RSS 2.0 Feed

» Welcome Guest Log In :: Register

Pages: (5) < [1] 2 3 4 5 >   
  Topic: Creating CSI with NS, H T T H H H T H T T H H H H T T T< Next Oldest | Next Newest >  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 06 2012,18:57   

As per this thread at TSZ:

http://theskepticalzone.com/wp....t....t-16375

I've created a version here:

http://complexspecifiedinformation.appspot.com/....pot....pot.com

Source code to follow shortly. Feature suggestions etc welcome. If it does not actually render for you then it's because it's not outputting actual HTML just yet, will fix asap.


images

Full data history of every step can be/will be made available. Initialized with random data.

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
rossum



Posts: 289
Joined: Dec. 2008

(Permalink) Posted: Oct. 07 2012,05:26   

Possibly off topic, but this is a piece I keep for when the CSI argument pops up:


CSI and Regular Processes

Dr Dembski asserts that it is not possible for regular processes to create Complex Specified Information (CSI). This is incorrect, it is perfectly possible for a regular process to create CSI in large amounts as I shall demonstrate.

The "S" in CSI stands for "Specified", so we need to have a specification for our example. I shall use, "The text of the King James Bible" as the specification for this demonstration.

Here is some text which meets the specification:

Quote
"In the beginning God created the heaven and the earth ... The grace of our Lord Jesus Christ be with you all. Amen."


For practical reasons I have elided much of the actual text.

Here is some text of the same length which does not meet the specification:

Quote
"Va gur ortvaavat Tbq perngrq gur urnira naq gur rnegu ... Gur tenpr bs bhe Ybeq Wrfhf Puevfg or jvgu lbh nyy. Nzra."


The first text contains information and also meets the specification, hence it contains specified information. The complete text, without the elision, would be long enough to be complex and so contain CSI, Complex Specified Information.

The second text contains the same amount of raw information; it is the same length and drawn from the same character set. The calculation of Shannon information will give the same value. However the second text is not specified information because it is not the King James Bible. The second text contains information but it contains zero CSI because it does not meet the specification. It only contains CUI: Complex Unspecified Information.

Now we will apply a regular process to the second text: an alphabetic barrel shift of 13 places, also known as ROT13. If we apply this regular process to the second text it changes to:

Quote
"In the beginning God created the heaven and the earth ... The grace of our Lord Jesus Christ be with you all. Amen."


This of course now contains exactly the same amount of CSI as the first text. Previously it contained zero CSI because it did not meet the specification. After applying the regular process it meets the specification, so we have increased the amount of CSI present purely by applying a regular process.

Contrary to what Dr Dembski has stated, this simple example shows that a regular process, such as ROT13, can create CSI. Since a regular process can create CSI it is therefore incorrect to assert that a regular process cannot be the origin of any CSI found in living organisms. This is a major problem for ID's attempts to use the presence of CSI as a marker of design. CSI can be generated by non-design processes such as ROT13.


rossum

--------------
The ultimate truth is that there is no ultimate truth.

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 07 2012,07:29   

It's all good. Let's collect it all here.

Made a few changes, now the app will generate a new generation every 60 seconds. Visiting the page will just display the current generation.

I'll have a graph of score/time up soon. Some issues recording each iteration currently, working on that.

Plus I'll put in a special "start with all tails" mode just for Mung as Mung is obviously incapable of programming his own version.  

Mung, if you grow some then by all means tell me on this thread what's wrong with the app I've written.

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
Wesley R. Elsberry



Posts: 4991
Joined: May 2002

(Permalink) Posted: Oct. 07 2012,08:14   

rossum wrote:

Quote

Contrary to what Dr Dembski has stated, this simple example shows that a regular process, such as ROT13, can create CSI.  [...]  CSI can be generated by non-design processes such as ROT13.


The are ways to demonstrate algorithmic processes generating CSI; Jeff Shallit and I have done that before. However, the example cited is not one of them. In fact, Dembski's "The Design Inference" devotes a chunk of text to the discovery of specifications for encrypted text.

Quote

Since a regular process can create CSI it is therefore incorrect to assert that a regular process cannot be the origin of any CSI found in living organisms.  This is a major problem for ID's attempts to use the presence of CSI as a marker of design.


This part I agree with, but it isn't justified by the ROT13 example.

--------------
"You can't teach an old dogma new tricks." - Dorothy Parker

    
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 07 2012,08:51   

For the time being here is a pastebin of the source:

http://pastebin.com/QLHPH0W....LHPH0Wf

It's an app engine project. I'll put the whole thing up as a bundle on one of those github soon.

Code is not exactly elegant, but hey, only a few hours in so far...

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
rossum



Posts: 289
Joined: Dec. 2008

(Permalink) Posted: Oct. 07 2012,11:26   

Quote (Wesley R. Elsberry @ Oct. 07 2012,08:14)
The are ways to demonstrate algorithmic processes generating CSI; Jeff Shallit and I have done that before. However, the example cited is not one of them. In fact, Dembski's "The Design Inference" devotes a chunk of text to the discovery of specifications for encrypted text.

It is always possible to find a specification for encrypted text: "The string which when decrypted with the Caesar cypher, key 13, gives the text of the King James Bible." The problem is knowing the specification in advance.

However, using that specification, the actual text of the KJV does not meet the specification, and so has zero CSI. The regular process of decryption will destroy CSI, but conversely, the regular process of encryption will create CSI.

If we are allowed to change the specification in mid-calculation then we can effectively set any value of CSI we want to zero; just switch the specification to: "A design for a working perpetual motion machine." Such a specification cannot be met. Hence it would be 'easy' to show that nothing at all had any CSI and there was no design to be found anywhere. Hardly the result that the ID side wants.

Dr Dembski's search for specifications for encrypted text, without knowing the key, is effectively a search for a universal code breaker. In cryptography, if the output of an encryption algorithm can be distinguished from random, then that encryption is considered to be broken. A mathematically perfect encryption cannot be distinguished from random, without the key. I am sure that both the NSA and GCHQ would be very interested indeed if Dr Dembski had made any progress in this area.

I agree that my piece is far from rigorous, but I think that it is at about right level for most internet discussion fora.

rossum

--------------
The ultimate truth is that there is no ultimate truth.

  
midwifetoad



Posts: 4003
Joined: Mar. 2008

(Permalink) Posted: Oct. 07 2012,15:50   

I'm afraid I don't understand the output of this program. My understanding is the fitness score for a 500 bit child is the product of the lengths of runs of ones.

The drop dead halting score is 1060.

I believe Lizzy's program ran for days at a time, doing many generations per second.

But my memory could be faulty.

Edit. After reading this I believe there's a typo in the TSZ thread, and the halting number should be 10^60. Sometimes my brain doesn't engage until after I've posted something.

Edited by midwifetoad on Oct. 07 2012,16:02

--------------
Any version of ID consistent with all the evidence is indistinguishable from evolution.

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 07 2012,15:59   

We has chart. http://complexspecifiedinformation.appspot.com/chart......t....rt



Quote
My understanding is the fitness score for a 500 bit child is the product of the lengths of runs of ones.


I have not been following the discussion closely. It's an attempt at the pseudo described in the OP in the link above to TSZ

Quote
My understanding is the fitness score for a 500 bit child is the product of the lengths of runs of ones.


This is the fitness code:
Code Sample
for toss in r:
if toss == 'H':
# As they do so, they record all runs of heads
counter+=1
if toss == 'T':
# so that if they toss H T T H H H T H T T H H H H T T T,
if counter > 0:
# they will record: 1, 3, 1, 4, representing the number of heads in each run.
score.append(counter)
counter = 0
round_score.append(score)


So given a string of heads and tails like so:

H,H,T,H,H,H,T,H,H

The score calculated would be

2,3,2

And then the product of that list would be the score, 12 in this example.

If there is now an alternative model being used, the bitstring length of 1's you mention, could you link me to the original description?

Or have I totally misunderstood this: ? Always possible!

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 07 2012,16:01   

Quote (midwifetoad @ Oct. 07 2012,15:50)
I believe Lizzy's program ran for days at a time, doing many generations per second.

Yeah, this is *much* slower but I can make the results available at every iteration and provide charts and that if anyone likes.

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 07 2012,16:08   

I can make it run much faster however. Much much faster. I'll probably attempt a mapreduce version or something, just to learn how that all works.

EDIT: I've sped it up some.

But what should the point mutation function be?

Currently (I've just changed it from the 1 in 10 per coin) it's set to one flip per row (set of 500) coin tosses. Probably too noisy 1 in 10 I thought.

Edited by oldmanintheskydidntdoit on Oct. 07 2012,17:06

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 08 2012,04:54   


images

A familiar pattern emerges. ~33135 generations.

What I find amazing is that such a simple construct can get so close to the target despite hardly any, well, anything really.

I implemented the TSP a while ago also (still under wraps for now) and again this procedure works amazingly well.

1) Generate a random population.
2) Sort by fitness (route length for TSP, or score here)
3) Delete the bottom % of the population.
4) Recreate the same % from the rest of the population
5) Throw in the odd mutation
6) Rinse and repeat.

And from the mess comes short TSP routes. From the mess comes high scoring patterns.

I can see why the ID crowd has such a problem with this sort of model. I can't imagine a realistic analogue to a biological population, say of bacteria, where the least fit make room for the fitter and information is exchanged to create the next generation with the odd mutation.

Just not going to happen is it?

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
Quack



Posts: 1961
Joined: May 2007

(Permalink) Posted: Oct. 08 2012,07:34   

Quote
Or have I totally misunderstood this: ? Always possible!

How many times > 0 have we heard something like that from a creationist?

--------------
Rocks have no biology.
              Robert Byers.

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 08 2012,08:35   

Quote (Quack @ Oct. 08 2012,07:34)
 
Quote
Or have I totally misunderstood this: ? Always possible!

How many times > 0 have we heard something like that from a creationist?

I'm currently in the process of finding out what I don't know about programming by attempting to answer other peoples questions on stackoverflow.

Turns out there are significant gaps in my (self)education, even on topics I really thought I had down pat.

EDIT: Getting corrected by the author of the programming language/system you've answered a question about (and been wrong on) is a novel experience.

Edited by oldmanintheskydidntdoit on Oct. 08 2012,08:36

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
midwifetoad



Posts: 4003
Joined: Mar. 2008

(Permalink) Posted: Oct. 08 2012,08:43   

Last night I did some calculator fiddling and discovered that the highest scoring pattern is runs of four. There may be a specific highest scoring pattern that includes a mixture, but I suspect the optimum is mostly fours with one five.

Math isn't my strongest subject, so it's likely I have screwed this up. But it appears there is a very small set of qualifying strings. Not quite as small as Mung's set of one string. But close.

The fitness function is similar to that used in the travelling salesman problem, and the GA has the ability to sum many variables into one score that has no knowledge of the target.

If it is to be considered a model of reality, then reality must allow fitness gradients. What the model demonstrates is that many dimensions of fitness can be addressed simultaneously.

--------------
Any version of ID consistent with all the evidence is indistinguishable from evolution.

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 08 2012,08:52   

Quote (midwifetoad @ Oct. 08 2012,08:43)
Last night I did some calculator fiddling and discovered that the highest scoring pattern is runs of four.

IIRC that's was the result of Lizzie's runs.

Currently mine won't get to that, bug in the distribution of mutations function, will fix and restart. Got it going at about 1 generation a ~second now but should be able to improve on that.

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
midwifetoad



Posts: 4003
Joined: Mar. 2008

(Permalink) Posted: Oct. 08 2012,09:47   

To my bemusement, the "winning" patterns are very simple and regular. As if you could have an oracle that sets "weasel" as the target without knowing the target string.

What remains of the ID argument is the unevidenced assertion that that are no fitness gradients or plateaus. Gpuccio in particular, treats fitness as entirely one dimensional, and selection always striving for a particular optimum sequence.

He seems unable conceptually to grasp the idea that natural selection sees all aspects of fitness simultaneously.

--------------
Any version of ID consistent with all the evidence is indistinguishable from evolution.

  
k.e..



Posts: 5432
Joined: May 2007

(Permalink) Posted: Oct. 08 2012,10:04   

Quote (midwifetoad @ Oct. 08 2012,17:47)
To my bemusement, the "winning" patterns are very simple and regular. As if you could have an oracle that sets "weasel" as the target without knowing the target string.

What remains of the ID argument is the unevidenced assertion that that are no fitness gradients or plateaus. Gpuccio in particular, treats fitness as entirely one dimensional, and selection always striving for a particular optimum sequence.

He seems unable conceptually to grasp the idea that natural selection sees all aspects of fitness simultaneously.

Indeed.

What those tards don't get is that a GA will generate non conceptual solutions which reach a target (or if you will, create new species) more effectively than most other searches AND can produce entirely different species accross multiple fitness landscapes.

Simply because designers are not as effective as (a random .....snikker)GA.

--------------
"I get a strong breeze from my monitor every time k.e. puts on his clown DaveTard suit" dogdidit
"ID is deader than Lenny Flanks granmaws dildo batteries" Erasmus
"I'm busy studying scientist level science papers" Galloping Gary Gaulin

  
OgreMkV



Posts: 3668
Joined: Oct. 2009

(Permalink) Posted: Oct. 08 2012,10:06   

Quote (midwifetoad @ Oct. 08 2012,09:47)
To my bemusement, the "winning" patterns are very simple and regular. As if you could have an oracle that sets "weasel" as the target without knowing the target string.

What remains of the ID argument is the unevidenced assertion that that are no fitness gradients or plateaus. Gpuccio in particular, treats fitness as entirely one dimensional, and selection always striving for a particular optimum sequence.

He seems unable conceptually to grasp the idea that natural selection sees all aspects of fitness simultaneously.

As I recall, there was a machine learning system that learned how to play checkers.  The programmers didn't tell it any rules or even what a winning condition was.  

All they did was, after 5 runs, gave the machine a win/loss ratio.  That's it.

The computer learned the rules of checkers and then learned optimal play.  When the microsoft gaming servers were online, they played the machine against live opponents there and found that it was nearly master level... something like an ELO score around 1480 (with 1500 being master).

So it's possible.

--------------
Ignored by those who can't provide evidence for their claims.

http://skepticink.com/smilodo....retreat

   
midwifetoad



Posts: 4003
Joined: Mar. 2008

(Permalink) Posted: Oct. 08 2012,10:52   

I would pay money to see that program play against Gil's I realize it is possible using a database to play a perfect end game, but I suspect the opening and middle would be interesting.

Is there any way to make that happen? I think Gil's can be purchased.

Edited by midwifetoad on Oct. 08 2012,10:53

--------------
Any version of ID consistent with all the evidence is indistinguishable from evolution.

  
OgreMkV



Posts: 3668
Joined: Oct. 2009

(Permalink) Posted: Oct. 08 2012,11:29   

Quote (midwifetoad @ Oct. 08 2012,10:52)
I would pay money to see that program play against Gil's I realize it is possible using a database to play a perfect end game, but I suspect the opening and middle would be interesting.

Is there any way to make that happen? I think Gil's can be purchased.

Gil's uses a database for the end game?  Probably, most of them do.

The machine learning one doesn't.  It very well may have developed it's own database internally, but a endgame database wasn't programmed into it.

Maybe just get on a chat with Gil and run both programs.  That's how the researchers did it.  They acted as an intermediary between the checkers learning program and the on-line gaming system.

--------------
Ignored by those who can't provide evidence for their claims.

http://skepticink.com/smilodo....retreat

   
OgreMkV



Posts: 3668
Joined: Oct. 2009

(Permalink) Posted: Oct. 08 2012,11:32   

Here's the link to the papers (and I had the ELO numbers all screwed up). http://www.natural-selection.com/NSIPubl....ine.htm

The papers are in 2000 and 2001.

Crap... those links are bad.  Here are some good links... http://red.cs.nott.ac.uk/~gxk....001.pdf

http://www.cs.nott.ac.uk/~gxk....ext.pdf

Edited by OgreMkV on Oct. 08 2012,11:35

--------------
Ignored by those who can't provide evidence for their claims.

http://skepticink.com/smilodo....retreat

   
Soapy Sam



Posts: 659
Joined: Jan. 2012

(Permalink) Posted: Oct. 08 2012,11:53   

Quote (midwifetoad @ Oct. 08 2012,09:47)
To my bemusement, the "winning" patterns are very simple and regular. As if you could have an oracle that sets "weasel" as the target without knowing the target string.

Yes, the fitnesss function 'smuggles in' something about products! 4x4 is bigger than 3 x 5, 2 x 6, 1 x 7 ... which is, for example, why equal masses give a higher gravitational force (proportional to m1 x m2) than any asymmetric distribution of the same total mass (actually, the cause is the unequal distribution of individual atomic interactions, but proportionality to m1 x m2 is the mathematical result).

So settling on same-sized runs with 1-bit separators maximises the product.

Edited by Soapy Sam on Oct. 08 2012,11:55

--------------
SoapySam is a pathetic asswiper. Joe G

BTW, when you make little jabs like “I thought basic logic was one thing UDers could handle,” you come off looking especially silly when you turn out to be wrong. - Barry Arrington

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 08 2012,12:41   

just playing around with a duplication_error function that takes a random length segment and repeats it, starting where the original segment ends.

It's at generation 290 and it's already beaten the single point mutation version which is on almost 50k iterations.

I turned charting off for speed, will get it back on and compare...

EDIT: Adding duplications makes it complete very quickly. Top15 shown here, not all 500 shown (rows truncated).  

free image hosting
Will have to make a roll your own parameter version of this.

Edited by oldmanintheskydidntdoit on Oct. 08 2012,17:02

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
oldmanintheskydidntdoit



Posts: 4999
Joined: July 2006

(Permalink) Posted: Oct. 13 2012,05:05   

re: compressibility, the hot topic of the moment.

Adding in charts for genome size after gzip compression is applied. More to follow.

photo storage

Edited by oldmanintheskydidntdoit on Oct. 13 2012,05:05

--------------
I also mentioned that He'd have to give me a thorough explanation as to *why* I must "eat human babies".
FTK

if there are even critical flaws in Gaugers work, the evo mat narrative cannot stand
Gordon Mullings

  
Jerry Don Bauer



Posts: 135
Joined: Nov. 2012

(Permalink) Posted: Nov. 19 2012,11:34   

Well, You guys may be missing the point just a bit.

Complex Specified Information can alway create further Complex Specified Information if by nothing more than rearrangement.

People, computers, software, etc. do this every day. RNA/DNA replicates quite readily.

The kicker is that CSI cannot arise naturally without Intelligent Design somewhere in the process.

Unfortunately for the ID detractors, they have never given an example where this has happened and it is, in fact, mathematically impossible.

That, indeed, is the point... :p

  
Henry J



Posts: 5787
Joined: Mar. 2005

(Permalink) Posted: Nov. 19 2012,11:49   

But the only way something can be mathematically impossible is for that to be proven as a theorem in a formal mathematical system (i.e., based on axioms).

But before doing that, one has to show that the physical phenomena is actually equivalent to the formal mathematical system.

But that would require first actually defining CSI in a meaningful manner, rather than as a political slogan used only for propaganda.

Henry

  
Jim_Wynne



Posts: 1208
Joined: June 2006

(Permalink) Posted: Nov. 19 2012,12:11   

Quote (Jerry Don Bauer @ Nov. 19 2012,11:34)
The kicker is that CSI cannot arise naturally without Intelligent Design somewhere in the process.

Unfortunately for the ID detractors, they have never given an example where this has happened and it is, in fact, mathematically impossible.

Joe Bob,

Did you know that the so-called argument from probability, aka argument from large numbers, pegs those who use it as profoundly ignorant of probability theory, and that it is, ultimately, an argument that assumes its own conclusion?

--------------
Evolution is not about laws but about randomness on happanchance.--Robert Byers, at PT

  
JohnW



Posts: 3217
Joined: Aug. 2006

(Permalink) Posted: Nov. 19 2012,12:46   

Quote (Henry J @ Nov. 19 2012,09:49)
But the only way something can be mathematically impossible is for that to be proven as a theorem in a formal mathematical system (i.e., based on axioms).

But before doing that, one has to show that the physical phenomena is actually equivalent to the formal mathematical system.

But that would require first actually defining CSI in a meaningful manner, rather than as a political slogan used only for propaganda.

Henry

CSI is defined as "that thing which cannot arise naturally without Intelligent Design somewhere in the process".  So: if it looks like GodDesignerdidit to me, then it has CSI.  And since it has CSI, then GodDesignerdidit.  All science so far!

--------------
Math is just a language of reality. Its a waste of time to know it. - Robert Byers

There isn't any probability that the letter d is in the word "mathematics"...  The correct answer would be "not even 0" - JoeG

  
Jerry Don Bauer



Posts: 135
Joined: Nov. 2012

(Permalink) Posted: Nov. 19 2012,12:55   

Quote (Henry J @ Nov. 19 2012,11:49)
But the only way something can be mathematically impossible is for that to be proven as a theorem in a formal mathematical system (i.e., based on axioms).

But before doing that, one has to show that the physical phenomena is actually equivalent to the formal mathematical system.

But that would require first actually defining CSI in a meaningful manner, rather than as a political slogan used only for propaganda.

Henry

Borel's law, which was a well accepted tenet of chemistry long before whoever you believe the "evil, science twisting; propagandists" that you think exist, came along, would beg to disagree with you concerning the math.

And why do you have trouble defining CSI? It is a well defined concept of modern ID thought.... is it information that calculates out above the upper probability bound? Is it specified information? Then, if it is both complex and specified it is therefore CSI.....

This too is just a take off of Borel's law that all of us who major in science tend to run across at some point as an undergrad.

It's actually not that complicated to comprehend.

  
Henry J



Posts: 5787
Joined: Mar. 2005

(Permalink) Posted: Nov. 19 2012,13:10   

But I am not one of those claiming that CSI has been defined, so I'm not the one who needs to supply a definition.

If it had a definition, you could have supplied it just now, but you didn't.

Keep in mind that it wouldn't work to define CSI simply as something that can't evolve, because that would produce a circular argument.

  
  128 replies since Oct. 06 2012,18:57 < Next Oldest | Next Newest >  

Pages: (5) < [1] 2 3 4 5 >   


Track this topic Email this topic Print this topic

[ Read the Board Rules ] | [Useful Links] | [Evolving Designs]