How to do ID: (1) Find a shark. (2) Jump it.

Posted 26 August 2009 by Richard B. Hoppe

Denyse O'Leary, she of the multiple blogs and little reliable knowledge of evolution, is offering a prize for the original code for Dawkins' cumulative selection demonstration program ('METHINKS ..."), described in The Blind Watchmaker, originally published in 1986. The winner actually gets to choose between two prizes, a copy of Stephen Meyer's new elaboration of the standard ID argument from ignorance, Signature in the Cell, or a copy of Dawkins' forthcoming The Greatest Show on Earth. (Actually, for the latter prize, O'Leary says she will ask Dawkins' publicist to provide the prize. Strange to offer a prize she can't herself deliver.) The comment thread is strangely reminiscent of the recent "birther" rhetoric in the U.S. A commenter called "kibitzer" replicates the birther script almost flawlessly. For example

It is simply unconscionable that over 20 years after the program has been out and used to argue for Darwinism, Dawkins still has not made this code publicly available.

and

But the program has been much discussed on the Internet in the last decade. So where is the code?

and

Then provide the original code. Repeat after me: WE WANT TO SEE THE ORIGINAL CODE, WE WANT TO SEE THE ORIGINAL CODE, WE WANT TO SEE THE ORIGINAL CODE ...

and

Of course, as programs go, Dawkins' WEASEL is trivial and it's easy enough to reconstruct something that's close to it. But given the controversy surrounding it, let's see the original program. Why is that so difficult?

and

We're all beating our gums. Please, let's see the original code. Why is that so much to ask? To paraphrase Ben Stein, Does anyone have it? Anyone?

Controversy? Only in the fevered imagination of Bill Dembski, who has now infected Robert Marks. Hat tip to Glenn Branch.

100 Comments

Eamon Knight · 26 August 2009

1986? If the source still exists, it's probably sitting on some floppy or tape that can't be read by any computer currently in production (not that there aren't geeks who save the old technology just for fun, and could read just about anything if offered enough beer for their trouble).

But about your headline: since when was D'Oh!Leary ever on the right side of the shark to begin with?

Mike Haubrich · 26 August 2009

May I ask a polite question? WHO CARES!

It was just a demonstration of a concept, but for some reason they think that if they can find a flaw in the program eevolluuion wil fall apart.

Do you have to have been dropped on your head to be a creationist, or is it just very helpul?

Mike Haubrich · 26 August 2009

Eamon Knight said: 1986? If the source still exists, it's probably sitting on some floppy or tape that can't be read by any computer currently in production (not that there aren't geeks who save the old technology just for fun, and could read just about anything if offered enough beer for their trouble). But about your headline: since when was D'Oh!Leary ever on the right side of the shark to begin with?

Or, they could just pull ou their twenty year old programming guides and re-create it to see for themselves how it runs.

DavidK · 26 August 2009

They have every intention of adulterating the code to suit their own ends, i.e., make genetic changes to it, evolve it, then claim it was theirs to begin with. I suspect they would end up with something like "creintelligent designationism."

RBH · 26 August 2009

I should have noted that Dawkins' birth certificate shows that he was born in Kenya. That settles it. Right?

eric · 26 August 2009

Lots of the UD comments appear to get it - i.e. its the algorithm, not the code, thats important; the algorithm is available; you can produce a code that makes the same point any time you want; etc.

Of course I fully expect that in a day or two all those comments will be gone....victims of intelligent design...

ObSciGuy · 26 August 2009

And how's he going to verify the code?? Anyway...

So why not hold a proper contest and do some science? The challenge would be to code up at least the two different versions of the thing and try and compare and contrast them w/ one another and with what (if anything) is in the book (I haven't read it). Top prize goes to the best comparison and (snarky) writeup??

The only rules I can think of are (1) to use a freely available language so that any and all can replicate your analysis, and (2) address the original quibbling honestly, and discuss it's (ir)relevance to the theory of evolution by natural selection.

Just a thought ;)

Mike Elzinga · 26 August 2009

Special invitation for Denyse O’Leary – but any civil ID/creationist is entitled to enter.

Write an essay demonstrating that you have a complete understanding of evolution and the role of natural selection without having to mischaracterize any of it

Stanton · 26 August 2009

Mike Elzinga said: Special invitation for Denyse O’Leary – but any civil ID/creationist is entitled to enter. Write an essay demonstrating that you have a complete understanding of evolution and the role of natural selection without having to mischaracterize any of it

Do you realize that you've just doubly disqualified all of the members of the Discovery Institute, as well as all of the Intelligent Design proponents on the Internet? I mean, I've never encountered a civil or even polite Intelligent Design proponent before, and malicious mischaracterization of evolution (and the rest of science) is the life-manna of Intelligent Design and Creationism.

Rolf · 26 August 2009

Mike Elzinga said: Special invitation for Denyse O’Leary – but any civil ID/creationist is entitled to enter. Write an essay demonstrating that you have a complete understanding of evolution and the role of natural selection without having to mischaracterize any of it

I have for some time tried buttonholing Ray Martinez to tell us in his own words, of course without having to believe any of it, his interpretation of the term "natural selection". His only response so far is something like "Concept of Natural Selection is not seen in nature."

DiEb · 26 August 2009

I could post one comment on O'Leary's thread:

Dawkins describes his algorithm in the following way: It again begins by choosing a random sequence of 28 letters, just as before: WDLTMNLT DTJBKWIRZREZLMQCO P It now 'breeds from' this random phrase. It duplicates it repeatedly, but with a certain chance of random error - 'mutation' - in the copying. The computer examines the mutant nonsense phrases, the 'progeny' of the original phrase, and chooses the one which, however slightly, most resembles the target phrase, METHINKS IT IS LIKE A WEASEL. [...] the procedure is repeated. This is enough to replicate his program: 1. chose random string 2. copy string n times with mutation. NOTE: at this step you don't know which letters are correct in place, so no letter is safe from being mutated! 3. chose best fitting string. NOTE: best fitting seems to be generally understood to be the string with the most correct letters, the fitting is expressed by a number between 1 and 28 4. Stop, if the number of correct letters is 28, otherwise 5. goto 2 The parameters which you can chose is the number of copies n, and the probability that a letter in a string is mutated, p. You may even chose another procedure of mutation - but keep in mind the NOTE of step 2. It's really basic to realize steps 1 - 5 in the programming language of your choice...

But now I seem to be blocked, so my follow-up didn't come through:

Just to elaborate: The interesting part is what Atom describes as oracle, i.e., the application of the fitness function. In Dawkins description we read
The computer examines the mutant nonsense phrases, the 'progeny' of the original phrase, and chooses the one which, however slightly, most resembles the target phrase
An oracle - a black box - which accomplishes this just hints to one string when presented with a population of strings. No further information has to be exchanged. It is not necessary to know how the oracle defines the best string: It could just give hint you to the one with the greatest number of correct letters - or perhaps the one which shares the longest substring with the target phrase.

Dembski - who emphasizes the importance of information - strangely (?) doesn't see that his - and Dawkins's - algorithm get different amounts of information from the oracle...

harold · 26 August 2009

Mike Elzinga -

Write an essay demonstrating that you have a complete understanding of evolution and the role of natural selection without having to mischaracterize any of it

During the years between 1999 and Dover, when open creationists still showed their faces on science blogs and used AIG style arguments, and when some borderline intelligent lay people had vague ideas that "intelligent design" was something that might be worthwhile (*although then as now, only because they confuse it to imply theistic evolution, as the duplicitous types who made up the name "ID" intended), this was always the first thing I said to creationists. None of them ever managed to do it. Most just ran away, and a few of the bloated, pompous, elderly lawyer "big word creationist" types would bloviate on and on without answering. However, I now prefer the Stevaroni approach - "Pretend that you never heard of evolution; how do you explain the pattern of life's diversity? Never mind what's wrong with the theory of evolution, what is your explanation for what we see?". At least as an initial question. I was going to comment on how asinine this Denise O'Leary thing is - the Dawkins program has nothing to do with the real evidence for evolution (I'm not sure if he ever even published it in an actual journal), and it's obvious that a program could be written to do exactly the same thing, using either contemporary or 1986 technology.

harold · 26 August 2009

DiEB -

It's so ridiculous it's unreal. You offered a rational solution to the "dilemma", so your subsequent comments were blocked.

SteveF · 26 August 2009

Speaking of ID, here's a new paper in today's PNAS with a nice ID baiting title:

Clements, A. et al. (2009) The reducible complexity of a mitochondrial molecular machine. PNAS, advance online.

http://www.pnas.org/content/early/2009/08/25/0908264106.abstract

SteveF · 26 August 2009

and the authors appear to be taking IDists on pretty directly:

Molecular machines drive essential biological processes,from protein synthesis transport to genome maintenance, expression and inheritance (1, 2). Good examples include bacterial flagella (3, 4), the RNA polymerase holo-complex (5), and various protein transport machines that selectively transfer protein molecules across biological membranes (6). Proponents of Intelligent Design have argued that these sophisticated machines are ‘‘irreducibly complex,’’ with this standing as the proof that, at the molecular level, Darwin’s principles of evolution cannot explain the complexity of living systems (7, 8). Our current investigation of the function and evolution of the protein transport machines in mitochondria provides an excellent, and perhaps unique, system to provide evidence that a sophisticated molecular machine can evolve from simpler components, in a process strictly adhering to Darwinian principles of evolution.

Mike Elzinga · 26 August 2009

DiEb said: I could post one comment on O'Leary's thread:

Your post was embarrassing to them for a number of reasons: 1) They are incapable of generating code from a detailed verbal layout of a workable algorithm. 2) The educations of most ID/creationists seldom, if ever, involve having to synthesize knowledge to accomplish something unique. 3) They are almost always are asked to regurgitate specific information on multiple-choice tests with double and triple negatives thrown in to make the test “difficult”. 4) They would have no way of knowing if they got something that was correct. What would they use for comparison? 5) They have presented no criteria by which they would judge the code to do what it was designed to do. 6) They would be required to have specific and correct understandings of scientific concepts and terms in order to understand an algorithm that mimics the natural world. There are many more such reasons, but I think this gives the general idea of the problem.

Doc Bill · 26 August 2009

Show me the "X" is a lot like the creationist argument based on "were you there?"

I wrote a program in 1982 that was used successfully and re-written and modified up to around 1996.

Today, the computers it ran on, all of the backup tapes, source code documentation and even the building that housed it are gone.

Who gives a weasel's ass?

I could recreate the program to run on the iPhone, but 20 years from now with the xFon Implant we'd be back to square one and DOL would still be jumping sharks, albeit in her walker.

stevaroni · 26 August 2009

Denyse yammers... It is simply unconscionable that over 20 years after the program has been out and used to argue for Darwinism, Dawkins still has not made this code publicly available.

Notice the ID mode of thought. The program. As if anybody who's actually capable of forensically examining the code couldn't simply write their own testbed in an hour. To the ID crowd, "Weasel" is not a concept. Instead, it has to be anthromorphosed into a single, pagan, object that everybody somehow worships. The program. It's just like all the ID tripe about calling evolution Darwinism. You have to personify (or at least objectify) the thing you attack, because a mere object is assailable, whereas attacking an easily demonstrated concept is a real problem. It's especially if you're scientifically illiterate. The ironic thing, readily apparent to anybody who actually cares, is that "Weasel" is such a strong demonstration precisely because there are so many versions - the vast majority developed in true "black box" environments, where the programmer knew something about the algorithm in general terms, but hadn't actually seen the underlying code at all. He didn't need to see the code, it's so damned easy to write these things. Consequently, there must be thousands of "fitness functions" out there. And yet all these disparate versions of "Weasel", all these variants that have nothing in common, still work. Just like all the experiments that measure gravity in all the "physics 101" labs all over the world still measure the same value. That should tell you something.

RBH · 26 August 2009

stevaroni said:
Denyse yammers... It is simply unconscionable that over 20 years after the program has been out and used to argue for Darwinism, Dawkins still has not made this code publicly available.
Notice the ID mode of thought. The program. As if anybody who's actually capable of forensically examining the code couldn't simply write their own testbed in an hour.

Well, what you gotta unnerstand is that scripture is inerrant only in its original 'autographs.' If you don't have that original program, no other program embodying the same algorithm can demonstrate the same point -- they're not the original autographs.

GuyeFaux · 26 August 2009

Once again, IDists miss another important point about science, which is that experiments which can be replicated are better than those which can't. The nice thing about Dawkins's experiment is that anybody posessing a modicum of coding skills and some curiosity can replicate the purported results. So what if the original experiment is lost, it's real strength is the number of times its results were replicated.

GuyeFaux · 26 August 2009

And honestly, who cares? Even if Dawkins is shown to have made the whole thing up in '86 after having too much coffee, which seems to be the O'Leary's insinuation, why does O'Leary, or anyone, care?

Mike Elzinga · 26 August 2009

GuyeFaux said: Once again, IDists miss another important point about science, which is that experiments which can be replicated are better than those which can't.

This is similar to those AiG idiots in Ken Ham’s cult who argue that they are looking at the same data from a different perspective. First of all, we always observe them filtering out most evidence and data, and then playing gestalt games with the little that remains. But simply making such a claim is blatant admission that they are not practicing science and don't know what science is. If we just look at technological spin-offs alone, scientific evidence and consequences are independent of “philosophical perspective”. Even plants and non-human animals can benefit from what has been learned from science.

Dan · 26 August 2009

stevaroni said:
Denyse yammers... It is simply unconscionable that over 20 years after the program has been out and used to argue for Darwinism, Dawkins still has not made this code publicly available.
Notice the ID mode of thought. The program. As if anybody who's actually capable of forensically examining the code couldn't simply write their own testbed in an hour.

Agreed. It reveals a lot about creationist thinking. You start with the Bible -- the perfect word of God. It gets miscopied, mistranslated, misinterpreted, so that what we have today is but a shadow of the magnificent original. The old is better than the current, information is always lost, things always get worse. Tell a creationist that we know more this month than we knew last month ... that new information is has come to light ... that in any given month 1000 pages of new information are available through Physical Review Letters alone (to pick a single small scientific journal) ... and the idea is so far outside of his ken that he doesn't even reject it -- the idea bounces off his mind before it even enters. To the creationist, it's not the ideas in the Bible that are important, it's the Bible itself, the perfect word of God. They are so trapped in this mode of thinking that they can't even recognize that the ideas in Dawkin's book are important, but the exact word choice is irrelevant. That the idea of the Weasel program is significant, but its implementation in BASIC is not. That the program itself is not even intended to simulate evolution, but to demonstrate a facet of evolution. Scientists are accused of worshiping Darwin, and of venerating Origin of Species, not because of any evidence (there is none), but because creationists cannot conceive of operating except through a venerated "font of wisdom".

Mike Elzinga · 26 August 2009

RBH said: Well, what you gotta unnerstand is that scripture is inerrant only in its original 'autographs.' If you don't have that original program, no other program embodying the same algorithm can demonstrate the same point -- they're not the original autographs.

And yet they keep making up new translations of the “original” when they don’t like the other translations. I guess that’s ok even though this process leases to constant splintering of “proper understanding”. Yet hundreds of computer platforms and computer languages implementing from scratch and by different individuals the same algorithm involving the natural world all get the same results that all can agree upon. Go figure.

DS · 26 August 2009

DOL wrote:

"It is simply unconscionable that over 20 years after the program has been out and used to argue for Darwinism, Dawkins still has not made this code publicly available."

Right. And it is simply unconscionable that after more than ten years, Dembski has yet to calculate the amount of complex specified information in a single thing, living or not. You first Sir Issac.

Wheels · 26 August 2009

DiEb said: I could post one comment on O'Leary's thread:
[snip]
But now I seem to be blocked, so my follow-up didn't come through:
[also snip]
Dembski - who emphasizes the importance of information - strangely (?) doesn't see that his - and Dawkins's - algorithm get different amounts of information from the oracle...

I read that earlier. At least you got to post something; I've tried to register twice and never got confirmation e-mails. Remember the PT thread where contributors had WEASEL-like programs coming out of the woodworks in nearly every significant language? You'd think those would be evidence enough that the basic premise is examinable. But even back then, Dembski and UD didn't seem inclined to actually follow up on anything.
Reminds me of the time Andrew Schlafly demanded the "original data" from Richard Lenski's long term E. coli evolution experiment: they wouldn't know what to do with anything even if they could have it.

Joe Felsenstein · 27 August 2009

GuyeFaux said: And honestly, who cares? Even if Dawkins is shown to have made the whole thing up in '86 after having too much coffee, which seems to be the O'Leary's insinuation, why does O'Leary, or anyone, care?

So let me summarize 1. Opponents of evolution go around saying "Darwinism" is a theory of random evolutionary change, that it assumes that adaptations just happen as a result of purely random processes. 2. Dawkins gets mad at this deliberate misrepresentation and makes his Weasel program as a simple teaching program to show that change by mutation and natural selection can achieve a cumulative, very nonrandom result. 3. It can get to the Weasel sentence in hundreds (or a few thousands) of tries instead of the 10⁴⁰ which would be needed for a purely random search such as a monkey with a typewriter. 4. Opponents say "but it has latching"! Supporters say "no it doesn't". Opponents say "let's see the program!" Supporters say "it doesn't matter!" Opponents say "but if it has latching then [dramatically horrified indrawn breath] it is doing partitioned search!" 5. The implication they try to give is that the latching is the only reason is succeeds in much less than 10⁴⁰ tries. 6. Everyone is hung up over this (non)issue. Nonissue because even if it did have latching it would not be all that much faster. Latching isn't why it does a lot better than purely random search. People like O'Leary who make a Big Deal about whether the program has latching are not explaining in detail why that matters all that much. Can they explain why it is the critical feature of the program? Can they show that latching is the only reason why the program takes less than 10⁴⁰ trials? (No, they can't). What is really funny about all this is O'Leary's offer to Dawkins that, if he will provide the program, she will reward him by (as one possible prize) asking Dawkins' publicist to send Dawkins a copy of his own book. That takes real imagination. And chutzpah.

Bob O'H · 27 August 2009

What is really funny about all this is O’Leary’s offer to Dawkins that, if he will provide the program, she will reward him by (as one possible prize) asking Dawkins’ publicist to send Dawkins a copy of his own book. That takes real imagination. And chutzpah.

Has anyone told Dawkins? I would, but I'm a nobody, so he won't even notice.

C.W · 27 August 2009

I think I'll write a Weasel in BrainFuck and submit it. Or perhaps Malbolge.

slang · 27 August 2009

Well, thanks. That lead me to LOLCODE, and I'm still snickering, getting frowns from coworkers.

Kevin B · 27 August 2009

I seem to recall having seen when Googling (back at the time of the previous WEASEL infestation) that Dawkins has actually said that he no longer has a copy of the original program.

The question is whether Ms O'Leary was aware of this when she set her "challenge."

eric · 27 August 2009

Joe Felsenstein said: People like O'Leary who make a Big Deal about whether the program has latching are not explaining in detail why that matters all that much. Can they explain why it is the critical feature of the program? Can they show that latching is the only reason why the program takes less than 10⁴⁰ trials? (No, they can't).

Joe, I think you're slightly mischaracterizing Denyse (you're being too kind). Weasel programs with both latching and non-latching options have been made available to them. For decades. This is not a situation where the ID folk fail to argue their case; its a situation where they blatantly ignore actual evidence, just flat-out pretend it doesn't exist. The experiment has been done. Depending on the other variables, latching changes the number of generations required from (approximately) 10^3-5 down to 10^2-3. Compared to the 10⁴⁰ generations required for random selection, cumulative selection of either sort provides ample evidence that (i) evolution had the time to occur and (ii) evolution provides an enormous advantage over truly random genetic change.

Ravilyn Sanders · 27 August 2009

RBH said: I should have noted that Dawkins' birth certificate shows that he was born in Kenya. That settles it. Right?

Come on. Everyone knows he does not have a birth certificate. All he has is a certificate of birth. Only in the feverish mind of evilutionists addled by Devil himself these two would be the same. We challenge you to write a program to change "birth certificate" to "certificate of birth" using the weasel algorithm. It cant be done.

snaxalotl · 27 August 2009

oh gee oh gee, if they won't show the code then weasel is a hoax: cumulative selection is maybe no faster than random guesses after all! my little creationist heart is giddy with excitement: the impending collapse of evolution without, as usual, creationists having to do any actual work

Frank J · 27 August 2009

Mike Elzinga said: Special invitation for Denyse O’Leary – but any civil ID/creationist is entitled to enter. Write an essay demonstrating that you have a complete understanding of evolution and the role of natural selection without having to mischaracterize any of it

I doubt that any one person has a "complete" understanding, but a request to demonstrate an understanding within the range of that of practicing evolutionary biologists (including beginners at the BS/MS level) would be a reasonable one. In that case my guess is that O'Leary cannot demonstrate it with or without mischaracterization, but that Behe, Meyer, Wells, and maybe Dembski, could, but would not dare do so.

arachnophilia · 27 August 2009

i've always had a few fundamental problems with "methinks".

the first being that evolution doesn't select based on a predetermined mold, comparing each successive generation to a particular "design" that an intelligent agency (in this case, dawkins) has hard-coded into the algorithm. it selects based on viability in (environmental) context. but it's much harder to write a program that does that. dawkin's program is much too directed.

the second is that it's way too selective. "methinks" only calculates the next generation based on the "fittest" sample. natural selection, on the other hand, allows all but the very weakest (or perhaps unluckiest) individuals to potentially reproduce. "survival of the fittest" was always a bit of misnomer, as i believe dawkins himself has pointed out before.

the third follows from the "directed" complaint. the outcomes of evolution don't necessarily have to make any kind of sense. they just have to work well enough to reproduce. imposing "approximation of an english phrase" just isn't a good analogy.

what "methinks" does, and what i think "methinks" was meant to do, is demonstrate the power of variation-and-selection over random chance, and how the appearance of design can come about through selection. it's been a while since i've read "the blind watchmaker" but that's the context i remember it being used in. it is not, and should not be used as an analogy for evolution by means of natural selection. or even artificial selection, really. and there are much more robust evolutionary algorithms out there that DO NOT front-load a design and test fitness, but rather are intended to output the most fit designs. those are probably much better examples...

ben · 27 August 2009

snaxalotl said: the impending collapse of evolution without, as usual, creationists having to do any actual work

And without, as usual, anything actually collapsing. Except perhaps the readership of UD.

raven · 27 August 2009

Creationism is just another lunatic fringe idea. It's been noted that polykookery is common. People who believe one impossible idea usually believe many. If you look at creationists such as O'Leary and the commenters, chances are they do believe Obama is a Kenyan, Moslem terrorist, UFOs exist and are piloted by demons (a very common fundie xian belief), satanists abduct children and kill them in rites, the Illluminati are doing something horrible any day, vaccinations are unhelathy, modern medicine doesn't work, and on and on. The latest Tinfoil Hat "theory" which I've seen multiple times yesterday.

* Did you know that criminal charges are being filed in multiple countries as well as with the FBI against the WHO / UN for conspiring to decimate the human population with a reverse engineered virus concoction to be delivered via a global mass mandatory vaccination program in the fall of 2009?

The swine flue vaccine is designed as a vehicle or deliver a fatal virus to reduce the world's population from 6.7 billion to 1/2 billion. Of course, when it doesn't happen, they will just forget this one and come up with something equally ridiculous. Maybe "the earth is 6,000 years old and geologists have been hiding that fact for a century."

Tom · 27 August 2009

I thought the proper way to do ID was

1) Find a dead horse.
2) Beat it.

raven · 27 August 2009

FWIW, it looks like Dawkin's Weasellike programs are available for free to anyone with an internet connection and a working brain. It took me 1 minute with google to find a bunch of them. That working brain requirement can be difficult, screens out most creationists.

http://health.adelaide.edu.au/Pharm/Musgrave/essays/whale.htm. WEASLE2.BAS is a program in QBASIC 1.1 that implements Dawkins's idea. It generates a series of random sentences, selects the closest sentence to a given target sentence ("methinks it is a weasel" is the default, but you can choose others), an breeds from that sentence, inducing mutations in the offspring. Currently the only selectable items are the target sentence, and the number of offspring. WEASLE5.BAS is a more sophisticated version which allows you to select initial population sizes, number of offspring and mutation rates. In this version ALL sentences are bred from, then a proportion "die". The probability of dying is proportional to the closeness of the strings to the target, and the amount of free population space left. WEASLE6.BAS takes a different approach to killing the population off, and allows you to re-run scenarios. Again, it's not a realistic simulation, but does give an idea of how population variation affects the outcome. WEASLE2.BAS, WEASLE5.BAS and WEASLE6.BAS run under the QBASIC 1.1 that comes free with DOS 5.0 and up (although it can be hard to find under Win95 and Win98, all bet are off with Win2000). They are also compatible with QuickBasic 4.x Microsoft BASIC PDS 7.x and QuickBasic 1.0 for the MAC (not 32 bit clean). If you use the /AH option in QB4.x and PDS, you can increase the maximum string population in WEASLE5 &6 (currently 250 strings). It would require some simple editing to run under TurboBasic, PowerBasic, TrueBasic or the Linux BASIC.

GuyeFaux · 27 August 2009

arachnophilia said: i've always had a few fundamental problems with "methinks". [snip]

Would you still have a problem with the Methinks experiment if its stated purpose weren't to model evolution completely, but to simply demonstrate that search with selection beats random search by orders of magnitude?

waynef · 27 August 2009

Mike Haubrich said: they think that if they can find a flaw in the program eevolluuion wil fall apart.

If that's the case will it rupture the space/time continuum and cause us all to transform into single cell organisms?

Ravilyn Sanders · 27 August 2009

raven said: It would require some simple editing to run under TurboBasic, PowerBasic, TrueBasic or the Linux BASIC.

If anyone is planning to take this code and port it more modern language, I suggest give javascript a try. The advantage is that, once implemented it will run on any browser directly for others to play with. My javascript skills are quite rudimentary. I probably could implement it in C++ very quickly but most people lack compilers to test the code themselves. Distributing the executable via the net is bad for security. My suggestion: 1. Seed string. The starting string. 2. Target string. Allow user input. 3. Latching: ON means once a letter matches the target string at the correct location, it is protected from further mutations. OFF means no letter is protected from mutations. 4. Number of progeny: Limit it to 40. 5. Mutation rate: Number of letters randomly shuffled 6. Comparison: Number of matching letters at the correct location. Display: Show target string at the top in bold face. Show the current best string below, aligned. Both in big fonts. Show generation count, latching status, current match value in the next line. Below show all the progeny for the next generation. Show matched letters in green. Show mutated letters in red. Highlight the selected string. If any matched letters are mutated blink those letters. This seems to be the big thing the creotards are alleging. Erase and repaint next generation page. I will try to implement it if I get the time.

a lurker · 27 August 2009

RBH said: I should have noted that Dawkins' birth certificate shows that he was born in Kenya. That settles it. Right?

Does that mean that Dawkins can't be Lynn Jenkins' "Great White Hope"?

eric · 27 August 2009

arachnophilia said: i've always had a few fundamental problems with "methinks"...

You could certainly set up a program with a higher population size (thousands instead of ten), lower selection pressure (e.g. +1% increase in chance to survive vs. 100%/0%), and multiple slowly-changing targets. But think about this: if the offspring under this new model took a billion times longer (than in weasel) to track reasonably close to the targets, that would mean the modeled search was still approximately 1,000,000,000,000,000,000,000,000 times faster than random search. The sheer size of the difference between Weasel and Dembski's random search should tell any reasonable person that there will be many, many, many different search algorithms that do significantly better than random search. Knowing that one search algorithm out-performs random search by 35 orders of magnitude means one of two things: Dawkins got incredibly lucky in his choice, or many weasel variants are orders of magnitude better than random search.

MartinDH · 27 August 2009

Rolf:

I haven't been on T.O. for a while (it was sucking up too much time) and your post mentioned Ray Martinez (the only *REAL* Christian left now that Dr. Scott is dead).

Has he finished his long promised, epic, evolution destroying paper yet?

--
Martin

jasonmitchell · 27 August 2009

I think an important aspect of O'Leary's 'challenge' is being overlooked - she really doesn't want the code - this is not a quest for data - it is a political statement/ publicity stunt. Like the 'birthers' demanding 'proof' that President Obama was born if Hawaii one has to ask if they really want the proof (or would accept any proof given as genuine). What are the possible outomes from O'Leary's point of view?

1) someone in Dawkins' camp provides the original code - O'Leary claims that they caved to pressure - declares 'victory'

2) someone in Dawkins' camp refuses to provide the code / explains that they don't have it/ ist lost etc. - O'leary claims Dawkins is be obstructionist/ hiding something/ whatever - O'leary gets a 'win' and continues conspiracy theory postings etc.

3) eveyone in Dawkins' camp ignores her - see #2

no matter what - she'll spin it to support her premise - facts don't matter

Mike Elzinga · 27 August 2009

arachnophilia said: i've always had a few fundamental problems with "methinks". the first being that evolution doesn't select based on a predetermined mold, comparing each successive generation to a particular "design" that an intelligent agency (in this case, dawkins) has hard-coded into the algorithm. it selects based on viability in (environmental) context. but it's much harder to write a program that does that. dawkin's program is much too directed.

It is easy to put in an outer loop that changes the target string as the inner loop converges to the current target. In fact, one could change the target at random times even before convergence occurred. You could make the target greatly different of slightly different. The inner loop will simply start converging to the new target. If the new target is not much different from the old, and if convergence to the old was close, it would not take long to converge to the new. This modification to the program would not illustrate extinction however.

RBH · 27 August 2009

GuyeFaux said:
arachnophilia said: i've always had a few fundamental problems with "methinks". [snip]
Would you still have a problem with the Methinks experiment if its stated purpose weren't to model evolution completely, but to simply demonstrate that search with selection beats random search by orders of magnitude?

As Dawkins made crashingly, painfully, explicitly clear in TBW, that's specifically what the WEASEL program was aimed at demonstrating. Nothing else. Dr. Dr. Dembski and his minions have misrepresented it for years, and Dembski at least apparently does so knowingly and deliberately. His minions, generally lacking two neurons to rub together, merely follow his lead.

Raging Bee · 27 August 2009

raven said: Creationism is just another lunatic fringe idea. It's been noted that polykookery is common. People who believe one impossible idea usually believe many. If you look at creationists such as O'Leary and the commenters, chances are they do believe Obama is a Kenyan, Moslem [and Black Panther Christian atheist] terrorist, UFOs exist and are piloted by demons (a very common fundie xian belief), satanists abduct children and kill them in rites, the Illluminati are doing something horrible any day, vaccinations are unhelathy, modern medicine doesn't work, and on and on. The latest Tinfoil Hat "theory"...

Their demonstrated "thought" processes on this one echo that of the birfers practically word-for-word: 1) Pick a totally irrelevant, not-even-tangential side-issue, and suddenly start pretending everything depends on it, even if you've never heard of it, or never considered it relevant, before. 2) Demand some document or other bit of information as "proof," whether or not it's available, whether or not it actually proves anything. 3) Pretend your own hysteria, willful ignorance, and total refusal to understand reality is entirely the fault of whoever has the "proof" you demand and won't cough it up; and swear all over the place that the entire "controversy" will vanish if only you get the "proof" you demand (even as you hedge your bets by speculating that things are so bad that the "proof" won't really clear anything up). 4) Ignore -- or censor whenever possible -- every single comment that actually addresses your demand or shows why it is unreasonable. 5) Repeat until dead.

Tomato Addict · 27 August 2009

After reading this last night I thought: What is the big deal? This is so simple I could code it in a spreadsheet.

This morning I did just that. Yes I know hundreds of others have done the same already, but I wanted to see for myself just how easy it is: METHINKS IT IS LIKE A WEASEL

Wheels · 27 August 2009

Tomato Addict said: After reading this last night I thought: What is the big deal? This is so simple I could code it in a spreadsheet. This morning I did just that. Yes I know hundreds of others have done the same already, but I wanted to see for myself just how easy it is: METHINKS IT IS LIKE A WEASEL

To answer your "What's the big deal?" question, Dembski has at times claimed that evolution does no better than a random search, so it can't possibly explain all the wacky and zany complexities of living things.
It used to be that he claimed "random mutation" does not better than a random search. After a while of people beating him over the head with the "YOU FORGOT ABOUT SELECTION!" hammer, he slightly modified his claims to say that random mutation + natural selection does little better than a random search, trying to justify his earlier conflation by saying we can pretty much ignore the role of selection.
The problem is, Dawkins wrote a simple program in Apple BASIC twenty-three years ago which demonstrates that random mutation + selection is incredibly better than random searching alone. So Dembski has been obsessing over it lately because the proof that contradicts his argument is far older than... his argument (at least since the time he's been arguing it).

Joe Felsenstein · 27 August 2009

eric said:
Joe Felsenstein said: People like O'Leary who make a Big Deal about whether the program has latching are not explaining in detail why that matters all that much. Can they explain why it is the critical feature of the program? Can they show that latching is the only reason why the program takes less than 10⁴⁰ trials? (No, they can't).
Joe, I think you're slightly mischaracterizing Denyse (you're being too kind). Weasel programs with both latching and non-latching options have been made available to them. For decades. This is not a situation where the ID folk fail to argue their case; its a situation where they blatantly ignore actual evidence, just flat-out pretend it doesn't exist. The experiment has been done. Depending on the other variables, latching changes the number of generations required from (approximately) 10^3-5 down to 10^2-3. Compared to the 10⁴⁰ generations required for random selection, cumulative selection of either sort provides ample evidence that (i) evolution had the time to occur and (ii) evolution provides an enormous advantage over truly random genetic change.

I was saying that O'Leary and Co. have not provided the other part of their argument -- I was not implying that they were being honest about that. I do think that the way that posters here responded to the D&M article exacerbated the problem. They read until they found the latching mistake, then they argued that it was a Big Deal and was the basic problem with the article. I have argued in two other threads here that it was a side issue, not worth the (virtual) ink to discuss. Now, when we ask O'Leary to explain why the issue of latching is so important, she has only to say "well you folks said it was yourself!". And as the Dawkins program seems not to be available, the ID types can act like it is a big coverup of an embarassing failure. Of course, latching doesn't matter, and (as you note) makes not much difference compared to monkey-with-a-typewriter search. So there is no need to go back to the Dawkins program. At least the issue of Obama's birth impacts whether he can be President. Whether Dawkins's program latches by contrast doesn't matter, and we have messed up an opportunity to make that clear, by treating latching as big issue.

Wheels · 27 August 2009

Joe Felsenstein said: Of course, latching doesn't matter, and (as you note) makes not much difference compared to monkey-with-a-typewriter search. So there is no need to go back to the Dawkins program. At least the issue of Obama's birth impacts whether he can be President. Whether Dawkins's program latches by contrast doesn't matter, and we have messed up an opportunity to make that clear, by treating latching as big issue.

I doubt whether we'd be able to "make that clear" even if there wasn't a hoopla about it on this end. O'Leary can't seem to even grasp the basic idea of the program, hence demands to produce the source code so they can focus on that instead of the easily replicable results from other Weasel programs. Latching isn't critical, but then again neither is O'Leary's thinking.

Joe Felsenstein · 27 August 2009

Wheels said:
Joe Felsenstein said: Whether Dawkins's program latches by contrast doesn't matter, and we have messed up an opportunity to make that clear, by treating latching as big issue.
I doubt whether we'd be able to "make that clear" even if there wasn't a hoopla about it on this end. O'Leary can't seem to even grasp the basic idea of the program, hence demands to produce the source code so they can focus on that instead of the easily replicable results from other Weasel programs. Latching isn't critical, but then again neither is O'Leary's thinking.

I think the objective should be to make it clear to undecided onlookers, not to find a way to convince O'Leary and Co.

Ian Musgrave · 27 August 2009

Joe Felsenstein said: I do think that the way that posters here responded to the D&M article exacerbated the problem. They read until they found the latching mistake, then they argued that it was a Big Deal and was the basic problem with the article.

No, the big deal was the issue of their scholarly honesty. They know that Dawkins program is not a latching program, it has been demonstrated incontrovertibly. The speed difference may be trivial to the point of being immaterial, but it is just not how Dawkins program works, and generations of geeks like me have been able to reconstruct Dawkins program just fine. See Dembski Weasles Out and Weasles on Parade for further deatils.

Ian Musgrave · 27 August 2009

raven said: FWIW, it looks like Dawkin's Weasellike programs are available for free to anyone with an internet connection and a working brain. It took me 1 minute with google to find a bunch of them. That working brain requirement can be difficult, screens out most creationists.
http://health.adelaide.edu.au/Pharm/Musgrave/essays/whale.htm. WEASLE2.BAS is a program in QBASIC 1.1 that implements Dawkins's idea. [snip]

Hey, you found my old WEASLE site! And there was a whole bunch generated here at PT see Dembski Weasles Out and Weasles on Parade for further examples in multiple languages (and our side by side comparisons of locking and Dawkins version). Took nearly a week for the UD crew to generate even one.

Tomato Addict · 27 August 2009

@Wheels: I was perhaps too subtle in my sarcasm, it is indeed freaking obvious, but thank you for the reply. We are in full agreement about the power of selection.

I have submitted my algorithm over at UD, just for fun.

Joe Felsenstein · 27 August 2009

Ian Musgrave said:
Joe Felsenstein said: I do think that the way that posters here responded to the D&M article exacerbated the problem. They read until they found the latching mistake, then they argued that it was a Big Deal and was the basic problem with the article.
No, the big deal was the issue of their scholarly honesty

I think the "they" you are referring to are the ID types at UD. I was referring by "they" to the people here at PT, who mistakenly concluded that "latching" was a Big Deal. They are not dishonest, just mistaken.

Wheels · 27 August 2009

Tomato Addict said: @Wheels: I was perhaps too subtle in my sarcasm, it is indeed freaking obvious, but thank you for the reply. We are in full agreement about the power of selection.

Ah, see, that's why html5 needs sarcasm tags more than video.

Venus Mousetrap · 27 August 2009

DiEb said: I could post one comment on O'Leary's thread:
Dawkins describes his algorithm in the following way: It again begins by choosing a random sequence of 28 letters, just as before: WDLTMNLT DTJBKWIRZREZLMQCO P It now 'breeds from' this random phrase. It duplicates it repeatedly, but with a certain chance of random error - 'mutation' - in the copying. The computer examines the mutant nonsense phrases, the 'progeny' of the original phrase, and chooses the one which, however slightly, most resembles the target phrase, METHINKS IT IS LIKE A WEASEL. [...] the procedure is repeated. This is enough to replicate his program: 1. chose random string 2. copy string n times with mutation. NOTE: at this step you don't know which letters are correct in place, so no letter is safe from being mutated! 3. chose best fitting string. NOTE: best fitting seems to be generally understood to be the string with the most correct letters, the fitting is expressed by a number between 1 and 28 4. Stop, if the number of correct letters is 28, otherwise 5. goto 2 The parameters which you can chose is the number of copies n, and the probability that a letter in a string is mutated, p. You may even chose another procedure of mutation - but keep in mind the NOTE of step 2. It's really basic to realize steps 1 - 5 in the programming language of your choice...
But now I seem to be blocked, so my follow-up didn't come through:
Just to elaborate: The interesting part is what Atom describes as oracle, i.e., the application of the fitness function. In Dawkins description we read
The computer examines the mutant nonsense phrases, the 'progeny' of the original phrase, and chooses the one which, however slightly, most resembles the target phrase
An oracle - a black box - which accomplishes this just hints to one string when presented with a population of strings. No further information has to be exchanged. It is not necessary to know how the oracle defines the best string: It could just give hint you to the one with the greatest number of correct letters - or perhaps the one which shares the longest substring with the target phrase.
Dembski - who emphasizes the importance of information - strangely (?) doesn't see that his - and Dawkins's - algorithm get different amounts of information from the oracle...

I just wrote a working weasel from this description, in Java, in fifteen minutes.

arachnophilia · 27 August 2009

GuyeFaux said: Would you still have a problem with the Methinks experiment if its stated purpose weren't to model evolution completely, but to simply demonstrate that search with selection beats random search by orders of magnitude?

no, of course not. and, while it's been a while since i've read "the blind watchmaker", that is rather distinctly the context in which i recall it being used by dawkins himself.

RBH said: As Dawkins made crashingly, painfully, explicitly clear in TBW, that's specifically what the WEASEL program was aimed at demonstrating. Nothing else. Dr. Dr. Dembski and his minions have misrepresented it for years, and Dembski at least apparently does so knowingly and deliberately. His minions, generally lacking two neurons to rub together, merely follow his lead.

yes, this is precisely what i mean. it's just not an analogy for evolution as a whole; rather a simple demonstration of how selection is vastly superior to random. the cdesign propentsists have tried to attack it as if it were evolution itself -- and even suckered a few of us along into treating it as such. and we shouldn't. they're missing the point, and we shouldn't indulge them. and really, if they knew enough programming the read the code, they probably could write it themselves. we're not talking complicated programming here.

Tomato Addict · 27 August 2009

VM writes:

... I just wrote a working weasel from this description, in Java, in fifteen minutes.

@VM: Yes. Creating screen-shots for my blog took longer than creating and testing a working algorithm, and mine has a manual step. @Wheels: [sarcasm=on]We don't need no stinkin' sarcasm tags![sarcasm=off]

DavidK · 27 August 2009

The Dishonesty Institute is already touting Dembski's paper.
Dembski's summary includes:

"Darwinian evolution is, at its heart, a search algorithm that uses a trial and error process of random mutation and unguided natural selection to find genotypes (i.e. DNA sequences) that lead to phenotypes (i.e. biomolecules and body plans) that have high fitness (i.e. foster survival and reproduction)."

Sorry, but I'm not the biologist, but this certainly doesn't sound right to me. Is he saying the goal is basically to find a "predetermined" result that yields survival, rather than a random variation of genetic material that might result in survival since it's an unguided process, i.e., the goal has already been established, and random variation cannot possibly reach that goal within the lifetime of the universe?

RBH · 27 August 2009

The theory of evolution takes as initial conditions a population of imperfect replicators with heritable variation that is fecund -- the replication rate of the population is higher than than the carrying capacity of the environment. Now, given that the population is replicating, it is from the beginning in a viable area in the fitness space -- it doesn't have to "search" for a viable volume in geno/pheno-space; it starts there.

Via various variation-generating mechanisms (e.g.,. mutations), the population "explores" the neighborhood of its current location in that space. That is, it does not search the whole fitness space randomly, but preferentially explores in the immediate neighborhood of a location known to be hospitable to the population. The population does not "search" for fitter phenotypes, but rather happens on to them in the course of producing randomly varying individuals, though most exploratory excursions are either neutral or deleterious. On occasion, a reproductively fitter variant arises, and then (because they are reproductively more successful) the population hangs on to those new variants with a probability that depends on the value of the differentially successful reproduction (the selection coefficient, s).

Casting evolution as a "search" for some specified genotype/phenotype is a snare and a deception, and leads many a creationist into the same swamp Dembski inhabits. Evolution is an exploratory process, not a search process, and the exploration of any given population is confined to a highly constrained volume in the total geno/pheno-space.

Joe Felsenstein · 27 August 2009

RBH said: Casting evolution as a "search" for some specified genotype/phenotype is a snare and a deception, and leads many a creationist into the same swamp Dembski inhabits. Evolution is an exploratory process, not a search process, and the exploration of any given population is confined to a highly constrained volume in the total geno/pheno-space.

Yes. We might think that this would work against the Weasel program, as it searches for a given goal. But another way to think of the example is that it models evolution in a region of the adaptive surface where each matching letter increases fitness, so the different letters do not interact much at all. In such a smooth region, the program shows that selection works well.

Robin · 28 August 2009

RBH said: Casting evolution as a "search" for some specified genotype/phenotype is a snare and a deception, and leads many a creationist into the same swamp Dembski inhabits. Evolution is an exploratory process, not a search process, and the exploration of any given population is confined to a highly constrained volume in the total geno/pheno-space.

I must confess, RBH, that I have a tough time distinguishing (from a marketing/propaganda perspective) between search and exploration in your explanation above. I know what you meant (at least I think I know what you mean) from a scientific analogy perspective, but the terms are so close in meaning in many areas that it's a tough analogy to wrap my head around. I think part of my problem with trying to think of evolution as an exploratory process is that to me exploration implies conscious comparative analysis. I also think of exploring as an activity based some understanding of a preset baseline; a base set that provides some understanding against which someone or something can be "curious" about things that are different". Maybe it's just my thinking on that. At any rate, I tend to think of evolution not so much as exploring, but rather as a facilitor of adaptation to opportunities.

harold · 28 August 2009

Robin -

I must confess, RBH, that I have a tough time distinguishing (from a marketing/propaganda perspective) between search and exploration in your explanation above.

Search - You know what you're trying to find. Exploration - You aren't trying to find a specific thing; you discover things you weren't looking for.

Kevin B · 28 August 2009

Robin said: I must confess, RBH, that I have a tough time distinguishing (from a marketing/propaganda perspective) between search and exploration in your explanation above.

"Search" is going to Wikipedia and putting "Lohengrin" into the Search box. "Explore" is finding "Lohengrin" as one of the "On this day..." items on the Wikipedia main page.

Mike Elzinga · 28 August 2009

Robin said:
RBH said: Casting evolution as a "search" for some specified genotype/phenotype is a snare and a deception, and leads many a creationist into the same swamp Dembski inhabits. Evolution is an exploratory process, not a search process, and the exploration of any given population is confined to a highly constrained volume in the total geno/pheno-space.
I must confess, RBH, that I have a tough time distinguishing (from a marketing/propaganda perspective) between search and exploration in your explanation above. I know what you meant (at least I think I know what you mean) from a scientific analogy perspective, but the terms are so close in meaning in many areas that it's a tough analogy to wrap my head around.

From a marketing and propaganda perspective, the terms can, and are, exploited by the ID/creationist crowd. I have tried, on the two threads about the Dembski/Marks paper, to make the same point RBH is making, and I did it by referring to the underlying physics. Physicists use the terms “search” and “explore” also when referring to physical systems settling into nearby potential wells. In physics, this process is somewhat of a condensation and anthropomorphizing of the process of energy dissipation. It encapsulates things like the second law of thermodynamics in colloquial language. But RBH also mentions initial conditions when he describes the population of individuals settling into nearby areas of fitness space (analogous to settling into potential wells). Often overlooked in this wrangling over genetic algorithms in biology – and their counterparts in physics – is the fact that all physical systems that retain any semblance of their former selves or states as they change in response to changes in the environment are doing so under constraints. If that were not the case, the system would “fly apart” and be destroyed. One of the first things anyone learns as he/she attempts to model a physical system on a computer is that the system has to be defined. That means describing in detail the initial conditions and constraints that apply throughout the problem. As the computer solution to the problem progresses, the constraints must always be maintained. Otherwise you are solving an entirely different problem every time the computer progresses through a solution loop. The ID/creationist crowd appears to know none of this.

I also think of exploring as an activity based some understanding of a preset baseline; a base set that provides some understanding against which someone or something can be “curious” about things that are different”.

I suspect this may be your correct intuition about initial conditions and constraints.

Blake Stacey · 28 August 2009

RBH said: I should have noted that Dawkins' birth certificate shows that he was born in Kenya. That settles it. Right?

Funniest thing I've read all day.

Robin · 28 August 2009

harold said: Robin -
I must confess, RBH, that I have a tough time distinguishing (from a marketing/propaganda perspective) between search and exploration in your explanation above.
Search - You know what you're trying to find. Exploration - You aren't trying to find a specific thing; you discover things you weren't looking for.

Ok...that's a good distinction. Thanks Harold.

Robin · 28 August 2009

Mike Elzinga said:
Robin said:
RBH said: Casting evolution as a "search" for some specified genotype/phenotype is a snare and a deception, and leads many a creationist into the same swamp Dembski inhabits. Evolution is an exploratory process, not a search process, and the exploration of any given population is confined to a highly constrained volume in the total geno/pheno-space.
I must confess, RBH, that I have a tough time distinguishing (from a marketing/propaganda perspective) between search and exploration in your explanation above. I know what you meant (at least I think I know what you mean) from a scientific analogy perspective, but the terms are so close in meaning in many areas that it's a tough analogy to wrap my head around.
From a marketing and propaganda perspective, the terms can, and are, exploited by the ID/creationist crowd. I have tried, on the two threads about the Dembski/Marks paper, to make the same point RBH is making, and I did it by referring to the underlying physics. Physicists use the terms “search” and “explore” also when referring to physical systems settling into nearby potential wells. In physics, this process is somewhat of a condensation and anthropomorphizing of the process of energy dissipation. It encapsulates things like the second law of thermodynamics in colloquial language. But RBH also mentions initial conditions when he describes the population of individuals settling into nearby areas of fitness space (analogous to settling into potential wells). Often overlooked in this wrangling over genetic algorithms in biology – and their counterparts in physics – is the fact that all physical systems that retain any semblance of their former selves or states as they change in response to changes in the environment are doing so under constraints. If that were not the case, the system would “fly apart” and be destroyed. One of the first things anyone learns as he/she attempts to model a physical system on a computer is that the system has to be defined. That means describing in detail the initial conditions and constraints that apply throughout the problem. As the computer solution to the problem progresses, the constraints must always be maintained. Otherwise you are solving an entirely different problem every time the computer progresses through a solution loop. The ID/creationist crowd appears to know none of this.
I also think of exploring as an activity based some understanding of a preset baseline; a base set that provides some understanding against which someone or something can be “curious” about things that are different”.
I suspect this may be your correct intuition about initial conditions and constraints.

Thanks for the elaboration Mike. I guess the underlying problem I have with the term explore is that it implies (at least to me) some level of preconceived intention - that level of curiosity I mentioned earlier. So to me the idea, even if only losely implied, that evolution embodies the "curious poking around" of organisms in a fitness landscape is a complete anathma. Evolution is passive with respect to the pressures and constraints placed on organisms by the environment; it doesn't explore the extent of those constraints as much as those constraints and pressures favor some changes over others. Oh well...I'll keep working with the terms, particularly as Harold provided, and try to wrap my head around it.

Mike Elzinga · 28 August 2009

Robin said: I guess the underlying problem I have with the term explore is that it implies (at least to me) some level of preconceived intention - that level of curiosity I mentioned earlier. So to me the idea, even if only losely implied, that evolution embodies the "curious poking around" of organisms in a fitness landscape is a complete anathma. Evolution is passive with respect to the pressures and constraints placed on organisms by the environment; it doesn't explore the extent of those constraints as much as those constraints and pressures favor some changes over others.

The term “explore” comes up frequently in statistical mechanics when discussing the microscopic energy states of a system. We often say “the system explores” all available microstates. Obviously that doesn’t mean any consciousness on the part of the physical system. It is a shorthand way of saying that, because of the interactions among the constituents of the system, these constituents can keep exchanging energy and momentum until every possible instance of the available permutations and combinations of microstates are realized. In a similar manner we could talk of a complicated system of rivers and tributaries as having explored all the accessible river valleys in the terrain in which they exist.

Midwifetoad · 28 August 2009

I must confess, RBH, that I have a tough time distinguishing (from a marketing/propaganda perspective) between search and exploration in your explanation above. I know what you meant (at least I think I know what you mean) from a scientific analogy perspective, but the terms are so close in meaning in many areas that it's a tough analogy to wrap my head around. I think part of my problem with trying to think of evolution as an exploratory process is that to me exploration implies conscious comparative analysis. I also think of exploring as an activity based some understanding of a preset baseline; a base set that provides some understanding against which someone or something can be "curious" about things that are different". Maybe it's just my thinking on that. At any rate, I tend to think of evolution not so much as exploring, but rather as a facilitor of adaptation to opportunities.

The terms search, explore and adapatation are all unfortunate when debating creationist. All of these terms are retrospective. They imply that something happened as a result of planning or foresight. The difference between Dembski's algorithm and Dawkins' is more than superficial. Dembski's algorithm requires the mutation engine to have knowledge of a "target" in order to exclude correct letters from being mutated. Dawkin's algorithm more nearly models biology because the mutation engine knows nothing of goals, needs, targets and such. The battle over this distinction lies at the very heart of the ID/evolution debate.

Robin · 28 August 2009

Mike Elzinga said:
Robin said: I guess the underlying problem I have with the term explore is that it implies (at least to me) some level of preconceived intention - that level of curiosity I mentioned earlier. So to me the idea, even if only losely implied, that evolution embodies the "curious poking around" of organisms in a fitness landscape is a complete anathma. Evolution is passive with respect to the pressures and constraints placed on organisms by the environment; it doesn't explore the extent of those constraints as much as those constraints and pressures favor some changes over others.
The term “explore” comes up frequently in statistical mechanics when discussing the microscopic energy states of a system. We often say “the system explores” all available microstates. Obviously that doesn’t mean any consciousness on the part of the physical system. It is a shorthand way of saying that, because of the interactions among the constituents of the system, these constituents can keep exchanging energy and momentum until every possible instance of the available permutations and combinations of microstates are realized. In a similar manner we could talk of a complicated system of rivers and tributaries as having explored all the accessible river valleys in the terrain in which they exist.

Yes...I understand how you are using the term now and it makes sense. I'm just not used to using it from a statistical mechanics perspective. Knee-jerk reaction to a far-too ingrained concept of the term from a space-exploration standpoint. I'll work on it. :-)

Robin · 28 August 2009

Midwifetoad said: The terms search, explore and adapatation are all unfortunate when debating creationist. All of these terms are retrospective. They imply that something happened as a result of planning or foresight. The difference between Dembski's algorithm and Dawkins' is more than superficial. Dembski's algorithm requires the mutation engine to have knowledge of a "target" in order to exclude correct letters from being mutated. Dawkin's algorithm more nearly models biology because the mutation engine knows nothing of goals, needs, targets and such. The battle over this distinction lies at the very heart of the ID/evolution debate.

Quite so. I think that's why I am having a hard time with explore - I have this ingrained association of the term with foresight and planning. And yes - the heart of the issue for me is the two distinct concepts - that either organisms are a "target" product (and some "perfect template" is "out there" that all these organisms are supposedly shooting for or are a direct reflection of) or that organisms are just the result of a vast number of ever-changing organic attribute variations, some of which just happened to contain a combination of attributes that fit a given environment at a given moment, thus generating the appearance of an organism that "works well".

Mike Elzinga · 28 August 2009

Robin said: And yes - the heart of the issue for me is the two distinct concepts - that either organisms are a "target" product (and some "perfect template" is "out there" that all these organisms are supposedly shooting for or are a direct reflection of) or that organisms are just the result of a vast number of ever-changing organic attribute variations, some of which just happened to contain a combination of attributes that fit a given environment at a given moment, thus generating the appearance of an organism that "works well".

If there is one thing we can count on from the ID/creationist crowd, it is that they will exploit every possible use and misuse of words. Word-gaming for political gain is their primary tactic. It follows naturally from their having played it for centuries in their wars of “proper reading” of their holy book. Hermeneutics, exegesis, etymology, medieval scholastic word games; this is what they do in order to win the wars that they themselves have started.

GuyeFaux · 28 August 2009

Instead of "explores" or "searches" w.r.t. fitness landscapes, may I suggest "wanders"? The primary definition of "wander" is "move about aimlessly", which seems to not leave the door so open to misrepresentation.

Joe Felsenstein · 28 August 2009

Kudos to Richard Hoppe for the two funniest moments of this thread:

1. Noticing that Denyse O'Leary had offered Richard Dawkins a free copy of his own book, and

2. Uncovering the dastardly conspiracy to cover up the place of birth of Richard Dawkins.

barkdog · 29 August 2009

RE the search vs explore discussion. The difference doesn't matter if the end result must be the exact modern flora and fauna. If an unguided process could have produced different species, God would have been thwarted, creation would be a lie, and evolution would lead inexorably to abortion and euthanasia. Or at least that is how I understand the point of view of the creationists I talk to.

Frank J · 29 August 2009

If there is one thing we can count on from the ID/creationist crowd, it is that they will exploit every possible use and misuse of words. Word-gaming for political gain is their primary tactic. It follows naturally from their having played it for centuries in their wars of “proper reading” of their holy book. Hermeneutics, exegesis, etymology, medieval scholastic word games; this is what they do in order to win the wars that they themselves have started.
— Mike Elzinga

And when even that fails, they just block comments:

But now I seem to be blocked, so my follow-up didn’t come through:
— DiEb

And IDers have the chutzpah to whine that "Darwinists" "censor" them!

VentureFree · 29 August 2009

Here's the conversation that has been repeated ad nauseum since The Blind Watchmaker was first published:

IDist (I): “You claim that the Weasel program acts exactly like evolution, so why doesn’t it?”

Evolutionist (E):”Actually it was never claimed that it acted exactly like evolution. It’s meant only to demonstrate the power of cumulative selection.”

I: “Then why do you claim that it acts exactly like evolution.”

E:”Umm…no one actually claims that.”

I: “Well then show us the original code so that we can see how it supposedly acts exactly like evolution.”

E:”Look, it doesn’t act exactly like evolution. In fact it’s made very explicit that it does NOT act exactly like evolution, but rather demonstrates only one aspect of it.”

I: “Then why won’t you show the original code so we can prove to you that it doesn’t act exactly like evolution like you claim.”

E:”First of all, the code is irrelevant. It’s the algorithm that’s important. Second, any halfway decent programmer can build their own version based on the very clear description in the book. Finally, the algorithm is not meant to act exactly like evolution. It’s a demonstration of cumulative selection. ”

I: “So you admit that you can’t produce the original code, and we’re just supposed to to take your word for it that it acts exactly like evolution.”

E: “This is ridiculous. I’m outta here!”

I: “See, they claim that it acts exactly like evolution but refuse to explain how when asked, so we win!”

Stanton · 29 August 2009

Frank J said:
But now I seem to be blocked, so my follow-up didn’t come through:
— DiEb
And IDers have the chutzpah to whine that "Darwinists" "censor" them!

"Blatant hypocrisy" is a much much more apt term than "chutzpah"

se-rat-o-SAWR-us · 2 September 2009

Efficient perl code for weasel. It allows for variable string lengths, which toggle back and forth, so it obviously doesn't lock, even on the length of the string. Arbitrary printable ASCII mutations are possible, with the hyperlinked starting point of a relevant Richard Dawkins quote. Output: % ./weasel
These people are so unbelievably stupid. --Richard Dawkins (http://bit.ly/Cq8EC)
These people are so unbelievably stupid. --Richard Dawkins (http://bit.ly/Cq8EC
These people are so unbelievably stupid. --Richard Dawkins (http://bit.ly/Cq8E
These people are so unbelievably stupid. --Richard Dawkins (http://bit.ly/Cq8
These people are so unbelievably stupid. --Richard Dawkins (http://bit.ly/Cq
These people are so unbelievably stupid. b-Richard Dawkins (http://bit.ly/Cq
These people a0e so unbelievably stupid. b-Richard Dawkins (http://bit.ly/Cq
These people a0e so unbelie%ably stupid. b-Richard Dawkins (http://bit.ly/Cq
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
METHINKS I]:IH LIKE A WEASEB
METHINKS I]OIH LIKE A WEASEB
METHINKS I]OIH LIKE A WEASE
METHINKS I]OIH LIKE A WEASEo
METHINKS I]!IH LIKE A WEASEo
METHINKS I]!IH LIKE A WEASEo
METHINKS I]-IH LIKE A WEASEo
METHINKS I]-IH LIKE A WEASE'
METHINKS I]-IH LIKE A WEASE
METHINKS I[-IH LIKE A WEASE
. . . . . . . . . . . . . .
METH0NKS IT IS LIKE A WEASEL
METH0NKS IT IS LIKE A WEASEL
METHtNKS IT IS LIKE A WEASEL
METHINKS IT IS LIKE A WEASEL
Perl code:

#!/usr/bin/perl -w

$|=1;
$s="These people are so unbelievably stupid. --Richard Dawkins (http://bit.ly/Cq8EC)";
$e="METHINKS IT IS LIKE A WEASEL";
$try=31; # New offspring per generation.

@e=unpack("C*",$e);

# Print the starting string
$n=0;
# Count the characters in the new string that mismatch the target string
@ns=unpack("C*",$s);
$j=0;
$mn = (($#e<$#ns)? $#e : $#ns)+1;
$mx = (($#e>$#ns)? $#e : $#ns)+1;
$mm = $mx;
while($j<$mn){$mm-- if $e[$j] == $ns[$j++]}
printf("Gen %5d, %-2d mismatches:\t\t\t%s\n",$n,$mm,pack("C*",@ns));

while($mm > 0) {
$i=-1;
# Make $try new strings
while($i++ < $try){
$mmi = $mm;

# Mutate one character of the new string
$chr = int(rand(126-32+1))+32;
$chr[$i] = $chr;

# Delete or add the character to the string
$p = int(rand($#ns+3))-1;
$p = 0 if ($p<0 && $#ns==0);
$p[$i] = $p;

# Count the characters in the new string that mismatch the target string.
if (0 <= $p && $p <= $#ns) { # $ns[$p] changes to chr($chr)
if ($p <= $#e) {
$mmi-- if ($e[$p] == $chr);
$mmi++ if ($e[$p] == $ns[$p]);
}
} elsif ($p == -1) { # $ns[$#ns] is deleted
$mmi-- if ($#ns > $#e);
$mmi++ if ($#ns <= $#e && $e[$#ns] == $ns[$#ns]);
} elsif ($p == $#ns+1) { # $ns[$#ns+1] appended with chr($chr)
$mmi++ if ($#ns >= $#e);
$mmi-- if ($p <= $#e && $e[$p] == $chr);
}
$mmc[$i] = $mmi;
}

# Find high scoring offspring strings.
@sc = sort {$mmc[$a]<=>$mmc[$b]}(0..$#mmc);

@new=(shift @sc);
while(@sc && $mmc[$sc[0]] == $mmc[$new[0]]){push @new,shift @sc}

# Set new string to a random offspring strings from among the high scoring offspring.
$i = int(rand(@new));
$j = $new[$i];
$mm = $mmc[$j];
$p=$p[$j];
$chr=$chr[$j];
if ($p<0) { $#ns = $#ns-1; } #delete
else { $ns[$p] = $chr; } #replace/append character

printf("Gen %5d, %-2d mismatches (\$p=%2d,\$chr=%s):\t%s\n",++$n,$mm,$p,chr($chr),pack("C*",@ns));
}

IanM · 9 September 2009

A friend has been playing with just such a computer model of natural selection. Enjoy!

Davrosfromskaro · 18 September 2009

"But the program has been much discussed on the Internet in the last decade. So where is the code?"

Surely by that logic Microsoft should by now offer Windows up as Open Source...

IE "Windows has been much discussed on the Internet in the last (nearly) 2 decades. So where is the code?"

Bill Gates must be trembling in fear...

Kevin B · 18 September 2009

Davrosfromskaro said: "But the program has been much discussed on the Internet in the last decade. So where is the code?" Surely by that logic Microsoft should by now offer Windows up as Open Source... IE "Windows has been much discussed on the Internet in the last (nearly) 2 decades. So where is the code?" Bill Gates must be trembling in fear...

Stop that right now! You'll have Ms O'Leary going after Dawkins with an anti-trust suit, like the DoJ used against IBM in order to force them to open up the market for the plug-compatible manufacturers.

stevaroni · 18 September 2009

So where is the code

I've been thinking about the red herring of front-loading for a while and one of the ironies that struck me is that you don't even need a computer to run the weasel algorithm. You could "easily" do it with a bingo hall full of 50 undergrads, each one of which had a bucket of balls with the letters A through Z and the numbers 1 through 23. (And if I remember my college days correctly, a hall full of students comes much closer to the idealized "monkeys with typewriters" model than any computer simulation ever could.) In the beginning, each student would randomly pluck letter balls out of the bucket 23 times to produce a random starting string. He's then tot up the string's score. A "bingo caller" in the front of the room would then call out "OK, anybody got a 1?, How about a 2? a 3?" Once the high score was identified, the string would be called out, and everybody would write down that "child". Then they'd reach into the bucket, pull out a random number and letter and "mutate" that position with that letter, repeating the process. Once people got good at it, you could probably do a cycle per minute, or 60 cycles an hour. The irony is that humans would be much worse than machines at adding up the scores and transferring the winning strings, thereby making errors and adding an entirely new level of randomness to the process. Given the extant computer simulations, even with that small sample, the odds are very good that you could get "Methinksitislikeaweasel" in one moderately long day. Although the method would be clunky (hey-evolution is clunky), the beauty is it would have the advantage of being in a very understandable medium. Most people can wrap their heads around the idea of reaching into a bucket and pulling out a ball. It would be very dificult even for the most rabid IDiots to make a charge of "algortihm bias" when the "algorithm" is the same one used to assure that the evening Lotto is a fair game.

Mike Elzinga · 18 September 2009

stevaroni said:
So where is the code
You could "easily" do it with a bingo hall full of 50 undergrads, each one of which had a bucket of balls with the letters A through Z and the numbers 1 through 23.

One could even look at the evolution of language within one’s lifetime. We see changes in slang just within a few decades. Just looking at cooption of words such as “cool”, “hot”, “gay”, “blown-away” etc., shows new uses for old words. Then there are all those acronyms and abbreviations from texting (How about that word texting?). I don’t know if it has anything to do with aging, but it seems to me that language and words are changing even more rapidly, just within American English alone, than it has in the past. Without the stabilizing effects of tradition, dictionaries, and scientific usage (all of which are quite conservative), it would be even more difficult for seniors to talk with their grandchildren and great grandchildren.

Mike Elzinga · 18 September 2009

And I almost forgot; look at what the ID/creationists are doing with their constant word games, distortions of meaning and their introduction of jargon and memes into their propaganda.

They don’t like evolution, yet it is the larger part of their shtick.

Kevin B · 18 September 2009

stevaroni said:
So where is the code
I've been thinking about the red herring of front-loading for a while and one of the ironies that struck me is that you don't even need a computer to run the weasel algorithm. You could "easily" do it with a bingo hall full of 50 undergrads, each one of which had a bucket of balls with the letters A through Z and the numbers 1 through 23. (And if I remember my college days correctly, a hall full of students comes much closer to the idealized "monkeys with typewriters" model than any computer simulation ever could.) In the beginning, each student would randomly pluck letter balls out of the bucket 23 times to produce a random starting string. He's then tot up the string's score. A "bingo caller" in the front of the room would then call out "OK, anybody got a 1?, How about a 2? a 3?" Once the high score was identified, the string would be called out, and everybody would write down that "child". Then they'd reach into the bucket, pull out a random number and letter and "mutate" that position with that letter, repeating the process. Once people got good at it, you could probably do a cycle per minute, or 60 cycles an hour. The irony is that humans would be much worse than machines at adding up the scores and transferring the winning strings, thereby making errors and adding an entirely new level of randomness to the process. Given the extant computer simulations, even with that small sample, the odds are very good that you could get "Methinksitislikeaweasel" in one moderately long day. Although the method would be clunky (hey-evolution is clunky), the beauty is it would have the advantage of being in a very understandable medium. Most people can wrap their heads around the idea of reaching into a bucket and pulling out a ball.

I've got a version of the algorithm you describe running on a 1960's mainframe. With a possibly more "aggressive" fitness function and 100 32-character strings per generation, I get a cycle time similar to your 1 minute, and a run time of around 40 minutes. Of course, this is probably not precisely Dawkins' algorithm, since precisely one position is mutated in each candidate string. What you need to do is discard the bucket of numbers and replace it with one containing a very large number of red and black balls, with the number of red balls wrt the total number being proportional to the mutation rate. Each student works along their string, pulling a ball out of the bag for each position. If they pull out a red ball they replace the letter at that position with a random letter drawn from the bucket. The problem is, of course, that the students will know what the correct letter for each position is, and so it is possible that latching (or as it should properly be called, "cheating") may occur due to this inherent imperfection.

stevaroni · 18 September 2009

Kevin B said: I've got a version of the algorithm you describe running on a 1960's mainframe.

I'm sure the algorithm is quite nice, but I'm actually more fascinated by the idea that you have a 1960's mainframe out in the garage...

wile coyote · 18 September 2009

stevaroni said: I'm sure the algorithm is quite nice, but I'm actually more fascinated by the idea that you have a 1960's mainframe out in the garage...

I'm still trying to get it running on my Babbage Engine. Machining the gears is taking some time ... Hey, if we're going to dream, let's dream Steampunk.

Mike Elzinga · 18 September 2009

Anyone know a good soroban lubricant? Mine is starting to smoke.

I have the one shown in the upper picture.

fnxtr · 18 September 2009

Try sesame oil. It might not work but at least it'll smell good.

sswitaj · 19 September 2009

Anyone know a good soroban lubricant?

No, but not long ago, I had to find some steam locomotive oil. My father was a retired machinist, and apparently, at some point in every machinist's life, they start dabbling with steam engines. Some of them even band together to dabble in really big steam engines. Impressive hobby, but dramatically inconvienent, at least from a garage-space point of view. Steam locomotives have very particular lubrication needs for their cylinders. Cylinders are (were) typically cast iron, with the resultant tendency to corrode unless protected, and the oil has to form clingy, films in an environment that's very hot, subject to mechanical scraping, and bathed in steam that's changing phase between dry and saturated five times a second. The odd part is that there were companies that still make the stuff, although you had to buy a 55 gallon drum of it at a time. On the other hand, if you have a locomotive around, you know you're going to use 55 gallons of locomotive oil, eventually, so you might as well keep it handy.

Mike Elzinga · 19 September 2009

sswitaj said:
Anyone know a good soroban lubricant?
Steam locomotives have very particular lubrication needs for their cylinders. Cylinders are (were) typically cast iron, with the resultant tendency to corrode unless protected, and the oil has to form clingy, films in an environment that's very hot, subject to mechanical scraping, and bathed in steam that's changing phase between dry and saturated five times a second.

:-) This is a particularly good context in which to bring up these properties of steam engine lubricants. The ID/creationists’ criticisms of the Dawkins Weasel program revolve around “smuggling the answer into the program”. But we in the science community know something they don’t. All systems in nature interact; not only with the surrounding environment, but the constituents of these systems interact with each other. So the IDiot perceptions of Dawkins program are dead wrong. The properties you describe for steam engine lubricants are a nice description of the gooey world of organic chemistry and biological systems.

Mike Elzinga · 19 September 2009

Kevin B said: I've got a version of the algorithm you describe running on a 1960's mainframe.

While I was flying recently on a trip, I started outlining a version for the HP48/49/50 series of calculators. Then I’ll also try to replicate it for the TI-89 and the TI-83/84 series. I’ll be using vectors or lists instead of strings; it’s faster on these machines. I think I may have to trim down the sizes of the target, parent, and offspring in order to keep from running the batteries down during the program execution. We’ll see. It might also be useful to allow different probabilities for the mutations of each element; i.e., a different probability for going from “incorrect” to “correct” than for vice-versa. That would allow for latching as well as for investigating more realistic conditions that would suggest it more probable for mutations in the “correct” direction than in the “incorrect” direction as the offspring approach greater fitness.

Kevin B · 19 September 2009

stevaroni said:
Kevin B said: I've got a version of the algorithm you describe running on a 1960's mainframe.
I'm sure the algorithm is quite nice, but I'm actually more fascinated by the idea that you have a 1960's mainframe out in the garage...

I don't, but I know a computer museum that does.... (Although "garage" is not a very good description of a building that was probably the first erected specifically to house an electronic computer.)