matter/energy?

imho the answer is no, but there is a book about this
topic if you are interested:

Stonier, Tom: 
  Information and the internal structure of the universe : 
  An exploration into information physics

> How will things change if we find something new in Information
>Theory?
>Will it also affect the rules of physics or chemistry? Are shannon's
>theorems tightly connected to the real world?


br,
andras

NO LONGER WRT:  Is Information really  matter/energy, (randomness and
entropy)


I was going to say something else, then I read what "Uncle Al" had to
say...
I *may* have met my match, that much is clear.  So instead I thought
I'd challenge this uncle Al fellow to a duel.  And not owning either
swords or weapons, I decided that the duel should be in the domain of
anti-randomness.

As anyone knows (whose been paying attention, non-randomness can be
expressed as
either compression or predictions.)  These days I don't bother to
compress, I predict.

I know who will be the next US president, and a whole lot of other
things that I have cajoled out of the depths of my computer.  But that
stuff is far off.  For today, (well, tomorrow) I will do something
more definite.

So if you accept my challenge Al, my first 'shot' will be to predict a
few stocks tomorrow.  Before the US markets open (at 9:30AM) I will
post the OPEN, HIGH, LOW, and CLOSE'ing prices for some major US
stocks.

And what can you do?

--jg

PS:  You'll need to post your actual name and email address if you
accept my challenge.  I want to know who you are.

Andras Erdei <ccg@freemail.c3.hu> wrote in message 

 I thought of entropy. Entropy in Shannon's domain is very similar to
the entropy in Boltzman's domain which is Thermodynamics i.e. Physics
and Chemistry.
 When the understanding of entropy and randomness in Information
Theory change in the future, will it also imply changes in
thermodynamics?
 Also there is a topic called "Molecular Information Theory"  -Theory
of Molecular Machines by Tom Schneider which is based on Shannon's
Info theory. This topic is a part of biology. Will also biology
change???
 
 
> imho the answer is no, but there is a book about this
> topic if you are interested:
> 
> Stonier, Tom: 
>   Information and the internal structure of the universe : 
>   An exploration into information physics
> 
> 
> 
> br,
> andras

julesg@financier.com (Jules Gilbert) wrote 

> As anyone knows (whose been paying attention, non-randomness can be
> expressed as
> either compression or predictions.)  These days I don't bother to
> compress, I predict.
> 
> I know who will be the next US president, and a whole lot of other
> things that I have cajoled out of the depths of my computer.  But that
> stuff is far off.  For today, (well, tomorrow) I will do something
> more definite.

Jaysesskryst Gilbert, somehow I was under the impression that you are
only one sorry deluded wretch of a crank. Now you are trying for a
full featured big time lunatic. Makes one feel really sorry for you.

[snip] 

> 
> --jg
> 
> PS:  You'll need to post your actual name and email address if you
> accept my challenge.  I want to know who you are.

Uncle Al is well known in usenet, you can google up his data. But why
bother? Predict it or uncompress it...

An aside to Gunther:  So you are sick and ugly -- don't make the rest
of us like you are.  Okay?

But for those people who are serious scientists, this is what I mean
when I say I can predict.  Here are some prices of major stocks and
index's for tomorrow.

2004/04/23          SPY:     $114.17     $114.64     $113.45    
$114.19
2004/04/23          QQQ:     $36.86      $37.09      $36.55     
$36.78
2004/04/23         BRKB:     $3046.92    $3056.33    $3023.34   
$3040.79
2004/04/23         BRKA:     $92232.26   $92663.70   $91844.04  
$92153.18
2004/04/23         INTC:     $26.42      $26.62      $26.05     
$26.31
2004/04/23         MSFT:     $25.92      $26.14      $25.68     
$25.87
2004/04/23           GE:     $30.73      $30.90      $30.38     
$30.69
2004/04/23         TASR:     $75.98      $79.24      $73.86     
$77.19
2004/04/23         CSCO:     $23.07      $23.32      $22.79     
$23.02

I am posting this about 11:40 PM or so, Thursday night.

julesg@financier.com (Jules Gilbert) wrote in 
news:9c7d0936.0404221930.116d5432@posting.google.com:

> An aside to Gunther:  So you are sick and ugly -- don't make the rest
> of us like you are.  Okay?
> 
> But for those people who are serious scientists, this is what I mean
> when I say I can predict.  Here are some prices of major stocks and
> index's for tomorrow.
> 
> 2004/04/23          SPY:     $114.17     $114.64     $113.45    
> $114.19
> 2004/04/23          QQQ:     $36.86      $37.09      $36.55     
> $36.78
> 2004/04/23         BRKB:     $3046.92    $3056.33    $3023.34   
> $3040.79
> 2004/04/23         BRKA:     $92232.26   $92663.70   $91844.04  
> $92153.18
> 2004/04/23         INTC:     $26.42      $26.62      $26.05     
> $26.31
> 2004/04/23         MSFT:     $25.92      $26.14      $25.68     
> $25.87
> 2004/04/23           GE:     $30.73      $30.90      $30.38     
> $30.69
> 2004/04/23         TASR:     $75.98      $79.24      $73.86     
> $77.19
> 2004/04/23         CSCO:     $23.07      $23.32      $22.79     
> $23.02
> 
> I am posting this about 11:40 PM or so, Thursday night.
> 

  But is it complete??  You have 4 numbers after each with
no labels.  Say what it is before tomorrow.  What do you claim
them to mean? High low average RMS average what??


David A. Scott
-- 
My Crypto code
http://cryptography.org/cgi-bin/crypto.cgi/Misc/scott19u.zip
http://cryptography.org/cgi-bin/crypto.cgi/Misc/scott16u.zip
http://www.jim.com/jamesd/Kong/scott19u.zip old version
My Compression code http://bijective.dogma.net/
**TO EMAIL ME drop the roman "five" **
Disclaimer:I am in no way responsible for any of the statements
 made in the above text. For all I know I might be drugged.
As a famous person once said "any cryptograhic
system is only as strong as its weakest link"

On 22 Apr 2004 20:30:20 -0700, julesg@financier.com (Jules Gilbert)
wrote:

>An aside to Gunther:  So you are sick and ugly -- don't make the rest
>of us like you are.  Okay?
>
>But for those people who are serious scientists

Funny :-D

>this is what I mean
>when I say I can predict.  Here are some prices of major stocks and
>index's for tomorrow.

[snip]

>$77.19
>2004/04/23         CSCO:     $23.07      $23.32      $22.79     
>$23.02
>
>I am posting this about 11:40 PM or so, Thursday night.

I think you've just reached a new low in your +/- 10 years of dubious
usenet postings :) 

permutationcombination7@hotmail.com (permutation) wrote in message news:<4e2be3e5.0404200900.5482ac76@posting.google.com>...
> Andras Erdei <ccg@freemail.c3.hu> wrote in message 
> 
>  I thought of entropy. Entropy in Shannon's domain is very similar to
> the entropy in Boltzman's domain which is Thermodynamics i.e. Physics
> and Chemistry.
>  When the understanding of entropy and randomness in Information
> Theory change in the future, will it also imply changes in
> thermodynamics?


  It will have zero impact as statistical dynamics
  also had zero impact. Since as philosphers 
  have been told since 1800, Themodynamics
  has nothing to do with random, idiots.

  It's not an average of something weirdos.

Sorry David:  I really am not HTML or web-savvy.  

The four numbers are the OPEN, HIGH, LOW, and CLOSE'ing prices.

Each night, after the US markets close, my programs run and compute
the OHLC for most traded stocks.

At the end of today, either I will be right or wrong.  It's rare that
I am accurate to the penny of course.

The key issue is:  Did my program calculate each price closely enough
for an investor to trade advantageously.  Many of us in my trading
group know the answer.  We use my program to trade.

I won't pretend that my system is perfect.  It's not.  And I am not a
particularly good trader, either -- though I have friends who, enabled
with my data, are spectacular.  Even I, until 9/11, after hundred's of
trades, had made at least 1% on all but six trades.  (When the market
was against the direction I was trading, I had to bail out.)  Today
though another program predicts the market for me.  So I always trade
with the market -- makes things a whole lot easier, though it's
impossible to predict the market perfectly -- it's a whole lot less
predictable than most individual stocks.  Because non-market events
affect the market.  But about 90% of the time, my market-forecasting
system just works.

Funny, in most highly dynamical systems, knowing the actions of the
underlying process components tells you what the overall change will
be.  Not so the market.  Too many unforeseen events.  I call them
non-market events.





"David A. Scott" <daVvid_a_scott@email.com> wrote in message news:<Xns94D3E4480E084H110W296LC45WIN3030R@130.133.1.4>...
> julesg@financier.com (Jules Gilbert) wrote in 
> news:9c7d0936.0404221930.116d5432@posting.google.com:
> 
> > An aside to Gunther:  So you are sick and ugly -- don't make the rest
> > of us like you are.  Okay?
> > 
> > But for those people who are serious scientists, this is what I mean
> > when I say I can predict.  Here are some prices of major stocks and
> > index's for tomorrow.
> > 
> > 2004/04/23          SPY:     $114.17     $114.64     $113.45    
> > $114.19
> > 2004/04/23          QQQ:     $36.86      $37.09      $36.55     
> > $36.78
> > 2004/04/23         BRKB:     $3046.92    $3056.33    $3023.34   
> > $3040.79
> > 2004/04/23         BRKA:     $92232.26   $92663.70   $91844.04  
> > $92153.18
> > 2004/04/23         INTC:     $26.42      $26.62      $26.05     
> > $26.31
> > 2004/04/23         MSFT:     $25.92      $26.14      $25.68     
> > $25.87
> > 2004/04/23           GE:     $30.73      $30.90      $30.38     
> > $30.69
> > 2004/04/23         TASR:     $75.98      $79.24      $73.86     
> > $77.19
> > 2004/04/23         CSCO:     $23.07      $23.32      $22.79     
> > $23.02
> > 
> > I am posting this about 11:40 PM or so, Thursday night.
> > 
> 
>   But is it complete??  You have 4 numbers after each with
> no labels.  Say what it is before tomorrow.  What do you claim
> them to mean? High low average RMS average what??
> 
> 
> David A. Scott
> -- 
> My Crypto code
> http://cryptography.org/cgi-bin/crypto.cgi/Misc/scott19u.zip
> http://cryptography.org/cgi-bin/crypto.cgi/Misc/scott16u.zip
> http://www.jim.com/jamesd/Kong/scott19u.zip old version
> My Compression code http://bijective.dogma.net/
> **TO EMAIL ME drop the roman "five" **
> Disclaimer:I am in no way responsible for any of the statements
>  made in the above text. For all I know I might be drugged.
> As a famous person once said "any cryptograhic
> system is only as strong as its weakest link"

On 23 Apr 2004 04:02:03 -0700, ZZBunker <zzbunker@netscape.net> wrote:
> permutationcombination7@hotmail.com (permutation) wrote in message news:<4e2be3e5.0404200900.5482ac76@posting.google.com>...
>> Andras Erdei <ccg@freemail.c3.hu> wrote in message 
>> 
>>  I thought of entropy. Entropy in Shannon's domain is very similar to
>> the entropy in Boltzman's domain which is Thermodynamics i.e. Physics
>> and Chemistry.
>>  When the understanding of entropy and randomness in Information
>> Theory change in the future, will it also imply changes in
>> thermodynamics?
> 
> 
>   It will have zero impact as statistical dynamics
>   also had zero impact. Since as philosphers 
>   have been told since 1800, Themodynamics
>   has nothing to do with random, idiots.

Since the unification of measure theory and probability, probability
can be used as description of the averaged effects of deterministic
dynamics.  

Kolmogorov was not an idiot, and the confluence of Shannon's entropy
and thermodynamic entropy is not a spurious coincidence, even though
assumed mechanisms might superficially appear different.  

>   It's not an average of something weirdos.

Dr Chaos <mbkennelSPAMBEGONE@NOSPAMyahoo.com> wrote in message news:<slrnc8iqg0.ndl.mbkennelSPAMBEGONE@lyapunov.ucsd.edu>...
> On 23 Apr 2004 04:02:03 -0700, ZZBunker <zzbunker@netscape.net> wrote:
> > permutationcombination7@hotmail.com (permutation) wrote in message news:<4e2be3e5.0404200900.5482ac76@posting.google.com>...
> >> Andras Erdei <ccg@freemail.c3.hu> wrote in message 
> >> 
> >>  I thought of entropy. Entropy in Shannon's domain is very similar to
> >> the entropy in Boltzman's domain which is Thermodynamics i.e. Physics
> >> and Chemistry.
> >>  When the understanding of entropy and randomness in Information
> >> Theory change in the future, will it also imply changes in
> >> thermodynamics?
> > 
> > 
> >   It will have zero impact as statistical dynamics
> >   also had zero impact. Since as philosphers 
> >   have been told since 1800, Themodynamics
> >   has nothing to do with random, idiots.
> 
> Since the unification of measure theory and probability, probability
> can be used as description of the averaged effects of deterministic
> dynamics.  

  averaged *effects* are not the issue, as physics
  tarts are advised. Since effects are not
  *thermal* RETARDS.

Dear Superfly:

The real question is how accurate are my predictions.  Remember,
compression and prediction are closely related (I say this making use
of the model of a single bit stream wherein one doesn't have to send a
partial bit if that bit can be predicted on the RCVE side of the
channel.)

You snipped away everything but CSCO, So let's see how good or bad my
program did with that stock.

CSCO opened at $23.25. whereas I said $23.07; now that's not perfect,
I agree, but it is less than 1% off.  (The difference is 18 cents.)

As for the high Friday, CSCO reached $23.44, whereas I said $23.32;
and again, that's not perfect, but the difference is only 12 cents,
and this is about 1/2 of one percent.  Not a lot.)

For the low, CSCO dropped to $22.87, whereas I has said $22.79, An
eight cent error this time, or about 1/3 of one percent.

Finally CSCO closed at $23.32, whereas my program had predicted a
CLOSE of $23.02.  You know what; This is a huge error.  Not
acceptable.

Every day my program produces predictions like these and this
information is published on the web for the benefit of other
investors.

But for me CSCO wasn't important yesterday.  Because yesterday my
program liked TASR the most.  Not CSCO.  TASR.  It popped up about %5.
 And people, even novice traders, knew this before the market opened
because they got on the web.

And no my program isn't very good at predicting the exact price of
stocks that are in "play".  But so what -- it does identify which
stocks are going to soar 'tomorrow' (over the course of the next
trading day.)  So price isn't everything.  (Though I am trying to fix
this problem, and if any of you want to help, please email me.)

Trading is about correctly predicting the change in short-term value. 
For CSCO, the difference between the HIGH and the LOW was 57 cents,
whereas I had predicted a difference in value of 53 cents.

--jg




SuperFly <no@mail.com> wrote in message news:<1orh80hhngbll805104g0tkd2og2j1n1ma@4ax.com>...
> On 22 Apr 2004 20:30:20 -0700, julesg@financier.com (Jules Gilbert)
> wrote:
> 
> >An aside to Gunther:  So you are sick and ugly -- don't make the rest
> >of us like you are.  Okay?
> >
> >But for those people who are serious scientists
> 
> Funny :-D
> 
> >this is what I mean
> >when I say I can predict.  Here are some prices of major stocks and
> >index's for tomorrow.
> 
> [snip]
> 
> >$77.19
> >2004/04/23         CSCO:     $23.07      $23.32      $22.79     
> >$23.02
> >
> >I am posting this about 11:40 PM or so, Thursday night.
> 
> I think you've just reached a new low in your +/- 10 years of dubious
> usenet postings :)

"Jules Gilbert" <julesg@financier.com> wrote in message
news:9c7d0936.0404221930.116d5432@posting.google.com...
> An aside to Gunther:  So you are sick and ugly -- don't make the rest
> of us like you are.  Okay?
>
> But for those people who are serious scientists, this is what I mean
> when I say I can predict.  Here are some prices of major stocks and
> index's for tomorrow.
>
> 2004/04/23          SPY:     $114.17     $114.64     $113.45
> $114.19
> 2004/04/23          QQQ:     $36.86      $37.09      $36.55
> $36.78
> 2004/04/23         BRKB:     $3046.92    $3056.33    $3023.34
> $3040.79
> 2004/04/23         BRKA:     $92232.26   $92663.70   $91844.04
> $92153.18
> 2004/04/23         INTC:     $26.42      $26.62      $26.05
> $26.31
> 2004/04/23         MSFT:     $25.92      $26.14      $25.68
> $25.87
> 2004/04/23           GE:     $30.73      $30.90      $30.38
> $30.69
> 2004/04/23         TASR:     $75.98      $79.24      $73.86
> $77.19
> 2004/04/23         CSCO:     $23.07      $23.32      $22.79
> $23.02
>
> I am posting this about 11:40 PM or so, Thursday night.

So how did these predictions turn out?

--
Tom Potter     http://tompotter.us

julesg@financier.com (Jules Gilbert) wrote in message news:<9c7d0936.0404240516.17e2f3b4@posting.google.com>...
> Dear Superfly:
> 
> The real question is how accurate are my predictions.  Remember,
> compression and prediction are closely related (I say this making use
> of the model of a single bit stream wherein one doesn't have to send a
> partial bit if that bit can be predicted on the RCVE side of the
> channel.)
> 
> You snipped away everything but CSCO, So let's see how good or bad my
> program did with that stock.
> 
> CSCO opened at $23.25. whereas I said $23.07; now that's not perfect,
> I agree, but it is less than 1% off.  (The difference is 18 cents.)
> 
> As for the high Friday, CSCO reached $23.44, whereas I said $23.32;
> and again, that's not perfect, but the difference is only 12 cents,
> and this is about 1/2 of one percent.  Not a lot.)
> 
> For the low, CSCO dropped to $22.87, whereas I has said $22.79, An
> eight cent error this time, or about 1/3 of one percent.
> 
> Finally CSCO closed at $23.32, whereas my program had predicted a
> CLOSE of $23.02.  You know what; This is a huge error.  Not
> acceptable.
> 
> Every day my program produces predictions like these and this
> information is published on the web for the benefit of other
> investors.
> 
> But for me CSCO wasn't important yesterday.  Because yesterday my
> program liked TASR the most.  Not CSCO.  TASR.  It popped up about %5.
>  And people, even novice traders, knew this before the market opened
> because they got on the web.
> 
> And no my program isn't very good at predicting the exact price of
> stocks that are in "play".  But so what -- it does identify which
> stocks are going to soar 'tomorrow' (over the course of the next
> trading day.)  So price isn't everything.  (Though I am trying to fix
> this problem, and if any of you want to help, please email me.)
> 
> Trading is about correctly predicting the change in short-term value. 
> For CSCO, the difference between the HIGH and the LOW was 57 cents,
> whereas I had predicted a difference in value of 53 cents.
> 
> --jg


TASR closed Friday at 81.15 up 4.00, which means that it closed
Thursday at 77.15.  But on Thursday you predicted that it would close
Friday at 77.19.

That's not much of a buy signal!

You predicted it would close .04 cents higher, and it actually closed
4.00 higher!

Other than for hyper-volatile TASR, it was a fairly flat day for the
market, which would explain why most of your other predictions ended
up close.  But you can't make money on a flat market day.

Double-A

julesg@financier.com (Jules Gilbert) wrote in message news:<9c7d0936.0404221930.116d5432@posting.google.com>...
> An aside to Gunther:  So you are sick and ugly -- don't make the rest
> of us like you are.  Okay?

Sorry Gilbert, I dint't mean to offend. It's just you're getting so
pathetic. I do really feel sorry for you, as do a lot others. A few
weeks ago I challenged you to prove your claims regarding compression
by individually compressing 16 different four bit binary strings. You
didn't, instead you tried to talk yourself out of the challenge by
posing as an idiot. Now I have a new challenge for you:
On April the 25th 2005, transfer one billion us dollars to my account.
Since you can predict the value of stocks, that shouldn't be much to
ask of you.
Now let me do some fortune tellin' of my own:
"On April the 25th 2005, Jules Gilbert will NOT transfer a billion us
dollars to my account". Now, that's a prophesy you can bet on. Hold
your breath ladies and gentlemen, you are being witnesses to the first
ever documented 100% accurate prediction of a future event!.


> But for those people who are serious scientists, this is what I mean
> when I say I can predict.  Here are some prices of major stocks and
> index's for tomorrow.
> 
> 2004/04/23          SPY:     $114.17     $114.64     $113.45    
> $114.19
> 2004/04/23          QQQ:     $36.86      $37.09      $36.55     
> $36.78
> 2004/04/23         BRKB:     $3046.92    $3056.33    $3023.34   
> $3040.79
> 2004/04/23         BRKA:     $92232.26   $92663.70   $91844.04  
> $92153.18
> 2004/04/23         INTC:     $26.42      $26.62      $26.05     
> $26.31
> 2004/04/23         MSFT:     $25.92      $26.14      $25.68     
> $25.87
> 2004/04/23           GE:     $30.73      $30.90      $30.38     
> $30.69
> 2004/04/23         TASR:     $75.98      $79.24      $73.86     
> $77.19
> 2004/04/23         CSCO:     $23.07      $23.32      $22.79     
> $23.02
> 
> I am posting this about 11:40 PM or so, Thursday night.

"Jules Gilbert" <julesg@financier.com> schrieb im Newsbeitrag

> An aside to Gunther:  So you are sick and ugly -- don't make the rest
> of us like you are.  Okay?
I think G�nther is very smart, nothing else.

> But for those people who are serious scientists, this is what I mean
> when I say I can predict.  Here are some prices of major stocks and
> index's for tomorrow.
You know, there is no such thing like a trend line. If you believe so, go 
play with your tarot cards.
The stock market has by far too many parameters than anyone can pack this in 
a "prediction" algorithm.
To approach scientists you need to do better. At least you failed here.

-- 
Regards
J�rgen

http://cpicture.de/en 

> The stock market has by far too many parameters than anyone can pack this in 
> a "prediction" algorithm.
> To approach scientists you need to do better. At least you failed here.

Does this mean I have to give back the money I made in IPIX and NOK
today?

And an aside to "double-a":  you misread what I published here a few
days ago.  The info was for Friday, not Thursday.


But fear not; when it suits my purposes, I do intend to give a demo of
my perpetual compressor program.  I am sure that you folks will
proclaim that it was a fraud -- or that the judges were biased, or
that the earth is flat -- or something!  But admit you were wrong? 
Never!

It comes down to this;  the "counting argument" is not wrong, but you
folks misuse what it actually implies.

My compressors don't accept ANY input, just previously compressed
input.  They are also a lot slower than conventional compressors.  Oh
yes, one more thing; they work again and again.  Over and over again. 
Don't you get it?

And while I sometimes think about just giving my secret up, so far at
least, I've thought again.  Who knows, maybe someday I will just give
it away.  But years ago I promised myself that none of you would
profit from my work.  As long as I get gruff, why should I change my
plan?

Please note that I am not saying that I have something that would
compete with any of the commonly used systems, such as zip, arj, rar,
bzip, or pkzip.  No, my system is much slower -- far too slow to be of
practical use unless the program were implemented in hardware.

And I have had some technical problems.  For example, for a while my
"latest and greatest" version would happily compress files, on
occasion creating a file it couldn't decompress.  But that is was an
easily fixed bug once I realized this problem existed.  My point is
that I am not saying I have something that would wipe away gzip or
other popular free compressors.  (I use gzip in my shop all the time. 
It just works.)

Their are two academics I trust sufficiently to tell what I do.  One
is an MIT prof. (a dept. head in LCS,) the other a senior prof at UNF
in EE/CS.

But neither will accept the information and agree not to tell what is
disclosed, which is what I would require.

So we are at this impasse. 



"J�rgen Eidt" <JurgenE@hotmail.com> wrote in message news:<c6ihg8$cc20e$1@ID-47020.news.uni-berlin.de>...
> "Jules Gilbert" <julesg@financier.com> schrieb im Newsbeitrag
> 
> 
> > But for those people who are serious scientists, this is what I mean
> > when I say I can predict.  Here are some prices of major stocks and
> > index's for tomorrow.
> You know, there is no such thing like a trend line. If you believe so, go 
> play with your tarot cards.
> The stock market has by far too many parameters than anyone can pack this in 
> a "prediction" algorithm.
> To approach scientists you need to do better. At least you failed here.

Jules Gilbert <julesg@financier.com> wrote:

> CSCO opened at $23.25. whereas I said $23.07; now that's not perfect,
> I agree, but it is less than 1% off.  (The difference is 18 cents.)

Hmmm....  you claimed it would open $0.08 lower than the previous
day's close, and yet it opened $0.10 higher.

....

> Finally CSCO closed at $23.32, whereas my program had predicted a
> CLOSE of $23.02.  You know what; This is a huge error.  Not
> acceptable.

Yes, you claimed the close would be a $0.13 drop from the previous
day's close, and yet it really had a $0.17 *gain*.  Yes, I'd say
that's pretty unacceptable.

-- 

That's News To Me!
newstome@comcast.net

Hi Tom;  

What an interesting website!  Could you email me a useable email
address, please.

My assumption is that you are in China presently.  Is that right? 
Anyway, a working email address would be appreciated.

--jg

I certainly don't dispute your math or your logic, but for short-term
players, what happened yesterday usually isn't relevant.  We normally
get in and get out of a stock within a single trading day.

I was interested in this thread because I think that our understanding
of 'randomness' is woefully inadaquate  -- and I, for one, don't
believe that such a quality actually exists.  I understand that we see
it, in compressed files, for example.

--jg

Jules Gilbert wrote:

> My compressors don't accept ANY input, just previously compressed
> input.  They are also a lot slower than conventional compressors.  Oh
> yes, one more thing; they work again and again.  Over and over again.
> Don't you get it?

I take it that during development you've tested this with any number of files.

Take three of these files and keep compressing them over and over again.
Each of these three files are now only one bit long.
(At least) two of these files are the same.

How do you get from this file to both the initial inputs?
Starting with just this file how do you know which one of the two starting files it was meant to be?

The same argument applies whenever you're arguing in favour of compressing every input file as you cannot
have a 1-1 mapping between input and output files without accepting that you cannot compress all files.
Previously compressed input is actually worse as you're saying you can only compress some files.

If you can't get every file down to 1 bit and only to (say) 100 bits then simply take more than 2^100 files
and you're shoving pigeons in more than one hole again...

This applies globally, there is NO way around this.

Regards,
Mike...

Jules wrote:
) It comes down to this;  the "counting argument" is not wrong, but you
) folks misuse what it actually implies.
)
) My compressors don't accept ANY input, just previously compressed
) input.  They are also a lot slower than conventional compressors.  Oh
) yes, one more thing; they work again and again.  Over and over again. 
) Don't you get it?

No.

Suppose there is a subset S of all files, this subset consists of all files
that your compressor can compress (i.e. previously compressed input).

To be able to compress over and over again, your compressor output has to
be in this subset S.  (Otherwise you can't recompress it.)

Simply substitute S for the set of possible files in the counting theorem,
and voila.  Your claim is refuted.  Have a nice day.


SaSW, Willem
-- 
Disclaimer: I am in no way responsible for any of the statements
            made in the above text. For all I know I might be
            drugged or something..
            No I'm not paranoid. You all think I'm paranoid, don't you !
#EOT

julesg@financier.com (Jules Gilbert) wrote in 
news:9c7d0936.0404262017.100bd38b@posting.google.com:

> 
> I was interested in this thread because I think that our understanding
> of 'randomness' is woefully inadaquate  -- and I, for one, don't
> believe that such a quality actually exists.  I understand that we see
> it, in compressed files, for example.
> 
> 

  Did you ever think about writting speeches for Kerry somehow this
seemd like something he would say.


David A. Scott
-- 
My Crypto code
http://cryptography.org/cgi-bin/crypto.cgi/Misc/scott19u.zip
http://cryptography.org/cgi-bin/crypto.cgi/Misc/scott16u.zip
http://www.jim.com/jamesd/Kong/scott19u.zip old version
My Compression code http://bijective.dogma.net/
**TO EMAIL ME drop the roman "five" **
Disclaimer:I am in no way responsible for any of the statements
 made in the above text. For all I know I might be drugged.
As a famous person once said "any cryptograhic
system is only as strong as its weakest link"

Hello Willem;

Thank you for making your remarks cogent.  I wasn't clear on what I
meant; let me try again.

the output my system produces is 'random' (yes, I don't believe that
this concept is primitive and valid -- but I use it anyway, it's so
convienient.)

Say, for a minute, that my program produced non-RAD
('random-appearing-data').  Such 'text' would not be suitable as input
to the next stage.

I admit I didn't say this; sorry...

--jg



Willem <willem@stack.nl> wrote in message news:<slrnc8snr0.2fgc.willem@toad.stack.nl>...
> Jules wrote:
> ) It comes down to this;  the "counting argument" is not wrong, but you
> ) folks misuse what it actually implies.
> )
> ) My compressors don't accept ANY input, just previously compressed
> ) input.  They are also a lot slower than conventional compressors.  Oh
> ) yes, one more thing; they work again and again.  Over and over again. 
> ) Don't you get it?
> 
> No.
> 
> Suppose there is a subset S of all files, this subset consists of all files
> that your compressor can compress (i.e. previously compressed input).
> 
> To be able to compress over and over again, your compressor output has to
> be in this subset S.  (Otherwise you can't recompress it.)
> 
> Simply substitute S for the set of possible files in the counting theorem,
> and voila.  Your claim is refuted.  Have a nice day.
> 
> 
> SaSW, Willem

The process of compression does not alway make files one bit in length
(as in never!)

The process *sometimes* makes files (ie., buffers, all files are
broken up into buffers) only a little smaller than the input buffer.

I use a small buffer, of 128 bytes.  Why?  Because a 'file' consists
of a serial list of buffers, and, as compression occurs. the length of
of this file gets shorter.  The compression process produces RAD
output (RAD == 'random appearing data') for each buffer WHICH IS
SMALLER, maybe 10% smaller, maybe 40% smaller.

Why 128 bytes?  Because this is the dominant controlling parameter on
the smallest size file we can emit.  With 128 byte buffer's, I can
produce final files as small as 150-300 bytes.

And no, I can not represent every possible input.  Some inputs would
fail (I XOR the input -- all the core process sees is RAD.)

I've have to go over this again and again, as new grad students come
through.

--jg




Michael Collins <mcollins@arm.com> wrote in message news:<408E5600.E936EE7D@arm.com>...
> Jules Gilbert wrote:
> 
> > My compressors don't accept ANY input, just previously compressed
> > input.  They are also a lot slower than conventional compressors.  Oh
> > yes, one more thing; they work again and again.  Over and over again.
> > Don't you get it?
> 
> I take it that during development you've tested this with any number of files.
> 
> Take three of these files and keep compressing them over and over again.
> Each of these three files are now only one bit long.
> (At least) two of these files are the same.
> 
> How do you get from this file to both the initial inputs?
> Starting with just this file how do you know which one of the two starting files it was meant to be?
> 
> The same argument applies whenever you're arguing in favour of compressing every input file as you cannot
> have a 1-1 mapping between input and output files without accepting that you cannot compress all files.
> Previously compressed input is actually worse as you're saying you can only compress some files.
> 
> If you can't get every file down to 1 bit and only to (say) 100 bits then simply take more than 2^100 files
> and you're shoving pigeons in more than one hole again...
> 
> This applies globally, there is NO way around this.
> 
> Regards,
> Mike...

Jules wrote:
) And no, I can not represent every possible input.  Some inputs would
) fail (I XOR the input -- all the core process sees is RAD.)

Small question:  What percentage of possible inputs are RAD ?  Roughly ?


SaSW, Willem
-- 
Disclaimer: I am in no way responsible for any of the statements
            made in the above text. For all I know I might be
            drugged or something..
            No I'm not paranoid. You all think I'm paranoid, don't you !
#EOT

The process of compression does not alway make files one bit in length
(as in never!)

The process *sometimes* makes files (ie., buffers, all files are
broken up into buffers) only a little smaller than the input buffer.

I use a small buffer, of 128 bytes.  Why?  Because a 'file' consists
of a serial list of buffers, and, as compression occurs. the length of
of this file gets shorter.  The compression process produces RAD
output (RAD == 'random appearing data') for each buffer WHICH IS
SMALLER, maybe 10% smaller, maybe 40% smaller.

Why 128 bytes?  Because this is the dominant controlling parameter on
the smallest size file we can emit.  With 128 byte buffer's, I can
produce final files as small as 150-300 bytes.

And no, I can not represent every possible input.  Some inputs would
fail (I XOR the input -- all the core process sees is RAD.)

I've have to go over this again and again, as new grad students come
through.

--jg




Michael Collins <mcollins@arm.com> wrote in message news:<408E5600.E936EE7D@arm.com>...
> Jules Gilbert wrote:
> 
> > My compressors don't accept ANY input, just previously compressed
> > input.  They are also a lot slower than conventional compressors.  Oh
> > yes, one more thing; they work again and again.  Over and over again.
> > Don't you get it?
> 
> I take it that during development you've tested this with any number of files.
> 
> Take three of these files and keep compressing them over and over again.
> Each of these three files are now only one bit long.
> (At least) two of these files are the same.
> 
> How do you get from this file to both the initial inputs?
> Starting with just this file how do you know which one of the two starting files it was meant to be?
> 
> The same argument applies whenever you're arguing in favour of compressing every input file as you cannot
> have a 1-1 mapping between input and output files without accepting that you cannot compress all files.
> Previously compressed input is actually worse as you're saying you can only compress some files.
> 
> If you can't get every file down to 1 bit and only to (say) 100 bits then simply take more than 2^100 files
> and you're shoving pigeons in more than one hole again...
> 
> This applies globally, there is NO way around this.
> 
> Regards,
> Mike...

> And no, I can not represent every possible input.  Some inputs would
> fail (I XOR the input -- all the core process sees is RAD.)

I assume this works recursively? in each iteration the input can
be XORed?

and when you *decode* this, you invert the xor-operation ... ?
but how do you know, which of the outputs have to be XORed?


I think you've forgotten to count the side-information that the 
decoder would need to work properly ..

- Number of iterations (how does the decoder know how often 
to decompress the buffer?) ... this could also be done by 
remembering the original-file-size ...

- XOR-tracking ... for each iteration you need to know if
the output has to be XORed ...


bye,
Michael

If your final compressed files are around 300 bytes, this means that your
decompression algorithm is wrong.  300 bytes means 2 ^ (8*300) = 2 ^ 2400
possible files that can result in a decompression, not a single one more.
And 2 ^ 2400 is a extremely small number of files, compared with, say, 2 ^
800000 which represent somewhere around the maximum number of files of size
1 MB. So in reality there are much more files that will never be succesfully
compressed/decompressed by your algorithm. Correct?



"Jules Gilbert" <julesg@financier.com> wrote in message
news:9c7d0936.0404271338.5387864b@posting.google.com...
> The process of compression does not alway make files one bit in length
> (as in never!)
>
> The process *sometimes* makes files (ie., buffers, all files are
> broken up into buffers) only a little smaller than the input buffer.
>
> I use a small buffer, of 128 bytes.  Why?  Because a 'file' consists
> of a serial list of buffers, and, as compression occurs. the length of
> of this file gets shorter.  The compression process produces RAD
> output (RAD == 'random appearing data') for each buffer WHICH IS
> SMALLER, maybe 10% smaller, maybe 40% smaller.
>
> Why 128 bytes?  Because this is the dominant controlling parameter on
> the smallest size file we can emit.  With 128 byte buffer's, I can
> produce final files as small as 150-300 bytes.
>
> And no, I can not represent every possible input.  Some inputs would
> fail (I XOR the input -- all the core process sees is RAD.)
>
> I've have to go over this again and again, as new grad students come
> through.
>
> --jg
>

Paul, 

Of all the comments and discussion posted here in years, I find your
analysis the most cogent, simply the finest.

Because of the manner in which you frame your argument.  Look, email
me (use my public email address, julesg@financier.com to give me a
secure and usable address.  I have some things for you.

--jg

PS:  At the very least I maintain a list of people to email should I
decide to make my invention public and I want a best-address from you,
too.

In the meantime, don't make too much of the argument you raise.  Yes,
bit transfer is a pretty fundamental part of communications and file
reading/writing, but this is just design engineering wrt
communications, and in engineering one sometimes gets around problems
by making use of a clever design.

And for the record, I do consult with a few MIT trained CS and EE
folks holding advanced degrees from MIT.  But the basics have been
solid for years, I want to insure that people know that I had this on
such-and-such a date.

--jg


"Paul Sc." <paul_sc47@hotmail.com> wrote in message news:<c6nnd2$bk4$1@news.wplus.net>...
> If your final compressed files are around 300 bytes, this means that your
> decompression algorithm is wrong.  300 bytes means 2 ^ (8*300) = 2 ^ 2400
> possible files that can result in a decompression, not a single one more.
> And 2 ^ 2400 is a extremely small number of files, compared with, say, 2 ^
> 800000 which represent somewhere around the maximum number of files of size
> 1 MB. So in reality there are much more files that will never be succesfully
> compressed/decompressed by your algorithm. Correct?
> 
> 
> 
> "Jules Gilbert" <julesg@financier.com> wrote in message
> news:9c7d0936.0404271338.5387864b@posting.google.com...
> > The process of compression does not alway make files one bit in length
> > (as in never!)
> >
> > The process *sometimes* makes files (ie., buffers, all files are
> > broken up into buffers) only a little smaller than the input buffer.
> >
> > I use a small buffer, of 128 bytes.  Why?  Because a 'file' consists
> > of a serial list of buffers, and, as compression occurs. the length of
> > of this file gets shorter.  The compression process produces RAD
> > output (RAD == 'random appearing data') for each buffer WHICH IS
> > SMALLER, maybe 10% smaller, maybe 40% smaller.
> >
> > Why 128 bytes?  Because this is the dominant controlling parameter on
> > the smallest size file we can emit.  With 128 byte buffer's, I can
> > produce final files as small as 150-300 bytes.
> >
> > And no, I can not represent every possible input.  Some inputs would
> > fail (I XOR the input -- all the core process sees is RAD.)
> >
> > I've have to go over this again and again, as new grad students come
> > through.
> >
> > --jg
> >

>>> At the very least I maintain a list of people to email should I decide
to make my invention public and I want a best-address from you, too.

Unfortunately I am not interested in any non-public
theories/algorithms/applications. But if you will publicly post an article
somewhere on the web or in a scientific paper then I might take a look at
it.

>>> Yes, bit transfer is a pretty fundamental part of communications and
file reading/writing, but this is just design engineering wrt
communications, and in engineering one sometimes gets around problems by
making use of a clever design.

Well, you have to use bits for communication, so here we are today. We are
living in a digital world simply because digital signals are way easier to
deal with compared with analog counterparts. For example, it is very easy to
straighten up a digital signals by fixing distorsions in a sequences of
approximative "1"s and "0"s on the wire. Compare this with an analog signal
that has all sorts of external interferences, different speeds for various
frequency components, different autenuation depending on the frequency, etc.
In the end, it is much easier to implement algorithms in digital computers
than in anything else. So if your idea depends on some sort of an
non-digital computer/protocol then you will have a hard time finding a use
for it...

Thanks, Paul


"Jules Gilbert" <julesg@financier.com> wrote in message
news:9c7d0936.0404290406.2f59770f@posting.google.com...
> Paul,
>
> Of all the comments and discussion posted here in years, I find your
> analysis the most cogent, simply the finest.
>
> Because of the manner in which you frame your argument.  Look, email
> me (use my public email address, julesg@financier.com to give me a
> secure and usable address.  I have some things for you.
>
> --jg
>
> PS:  At the very least I maintain a list of people to email should I
> decide to make my invention public and I want a best-address from you,
> too.
>
> In the meantime, don't make too much of the argument you raise.  Yes,
> bit transfer is a pretty fundamental part of communications and file
> reading/writing, but this is just design engineering wrt
> communications, and in engineering one sometimes gets around problems
> by making use of a clever design.
>
> And for the record, I do consult with a few MIT trained CS and EE
> folks holding advanced degrees from MIT.  But the basics have been
> solid for years, I want to insure that people know that I had this on
> such-and-such a date.
>
> --jg
>
>
> "Paul Sc." <paul_sc47@hotmail.com> wrote in message
news:<c6nnd2$bk4$1@news.wplus.net>...
> > If your final compressed files are around 300 bytes, this means that
your
> > decompression algorithm is wrong.  300 bytes means 2 ^ (8*300) = 2 ^
2400
> > possible files that can result in a decompression, not a single one
more.
> > And 2 ^ 2400 is a extremely small number of files, compared with, say, 2
^
> > 800000 which represent somewhere around the maximum number of files of
size
> > 1 MB. So in reality there are much more files that will never be
succesfully
> > compressed/decompressed by your algorithm. Correct?
> >
> >
> >
> > "Jules Gilbert" <julesg@financier.com> wrote in message
> > news:9c7d0936.0404271338.5387864b@posting.google.com...
> > > The process of compression does not alway make files one bit in length
> > > (as in never!)
> > >
> > > The process *sometimes* makes files (ie., buffers, all files are
> > > broken up into buffers) only a little smaller than the input buffer.
> > >
> > > I use a small buffer, of 128 bytes.  Why?  Because a 'file' consists
> > > of a serial list of buffers, and, as compression occurs. the length of
> > > of this file gets shorter.  The compression process produces RAD
> > > output (RAD == 'random appearing data') for each buffer WHICH IS
> > > SMALLER, maybe 10% smaller, maybe 40% smaller.
> > >
> > > Why 128 bytes?  Because this is the dominant controlling parameter on
> > > the smallest size file we can emit.  With 128 byte buffer's, I can
> > > produce final files as small as 150-300 bytes.
> > >
> > > And no, I can not represent every possible input.  Some inputs would
> > > fail (I XOR the input -- all the core process sees is RAD.)
> > >
> > > I've have to go over this again and again, as new grad students come
> > > through.
> > >
> > > --jg
> > >