Uncommon Descent Serving The Intelligent Design Community

Uncommon Descent Contest Question 10: Provide the Code for Dawkins’ WEASEL Program

Share
Facebook
Twitter
LinkedIn
Flipboard
Print
Email

Special invitation for Richard Dawkins – but any civil person is entitled to enter.

There’s been some discussion here and elsewhere whether the the recent IEEE article by Dembski and Marks correctly characterizes Richard Dawkins’ famous METHINKS IT IS LIKE A WEASEL program.

Does the program ratchet correct letters or does it let them vary? One is a partitioned or stair-step search, the other a more realistic evolutionary search. From The Blind Watchmaker, where Dawkins describes the program, its performance suggests that it could be either of these options (though he doesn’t say).

On the other hand, from a (video-run of the program , go to 6:15), it seems to be the latter.

It’s easy enough to settle this question: Make the code for the program public. Perhaps Richard Dawkins himself or his friends at RichardDawkins.net can finally provide this code (apparently a program written in BASIC).

The prize is a copy of either Stephen Meyer’s new Signature in the Cell or Richard Dawkins’ soon-to-be-out The Greatest Show on Earth.

Should the winner choose the latter, I will ask Dawkins’s publicist to mail the copy. Given that at his site, he calls himself “the most formidable intellect in public discourse,” I would assume that if he signs the copy, it will be worth millions.

But wait. Let’s see that code first.

Comments
-kf,
This, too relies on the silly strawman algorithm manufactured out of a didactic example intended to illustrate what partitioning does. Namely: once a letter goes correct in a given generational champion, it remains that way all the way tot he target – as Weasel o/ps showcased in 1986 credibly show.
Dembski and Marks describe the algorithm they use to get their didactic example fairly well, I think. But I don't even have to know their algorithm (and therefore, don't have to come up with a strawman algorithm), as they give me their first and second generation in their paper. And this second generation looks quite different to a second generation derived by Dawkins's algorithm, implying that there is a difference in the design of the algorithms used. For me, that's the relevant context.DiEb
September 9, 2009
September
09
Sep
9
09
2009
03:20 AM
3
03
20
AM
PDT
Dieb: It is fair comment that you know the relevant context. G'day. GEM of TKIkairosfocus
September 9, 2009
September
09
Sep
9
09
2009
01:50 AM
1
01
50
AM
PDT
--kf
This of course would give rhe appearance of a sequence of generational champions that has high reversions, so no latching or quasi-latching behaviour. The argument is an insincere strawman intended to create a “gotcha.”
You lost me. I introduced the ten strings in 115, fully explaining them as
I took the string SCITAMROFN*IYRANOITULOVE*SAM and calculated a next generation using Dawkins’s algorithms with populations of 10,50 and 100 – and mutation rates of .04, .05 and .1. The tenth string in the list is the second generation given in the paper of Mark and Dembski.
and I asked:
Can anyone spot a difference in the design of the strings? Anyone? KF? Anyone?
I repeated this question in 263, 288, and, in a shorter form in 291 and 297 until you came up with your reply. I don't think that I was misleading any attentive reader.
DiEb
September 8, 2009
September
09
Sep
8
08
2009
11:11 AM
11
11
11
AM
PDT
PPS: All A's are B's is a set relationship. Some A's are B's is a set relationship with existential import -- at least one member of A, a exists, and is also a member of B, another set. No A's are B's is a statement about non-overlap of sets, and some A's are not B's is a statement that at least one member of A, a exists, and is not in set B as well. (And I am leaving off for the moment the issue of the empty set, the division between classical and modern logic.) The undistributed middle term is a case where relevant overlap fails. In the case of run B, it is an existing example of an implicitly latched, ratcheting, partitioned search. And, there is a dynamical connexion such that latched searches will be partitioned. So, if a is a member of A then it will be partitioned search due to a mechanism that causes ratcheting behaviour. So, in terms of the sets at work:
All partitioned searches latch: --> partitioned seqarches, P are a subset of latched searches, L --> but, suppressed context: due to the involved dynamics, we can also see that Latched searches L, are a subset of partitioned searches P --> That is, the two sets are equivalent, due to the dynamics of ratceting Kf’s imagined Weasel [Run B] latches. --> there exists b, such that it is a member of L --> but also we know though the dynamics of ratcheting, that members of L are also members of P, the two sets being equivalent. Therefore kf’s imagined Weasel [run B] is a partitioned search --> As, we have reason to see why members of L will be members of P. --> In short the imagined fallacy is begging the question of the relationship between sets L and P. And, we have reason to see that L is not a proper subset of P but an equivalent set to P. --> And, see how thinking in terms of sets untangles the complexities of syllogistic reasoning? [Thank you, Irving Copi!]
kairosfocus
September 8, 2009
September
09
Sep
8
08
2009
02:24 AM
2
02
24
AM
PDT
PS: Onlookers, note that the ratcheting and latching o/p behaviour to be explained is that for generational champions as Weasel 1986 showcased. This can be achieved explicitly by using say a mask register to lock up successful letters from further change. It may also be achieved implicitly through a mechanism that uses a population size big enough and a per letter mutation rate small enough such that by overwhelming probability, each generation seeded from a previous champion or the original string will have in it unchanged members [and/or those with a letter changed from one incorrect form to another that does not affect already correct letters.] In such a generation, with a relatively small size such that double etc letter mutations are rare enough not to show up among generation champions in a significant fraction of runs, no increase in proximity and single step increases will dominate behaviour. Under these conditions it will not be hard to find runs that show cumulative progress to target without reversions of successful letters of generational champions. Such runs are implicitly latched, and show ratcheting without slips (NB: as the "dog" wears in say a baitcaster reel, the anti-reverse action can show occasional slips). Run B above is a case in point of such implicit latching. (Beware of the rhetorical tactic of objectors who will try to slip in the definition that latching can only mean explicit latching.)kairosfocus
September 8, 2009
September
09
Sep
8
08
2009
02:03 AM
2
02
03
AM
PDT
Onlookers: We first and foremost observe above no compunction over abusive, manipulative and disrespectful rhetoric and many misleading arguments, even in the face of step by step corrections. Just, rushing on to he next objections as though nothing more serious than a debate game were at stake, or vanishing behind the cloud of poisonous smoke from burning, ad hominem soaked strawmen. We should therefore take the prudence to act in light of the clear observation of High Machiavellian tendencies at work. (And, we should note how the widespread acceptance as "science" of the amoral philosophy of evolutionary materialism naturally promotes such tendencies, as was discussed yesterday in the other thread.) Now, let us focus a few moments on several further rhetorical gambits: 1] Dieb, 310: I calculated a next generation using Dawkins’s algorithms with populations of 10,50 and 100 – and mutation rates of .04, .05 and .1. This of course would give rhe appearance of a sequence of generational champions that has high reversions, so no latching or quasi-latching behaviour. The argument is an insincere strawman intended to create a “gotcha.” Sad. 2] 310, The string Dembski’s and Marks’s algorithm produces as a second generation is quite different in design from strings which nine runs of Dawkins’s program produce with various parameters as the best string of the second generation. Of course, irt was already correctivley pointed out that there is no one “Dembski’s and Marks’s algorithm”, as the GUI hosted at EIL shows: several algorithms, taking a diversity of apporaches to Weasel. Instead, despite this being already corrected, we have a strawman mischaracterisation of the didactic example to illustrate what partitioning of a search means. (Observe onlookers, Dieb NEVER engages the import of the cluster of Weasels sponsored by EIL, which would at once show that there is no reasonable reading of p. 1055 of the IEEE paper that would support the conclusions being advanced through a patent – and now plainly willful -- misrepresentation.) And, when the silly strawman algorithm is applied to the string used by Dawkins, then it will produce the intended silly results. A cheap gotcha but not an honest one. Inexcusable. And, ever so sad. 3] DNA-J, 311: DiEb’s point: you cannot get a Weasel search to look like a partitioned search. This, too relies on the silly strawman algorithm manufactured out of a didactic example intended to illustrate what partitioning does. Namely: once a letter goes correct in a given generational champion, it remains that way all the way tot he target – as Weasel o/ps showcased in 1986 credibly show. And, Weasels with the implicit latching and ratcheting to the target as described by Dembski and Marks are long since DEMONSTRATED, on EIL's Weasel ware. It seems I need to reproduce, e.g. Run B from April 9, 2009 – the day when the issue was definitively settled by demonstration of latching, quasi-latching, and non-latching behaviour using the EIL “proximity reward” search algor (one of their stable of weasel type algors): ____________ RUN B, 500 pop/gen, 4% per letter mut rate: 1. MEL LSI YHXMAJLMDGMVKTSKGW 2. MEL LSI YHXIAJLMDNMVKTSKGW 3. MEL LSI YHXISJLMDNMJKTSKGW 4. MEL LSI YHXISJLMDN JKTSKGW 5. MEL LNI YHXISJLDDN JKTSKGW 6. MEL LNI YHXISJLDDN JKTEKGW 7. MEL LNB BHXISJLDDN JKTEKGE 8. MEL LNB BHXISJLIDN JKTEKGE 9. MEL LNB BHXISJLIDN JKTEKSE 10. MEL LNB BHXISJLIDN JKTEKSEL 11. MEL LNK BHXISJLIDN JKTEKSEL 12. MEL LNK BHXIS LIDN JKTEKSEL 13. MET LNKV BHXIS LIDN JKTEKSEL 14. MET LNKV BHXIS LIDN AKTEKSEL 15. MET LNKV BHXIS LIDE AKFEKSEL 16. MET LNKV BHXIS LIKE AKFEKSEL 17. MET LNKS BHXIS LIKE AKFEKSEL 18. MET LNKS BH IS LIKE AKFEKSEL 19. MET LNKS BH IS LIKE AKFEKSEL 20. MET LNKS BH IS LIKE AKWEKSEL 21. MET INKS BH IS LIKE AKWEKSEL 22. MET INKS BH IS LIKE AKWEKSEL 23. MET INKS BH IS LIKE AKWEKSEL 24. MET INKS IH IS LIKE AKWEKSEL 25. MET INKS IH IS LIKE A WEKSEL 26. MET INKS IH IS LIKE A WEASEL 27. MET INKS IH IS LIKE A WEASEL 28. METHINKS IH IS LIKE A WEASEL 29. METHINKS IH IS LIKE A WEASEL 30. METHINKS IH IS LIKE A WEASEL 31. METHINKS IT IS LIKE A WEASEL ______________ This is a case of implicit latching or at least quasi-latching. (As I recall, the correct letters went red, and no red letter reverted to black.) You will see that the 31-gen run of generation champions to the weasel target is actually a little better than the 43 gen one published by Dawkins in 1986 in BW, probably due to the population size. (4% odds of mutation per letter would give about 1 letter mutated per member of a generation on average.) This – contrary to DNA-J's confident declarations -- is an implicitly latched, ratcheting, cumulatively selected, progressive run of generational proximity champions to target. And, it looks rather like the two runs published by Dawkins in 1986, except that this is not every tenth generation shown, but every generation shown. QED. Since April 9th, and linked directly from point 15 my App 7, the always linked. (That is, the claims I am correcting were made in the context of refusing to first check easily accessible facts.) 4] Repeating a fallacy: I will slightly update my correction of DNA-J's fallacy, as I have given a concrete example -- run B of April 9 -- from an undeniable Marks and Dembski- approved, EIL-sponsored algorithm:
All partitioned searches latch. [true by definition] Kf’s imagined Weasel [Run B above, based on EIL's real world Weasel program] latches [implicitly]. [SUPPRESSED, MATERIAL CONTEXT: Latching and ratcheting, in the context of Weasel algorithms with the relevant behaviour similar to that credibly observed in the showcased examples by Dawkins c 1986, are two aspects of the same process. "Latching" focusses on how in generational champions, once letters go correct under certain circumstances, they remain that way in future generations all the way to target. "Ratcheting," on how such locked in successful letters then provide the baseline for progress as futher letters are discovered and similarly locked in as the search homes in on the target through hotter-colder proximity signals. The search is therefore partitioned so that successful letters in generational champions to date are preserved from loss of correct state. This can be done explicitly or implicitly, as Run B demonstrates. In either case, the result is cumulative progress to target – i.e. By successive addition of successful letters. And, once we observe the latching effect, the mathematical analysis presented in the M & D IEEE paper, p. 1055 applies, as was already discussed by the undersigned, above. Thus, ratcheting and latching cannot be dynamically separated in the context of such a weasel run. And, Run B above is an illustration of this in action.] Therefore kf’s imagined Weasel [Run B above, an implicitly latched, ratcheting, cumulatively progressive, partitioned and selected search, based on the Marks and Dembski EIL "proximity reward search" algorithm] is a partitioned search [by virtue of the dynamics or mechanism of a latched, ratcheted search]
In short, the attempt to force-fit case B etc into the imagined fallacy of the undistributed middle term [forgive my error of memory on this . . . it's been a long time since I thought in terms of syllogistic terminology – I usually just draw the relevant Venn diagram and directly see the set relationships instead of bothering with the complexities of classical syllogistic arguments] is based on a misunderstanding of the observable dynamical relationship between latching, ratcheting and cumulative selection in weasel type algorithms. In short, one cannot properly separate latched searches from partitioned searches, as ratcheting behaviour and associated cumulative progress to target are the dynamical bridge that inherently and inseparably joins the two. That is, the set of partitioned searches is here EQUIVALENT to the set of latched, ratcheting searches: it is necessary and sufficient for a search to be partitioned in the sense used by Marks and Dembski on p. 1055 of their IEEE paper, that it be (latched and ratcheting). ____________ GEM of TKIkairosfocus
September 8, 2009
September
09
Sep
8
08
2009
01:49 AM
1
01
49
AM
PDT
kairosfocus is very eloquent, but he continues to get everything wrong.
It seems I will only be able to get Dieb off this irrelvancy by being more detailed. The ten sequences selected of course are in part a run of a Weasel type program with reversions.
Err. No they are not. Dieb's post 115 describes how he ran 9 Weasels that cover the range of typical input values for the parameters Generation Size and Mutation Rate,to see how many letters would change in the first generational champion for a Weasel. The tenth row is the result for a partitioned search, as accurately described in D&M. DiEb's point: you cannot get a Weasel search to look like a partitioned search. And, as I pointed out mathematically in post 305, you cannot get a partitioned search to look like the TBW example without millions of computer.years. I note that kairos "it's just a flesh wound" focus has failed to provide any substantive response to any of the points raised in post 305. If any of the onlookers doubt this, I will happily clarify for them. But there is little point in arguing with someone who, when a logical fallacy in their argument is pointed out to them, responds with the same fallacious argument... From post 285:
Oh dear. We can agree that partitioned search, as accurately described by D&M in eqn22, latches. Unfortunately, you are using this fact to conclude that a (hypothetical) latching search is a partitioned search. All partitioned searches latch. Kf’s imagined Weasel latches. Therefore kf’s imagined Weasel is a partitioned search All ostriches have two legs Kf has two legs Therefore Kf is an ostrich.
It is a famous logical fallacy. Surely you can see this. You cannot (e.g. 308) use it to try to prove anything kf.DNA_Jock
September 7, 2009
September
09
Sep
7
07
2009
12:08 PM
12
12
08
PM
PDT
--kf,
The ten sequences selected of course are in part a run of a Weasel type program with reversions.
No, they are not. Especially not the tenth. I'm afraid you misunderstood my post. As is said in 288:
Dembski and Marks start in their paper with the phrase: SCITAMROFN*IYRANOITULOVE*SAM I calculated a next generation using Dawkins’s algorithms with populations of 10,50 and 100 – and mutation rates of .04, .05 and .1. The tenth string in the list is the second generation given in the paper of Mark and Dembski.
So, it's not about latching, quasi-latching, or whatever. It's just a comparison of the best strings of a second generation. And to spell it out: The string Dembski's and Marks's algorithm produces as a second generation is quite different in design from strings which nine runs of Dawkins's program produce with various parameters as the best string of the second generation. This difference in design shows that the algorithms are different in design...DiEb
September 7, 2009
September
09
Sep
7
07
2009
05:41 AM
5
05
41
AM
PDT
5] 305, All at once begging the question, and (brilliantly) re-defining the word “partitioning” (which has a specific mathematical meaning in D&M’s paper) to mean something entirely different. And thus, since HE thinks latching and partitioning are synonymns, any attempt to argue that there is a difference between them, or between mechanism and behavior, is “driving a rhetorical wedge between latching and ratcheting”, apparently. Of course, I have already shown how M & D defined partitioning, in terms of once a letter becomes on-target, the search for that letter is over. That is, it is locked in. progress to date is preserved, and further progress comes as new letters are guessed, resulting in ratcheting progress and cumulative selection to the target. But, the unwary, overly trusting reader might not check up to see that this is the case. 6] 305, you invent totally novel meanings for words in order to avoid admitting defeat. ‘Mechanism’, ‘partitioning’, and the strictly telic use of ‘cumulative’. Outright, brazen falsehood. First, on two of the terms, I have simply cited the AmH Dict, a well known generally used, respected dictionary, and on the third I have demonstrably used the sense of the authors in question: CUMULATIVE: 1. Increasing or enlarging by successive addition. (In context, note how the showcased runs used by Dawkins c 1986 NEVER show a correct letter reverting. If you say cumulative in a context of such successive increases in proximity without exception, that makes the meaning clear but to those determined to wriggle out of the implications.) MECHANISM: 3. An instrument or a process, physical or mental, by which something is done or comes into being: (Here, I am the original person to speak. This is the sense in which I spoke of mechanism, a sense tat would be familiar to any Physicist, chemist or engineer.) PARTITIONING: Am H D. 2. To divide or separate by means of a partition. M 7 D, p. 1055: “In partitioned search, our search for these letters [once they have gone correct] is finished. ” (this one I did not use a dictionary on, but Marks and Dembski's terms. THEY are the ones using the word in their context, and we need to identify THEIR sense. And the dictionary meaning makes their use seem very reasonable.) And, onlookers, it is clear that I have sought instead to correctly understand and address the situation, so the demonising ad hominems are utterly uncivil and unwarranted. DNA-J, you owe me an apology and retraction. 7] 305, Citation with approval from BillB: Partitioning means to divide up the search into separate, independent units. WEASEL is not an algorithm that searches for individual letters, it is not partitioned. First, the facts to be explained are the sampled and showcased runs of generational champions, c. 1986:  1 WDL*MNLT*DTJBKWIRZREZLMQCO*P 2? WDLTMNLT*DTJBSWIRZREZLMQCO*P 10 MDLDMNLS*ITJISWHRZREZ*MECS*P 20 MELDINLS*IT*ISWPRKE*Z*WECSEL 30 METHINGS*IT*ISWLIKE*B*WECSEL 40 METHINKS*IT*IS*LIKE*I*WEASEL 43 METHINKS*IT*IS*LIKE*A*WEASEL  1 Y*YVMQKZPFJXWVHGLAWFVCHQXYPY 10 Y*YVMQKSPFTXWSHLIKEFV*HQYSPY 20 YETHINKSPITXISHLIKEFA*WQYSEY 30 METHINKS*IT*ISSLIKE*A*WEFSEY 40 METHINKS*IT*ISBLIKE*A*WEASES 50 METHINKS*IT*ISJLIKE*A*WEASEO 60 METHINKS*IT*IS*LIKE*A*WEASEP 64 METHINKS*IT*IS*LIKE*A*WEASEL These runs show that in over 200 cases where letters go correct, they never once revert in the samples, in a context where it is common for incorrect letters to remain incorrect form decades of generations. So, it is credible – especially given the enthusiastic words in BW on “cumulative selection” [remember we have two multiplying lines of evidence here]-- to see that the showcased runs did not have any reversions. We have been able to replicate such an effect of ratcheting-latching progress to target by generational champions by two algorithmic mechanisms: explicit latching and implicit latching, as has been demonstrated. These two mechanisms have the same effect on the relevant generational champions, and they will see the same pattern of accelerated progress to target relative to baseline random walk search. So, on explaining what was to be explained [ratcheted progress of generational champions], the two are plainly valid. Further to this, the D & M analysis as presented – and as already discussed by the undersigned -- pivots on the latched effect; which as we note applies to generational champions. Therefore also, the internal processing difference between the two classes of mechanism [explicit latching mechanisms protect all mutants from varying successful letters, implicitly latched cases do not] – which is what is being pounced on and highlighted – is an irrelevancy to what was to be explained. But, a very convenient one for those determined – regardless of costs to truth, reasonableness and civility -- to portray a mythical Dembski algorithm that diverges from the Dawkins algorithm sufficiently that Dembski can be impugned as making misrepresentations. But -- as the very un-awarded status of the prize for this thread testifies -- to begin with there is no known THE Dawkins Weasel algorithm c 1986. Just, reasonable reconstructions, and they can come in varieties that latch letters explicitly or implicitly. AND THE EIL SPONSORED BY DEMBSKI AND MARKS OFFERS A CLUSTER OF SUCH ALGORTIHMS THAT COVERS THE BASES. 8] Dieb, 207: Isn’t that like gluing moths to trees – for tutorial reasons? Turnabout insinuation that tends to create an unwarranted perception of immoral equivalency. BACKDROP: One of the misleading icons of evolution was that he pepper moths presented from the 1950's as illustrating how moths rested on tree trunks and it was contrasting colour that explained population shifts due to differential predation, as pollution darkened the trunks. Turns out moths don't usually rest on trunks and the pictures were based on moths glued t three trunks where they did not usually rest. By contrast, Dembski and Marks on p. 1055 of the IEEE paper simply say that they are illustrating what partitioning is about, emphasising that the point is that when letters go correct [in context, for generational champions], they remain that way in future champions till the target is hit. This holds for explicitly latched and implicitly latched Weasel algorithm mechanisms. Their analysis – as was previously shown in 302 -- is premised on the fact of latching, not the mechanism at work underneath. When the EIL sponsored by these men does present definite Weasel algorithms [with the source code available through a zip], it presents a cluster that covers the bases. So, there is no good reason to infer to misrepresentation of reality on M & D's part. (by contrast, when he presented Weasel in BW, Dawkins had to admit that it was misleading in important ways ,” which I have discussed already.] _____________ The pattern of distraction, distortion, demonisation and dismissal by darwinist advocates, sadly, plainly continues. That speaks volumes on the implications and impact of the amorality that is inextricably intertwined with the institutionally predominant evolutionary materialism that ever so plainly dominates Darwinland [formerly known as Christendom]. GEM of TKIkairosfocus
September 7, 2009
September
09
Sep
7
07
2009
04:40 AM
4
04
40
AM
PDT
Onlookers I have already had occasion to speak this morning to the turn that the blog has taken over the weekend, as we see the from Darwinist unresponsiveness to the responsibility issues that were addressed in the comments at 301 - 303 above. So, I will just link there (as done above), then call us back to footnotes on points that need a further remark or two, as I proposed to do yesterday: 1] Dieb, 304: ID is about the detection of design, therefore I’m not surprised that you have spotted a difference in the design of the strings below, and that it annoys you. Declaring it irrelevant won’t make it get away… It seems I will only be able to get Dieb off this irrelvancy by being more detailed. The ten sequences selected of course are in part a run of a Weasel type program with reversions. Such is of course a distinct case that had Dieb troubled to read my App 7 the always linked, he would have seen explicitly addressed in point 16, where I discuss quasi-latching. (Having already linked demonstrated cases of implicit latching. Indeed, in the just linked thread form March-April, I then went on to give several runs that showed various other types of behaviour. All duly accessible and ignored in the haste to set up and knock over a strawman.) Beyond quasi-latching, there is also the – discussed as well -- case where there is no apparent latching at all. And, it is explicitly the case that Weasel latches for “good runs” implicitly when the pop per gen, mut rate per letter and proximity to target filter interact to give that effect. Remember: DEMONSTRATED. Nowhere have I said that only explicit latching and implicit latching are possible. Just the opposite. Going further, the rhetorical point being made is that there is a Demsbki algorithm that is ever so distinct from the Dawkins algorithm, that it is a case of misrepresentation. But in fact the Marks-Dembski EIL, as already noted, hosts a GUI that gives a set of various Weasel algorithms, covering the bases for the various types and sub types that have been mostly discussed. Similarly, from the 1986 descriptions and showcased run excerpts the only thing we can see is that we can determine no one THE Dawkins algorithm. Explicit and implicit latching interpretations are reasonable on the evidence c 1986 [why there is a prize for actual credible code c 1986 – not forthcoming!], and we have no reason to immediately accept that the video runs of 1987 are the same algorithm as the one showcased in 1986. It is on subsequent reported statements c 2000 that it is held on balance of evidence that implicit ratcheting-latching is the likeliest explanation for the showcased behaviour c 1986, on which the 1987 video would be either a run of individual mutants or a detuned Weasel run that shows non-latched behaviour. All of which has been long since discussed and just as long since repeatedly ignored or distorted as inconvenient to the rhetorical agenda being pushed by the darwinists. 2] DNA-J, 305: It takes a very special kind of person to simultaneously maintain that the opening pair:SCITAMROFN?IYRANOITULOVE?SAM. (20) OOT?DENGISEDESEHT?ERA?NETSIL. (21) is too “good” to be a real example, and is instead a “didactic example”, because it goes from 2 hits to 7 in one generation This neatly side-steps the fact that M & D explicitly introduced the example as a didactic one,as already cited at : E Partitioned Search Partitioned search [12] is a “divide and conquer” procedure best introduced by example. Consider the L =28 character phrase . . . And, it is not that it is “too good” an example that the case goes forward five letters in one gen, but that multiple go-corrects are rare in the config space, so it is unlikely that the didactic example would be a real world run; especially when we know as well that EIL sponsors a clutch of Weasel algorithms, none of which will typically behave like that. This is an agenda-serving failure to read reasonably and charitably in light of relevant context. (I would not want to be under inquisition at the hands of such men, nor should you, dear onlooker. Such, plainly will twist anything you say or do or do not say to suit themselves. Do you see why it is therefore dangerous to cede real institutional power to such?) 3] 305: Unfortunately for kf, the first two generations in TBW are:WDL*MNLT*DTJBKWIRZREZLMQCO*P WDLTMNLT*DTJBSWIRZREZLMQCO*P Which only changes in two places, out of 25 incorrect letters. The run in question is: >> 1 WDL*MNLT*DTJBKWIRZREZLMQCO*P 2? WDLTMNLT*DTJBSWIRZREZLMQCO*P 10 MDLDMNLS*ITJISWHRZREZ*MECS*P 20 MELDINLS*IT*ISWPRKE*Z*WECSEL 30 METHINGS*IT*ISWLIKE*B*WECSEL 40 METHINKS*IT*IS*LIKE*I*WEASEL 43 METHINKS*IT*IS*LIKE*A*WEASEL>> In the initial phrase, 1, three letters are already correct: N, *, T. in the next one just one letter goes correct in addition, S, as highlighted. The first asterisk also goes to an M. So, we see a double letter change, one going correct – which explains the advance towards target. To get his “unfortunately,” DNA-J imposes the idea that M & D have an algorithm working in which all incorrect letters vary but all but two them happen to pick the original state out of the 27 available. But this is a case of assuming that there is an algorithm in which all such letters change in every generation, and putting it in my mouth (not to mention those of M & D). My argument, however, was precisely that there is no such algorithm at work on M & D's part, but an explicitly illustrative example of what partitioning is. [Cf 1 and 2 above.] In short this is a strawman game at work. It is easy to set up a strawman to look silly, but that has nothing to do with the reality of what is being pointed out. In this case, an illustration of the EFFECT of partitioning: In partitioned search, our search for these letters [once they have gone correct] is finished. , as M & D said, and as I cited above and highlighted. Note again: M & D do sponsor the publication of Weasel Algorithms, which can be genuinely deemed M & D Weasel Algorithms. NONE OF THE ALGORITHMS THEY SPONSOR SHOW THE PATTERN OF BEHAVIOUR THAT IS BEING STRAWMANNISHLY CARICATURED TO MAKE ME LOOK SILLY AND TO MAKE THEM SOUND DISHONEST. NONE: NADA, ZIP, ZILCH. It is fair comment to note, therefore, that M & D , when they do make Weasel type algors [through Atom acting as their agent], do not make anything that looks like the caricature being set up and knocked over. And, since this has been repeatedly pointed out but ignored in the rush to gleefully pummel a strawman, to insist on the strawman in the teeth of such a reasonable alternative is less than fair-minded and I daresay, less than honest. But then, ever since the days of Alcibiades [cf. my link to the other thread this morning], for evolutionary materialists the inescapable is-ought gap and cosequent inherent amorality and relativism of such avant garde atheism have always meant that in the end, might makes right. And we onlookers should take due note of how that leads such men to behave when they are in a position to owe duties of care but heeding the voice of duty runs counter to the agenda. Then, we should think twice about ceding further institutional or general policy making power to such men. 4] 305, KF accuses me of “trying to drive a rhetorical wedge between latching and ratcheting”. I call him on this, saying . . . . To prove his point (that I drive a rhetorical wedge between ‘latching’ and ‘ratcheting’) he quotes my post highlighting his logical fallacy, which does not even contain the word ‘ratcheting’ : . . . Contrary to how the matter is presented in 305, a glance up at 302, point 5 will show:
I am being accused of making false accusations. However, a scrollup will show that a primary reference is e.g. DNA-J, 285:
Oh dear. We can agree that partitioned search, as accurately described by D&M in eqn22, latches. Unfortunately, you are using this fact to conclude that a (hypothetical) latching search is a partitioned search. All partitioned searches latch. Kf’s imagined Weasel latches. Therefore kf’s imagined Weasel is a partitioned search
But, instead of this strawman and red herring muddles on undistributed middles etc, the defining characteristic of partitioning of search as just highlighted, as used by Marks and Dembski, is that correct letters are latched and ratcheting progress builds on what is achieved by making successive additions of correct letters through searching, until the whole target is cumulatively achieved. And, I am taking “partitioning” in the light of that usage, whatever other uses there may be in other contexts . . .
In short, I first highlighted that the strawmannish syllogism DNA-J set up was about the undistributed middle. Then, I showed that “latching” and “ratcheting” spoke to the fact that PARTITIONED SEARCH, as used by M & D, is about the locking in of successful letters in generational champions, and the cumulative progress to target as more and more letters become latched. Thus, we see the evident ratcheting and latching effects in Weasel showcased run excerpts, c. 1986. (I then went on to discuss how the M & D analysis of the performances of such partitioned search was PREMISEED on the latching of successful letters.) The suppression of context therefore twisted my words into yet another silly strawman, which was duly pummelled. And the non-mention of “ratcheting” in the specific excerpt was just a distractive red herring. For, I SHOWED that ratcheting and latching are simply two sides of the same coin. To see why that is so, let's re-examine the strawman, given that latching and ratcheting are two views of the same process:
All partitioned searches latch. [True, by definition] Kf’s imagined Weasel latches. [Weasel c. 1986 credibly latches.] [Latching and ratcheting are two sides of the same process, whereby successful letters in a search will be retained in further generations of champions until the target is hit. This, being by M & D's definition for this context, partitioned search.] Therefore kf’s imagined Weasel [a weasel that latches] is a partitioned search
See the difference? (And, see how the strawman was laced with loaded word ad hominems, too; with what rhetorical effect?) [ . . . ]kairosfocus
September 7, 2009
September
09
Sep
7
07
2009
04:34 AM
4
04
34
AM
PDT
I have pointed out in particular that it is a reasonable reading of the Dembski-Marks paper on p. 1055 that they intend to provide a brief tutorially oriented explanation, rather than realistic output of an algorithm.
Isn't that like gluing moths to trees - for tutorial reasons?DiEb
September 6, 2009
September
09
Sep
6
06
2009
05:46 AM
5
05
46
AM
PDT
DNA-J: Onlookers will be able to see that you have unfortunately distorted the matter, by comparing your remarks to mine in light of the cited section of the IEEE paper. I have pointed out in particular that it is a reasonable reading of the Dembski-Marks paper on p. 1055 that they intend to provide a brief tutorially oriented explanation, rather than realistic output of an algorithm. [Your remarks and too much of the above remind me of Kipling's poem If, and not happily.] Onlookers will be able to tell for themselves what is more reasonable. Having had to take up BL this morning already point by point, I will deal with your further remarks here anon. G'day. GEM of TKIkairosfocus
September 6, 2009
September
09
Sep
6
06
2009
04:06 AM
4
04
06
AM
PDT
It takes a very special kind of person to simultaneously maintain that the opening pair:
SCITAMROFN?IYRANOITULOVE?SAM. (20) OOT?DENGISEDESEHT?ERA?NETSIL. (21)
is too “good” to be a real example, and is instead a “didactic example”, because it goes from 2 hits to 7 in one generation (Probability = 1 in 419, as previously pointed out to kf in post 164) Yet he continues to maintain:
Here, it is obviously not practically feasible to have a single member per generation case that varies 26 letters and is sufficiently likely to be observable, to get 5 new letters at one go. Instead we would be looking at a very large population of mutants per generation with an aggressive filter that will pick up far tail of distribution effects such as multiple letters going correct that can see a case of 5 letters going correct.
and
I have also shown why (most recently from 279 on) it is that on p. 1055 of the IEEE paper, it is proper to conclude that Dembski and Marks used a didactic illustration of what ratcheting-latching action is like — with a link to BW that gives the real-world run in view!
Unfortunately for kf, the first two generations in TBW are:
WDL*MNLT*DTJBKWIRZREZLMQCO*P WDLTMNLT*DTJBSWIRZREZLMQCO*P
Which only changes in two places, out of 25 incorrect letters. The probability that a partitioned search (as described by D&M) would do this is = 3.3 x 10^-31 For those who dislike sceintific notation, that’s one in 2.9 million million million million million. So
SCITAMROFN?IYRANOITULOVE?SAM. (20) OOT?DENGISEDESEHT?ERA?NETSIL. (21)
Is too unlikely to be a real example, but
WDL*MNLT*DTJBKWIRZREZLMQCO*P WDLTMNLT*DTJBSWIRZREZLMQCO*P
Which is a trillion trillion times more unlikely, is just fine as an example. Note that the unusual behavior of generation 2 (given kf’s assumptions) was pointed out to kf in post (34), and many times since (190,199, etc). I even gave D&M the benefit of the doubt, allowing a mutation rate less than 100%. But you still cannot get a run that looks like the first run in TBW, using a partitioned search. (I did have fun though – the optimum occurs close to a 49% mutation rate – at which point generation #2 is fairly likely (1 in 32,471 runs), but the odds against finishing within 43 queries has grown to 1 in 23,992,787. So even with the generous assumption that these two are independent, we still have a probability of getting a partitioned search that changes 2 or fewer out of 25 at query two and hits within 43 queries = one in 0.7 million million. If we assume 2 minutes per weasel run, CRD would need to run a thousand computers continuously for nearly 3,000 years to get a partitioned search this weird. So I don’t think TBW is a good exemplar of the search described by equation 22. I will anticipate kf’s red herring: (remember that kf is maintaining that D&M were justified in citing TBW as an example of partitioned search, because of the behavior of the outputs), so any suggestion that CRD messed with the results would 1) be an completely unfounded accusation of dishonesty 2) concede our point that the citation is obviously inappropriate And now for a quick recap of the semantics battle – KF accuses me of “trying to drive a rhetorical wedge between latching and ratcheting”. I call him on this, saying
if you were referring to me, then you are just making stuff up….ellipsis… I would be happy if we could agree to use one of the “latching”/”ratcheting” pair to refer to mechanisms, and the other to refer to behaviors, but when I tried to make this behavior/mechanism distinction, you promptly re-defined the word “mechanism”.
[see footnote 1] To prove his point (that I drive a rhetorical wedge between ‘latching’ and ‘ratcheting’) he quotes my post highlighting his logical fallacy, which does not even contain the word ‘ratcheting’ :
Oh dear. We can agree that partitioned search, as accurately described by D&M in eqn22, latches. Unfortunately, you are using this fact to conclude that a (hypothetical) latching search is a partitioned search. All partitioned searches latch. Kf’s imagined Weasel latches. Therefore kf’s imagined Weasel is a partitioned search
(I went on to use the same fallacious logic to prove that kf is an ostrich, but no matter.) And then kf says :
But, instead of this strawman and red herring muddles on undistributed middles etc, the defining characteristic of partitioning of search as just highlighted, as used by Marks and Dembski, is that correct letters are latched and ratcheting progress builds on what is achieved by making successive additions of correct letters through searching, until the whole target is cumulatively achieved. And, I am taking “partitioning” in the light of that usage, whatever other uses there may be in other contexts.
All at once begging the question, and (brilliantly) re-defining the word “partitioning” (which has a specific mathematical meaning in D&M’s paper) to mean something entirely different. And thus, since HE thinks latching and partitioning are synonymns, any attempt to argue that there is a difference between them, or between mechanism and behavior, is “driving a rhetorical wedge between latching and ratcheting”, apparently. Note that kf has previously quoted from post 269, so we may assume that he read it and is therefore aware of the concept that a latching search may not be a partitioned search. Yet he continues to hammer away, repeatedly asserting what he knows to be a logical fallacy. So, in conclusion, it appears that anytime you lose an argument, kairosfocus, you invent totally novel meanings for words in order to avoid admitting defeat. ‘Mechanism’, ‘partitioning’, and the strictly telic use of ‘cumulative’. “Well, there’s glory for you” H.Dumpty BillB took 25 words to nail the issue at hand:
Partitioning means to divide up the search into separate, independent units. WEASEL is not an algorithm that searches for individual letters, it is not partitioned.
My own summary:
D&M’s Eqn22 CANNOT describe a search that has generational champions, and therefore it CANNOT describe Weasel (irrespective of pseudo-quasi-implicit-latching)
Foot Note 1: kf devised a novel definition for “mechanism” in his post 233 , which included this GEM:
I firmly believe -- on good grounds -- that unless necessary causal factors are present an event CANNOT happen, and unless sufficient ones are present it WILL not happen. So if it happens, we can identify causal factors and how they work -- i.e mechanisms. And given the issue of synergy -- effects due to interaction -- mechanisms do not have to be explicitly built and labelled as such. I do not believe in magical poofery!
Reeeeally.DNA_Jock
September 5, 2009
September
09
Sep
5
05
2009
01:00 PM
1
01
00
PM
PDT
I have had no good reason to debate resemblances or non-resemblances among the strings listed by Dieb, which at best are distractingly irrelevant.
Well, ID is about the detection of design, therefore I'm not surprised that you have spotted a difference in the design of the strings below, and that it annoys you. Declaring it irrelevant won't make it get away... 1. SCITAMROFN*IYRANOIEULOVE*SAM 2. SCITAMROFN*IYRANOITULOGE*SAM 3. ECITAMRI*N*IYZANOITULOVE*SAM 4. SCITAMROFN*IYRANOITUL*VE*SAM 5. SCITAMROFN*IYRANOITULOVE*SEM 6. SCITAMOOLNOIYRAMOITULOVE*SEM 7. SCITANROFN*IYYANOITULOVE*SAM 8. SCITIMROFN*JYRANOITULOVE*SAM 9. SCITAMROFN*ICRHNOITSLOWE*SAV 10. OOT*DENGISEDESEHT*ERA*NETSILDiEb
September 5, 2009
September
09
Sep
5
05
2009
07:11 AM
7
07
11
AM
PDT
7] I have a better FSCI link for you We need not bother with the hurled elephant tactic of a glitteringly general claim based on a substantially irrelevant literature bluff and dump, as the first article by Hazen – as discussed several times at UD -- suffices to show what is going on. Excerpting, starting with the opening words: ____________ >> Complex emergent systems, in which interactions among numer-ous components or ‘‘agents’’ produce patterns or behaviors not obtainable by individual components, are ubiquitous at every scale of the physical universe, for example in neural networks (1), turbulent fluids (2), insect colonies (3), and spiral galaxies (4). Complex systems also appear in a range of artificial symbolic contexts, including genetic algorithms (5), cellular automata (6), artificial life (7), and models of market economies (8) . . . . complex system. Furthermore, the ancient transition from a geochemical world to a living planet may be modeled as a sequence of emergent events, each of which increased the chemical complexity of the prebiotic world (9–11) . . . . The function of some emergent systems is obvious: a sequence of letters communicates a specific idea, a computer algorithm performs a specific computation, and an enzyme catalyzes at least one specific reaction. Less obvious are the functions of systems of many interacting inanimate particles, such as mole- cules, sand grains, or stars, but these systems may also be described quantitatively in terms of function, for example, in terms of their ability to dissipate energy or to maximize entropy production through patterning (e.g., refs. 26–29). Living systems, by contrast, typically display multiple essential functions (21, 30, 31). This consideration of complexity in terms of the function of a system, as opposed to some intrinsic measure of its patterning or structural intricacy, distinguishes our treatment from many previous efforts. >> ________________ i --> Immediately, the article conflates complexity of order or even randomness with complexity of functional organisation; something that was wared against by Orgel all the way back to 1973:
In brief, living organisms are distinguished by their specified complexity. Crystals are usually taken as the prototypes of simple well-specified structures, because they consist of a very large number of identical molecules packed together in a uniform way. Lumps of granite or random mixtures of polymers are examples of structures that are complex but not specified. The crystals fail to qualify as living because they lack complexity; the mixtures of polymers fail to qualify because they lack specificity. [Leslie Orgel (1973). The Origins of Life, p. 189 ]
ii --> Similarly, Hazen fails to recognise at the outset characteristic distinctions between the actions of agents and those of chance and or necessity. That is, he is begging the question by failing to see that design is an empirically evident, characteristic and distinct phenomenon. iii --> Almost immediately thereafter, he fails to distinguish between a MODEL whose plausibility rests on the imposition of Lewontinain materialism on science, and the observational reality and challenges of getting form any plausible prebiotic geochemical environment and the organised, functional complexity of life. iv --> Along the way, he manges to assume – without providing a clear and balanced discussion of he evidentiary challenges and the real world frustrating empirical findings of OOL researches since 1953 -- that there is a smooth pathway of small steps from simple chemicals to organised life forms which reproduce themselves using a von Neumann replicator: coded blueprint, code system with associated algorithms and data structures, reader, effector mechanisms. v --> He then diffuses the focused meaning of specific functionality by introducing dis-analogous cases: any old pile of sand will function as a pile of sand, but not any old string will function in a given context as a meaningful, code and information bearing functional entity that says something coherent or properly instructs the assembly of a protein's amino acid sequences. And modest perturbation of the one will have but little effect: you still have a sand pile. For the other, modest perturbation destroys function, i.e we have sharply distinct islands of specific function. That is why the work of Abel, Trevors and others on functional sequence complexity and its quantification, is so important. (Cf my discussion of FSCI and related ideas here onlookers. Also cf the Weak Argument cor4ectives above.) 8] TA, 299: I have been accused of being a “Darwinist”; a baseless and unfounded slur for which not a shred of evidence has been provided. If you choose to run with the wolves, you are hereby provisionally identified and treated as one of same. Sorry if you end up as a pet thats gets shot while running with the wolves. (Cf here, the current media trumpeted case of "civilians" taking gasoline from Taliban, who have just hijacked tankers and murdered the civilian crews.) 9] KF offers examples of latching and claims this as implicit latching, but is unable to offer a clear definition of implicit latching that I can detect This has of course been given many, many times over, starting with the App 7 and being again given above:
13 --> Letterwise partitioned search is also a very natural way to understand the Weasel o/p in light of Mr Dawkins' cited remarks about cumulative selection and rewarding the slightest increment to target of mutant nonsense phrases. As such, it has long been and remains a legitimate interpretation of Weasel. However, on recently and indirectly received reports from Mr Dawkins, we are led to understand that he did not in fact explicitly latch the o/p of Weasel, but used a phrasewise search. 14 --> Q: Can that be consistent with an evidently latched o/p? ANS: yes, for IMPLICIT latching is possible as well.   15 --> Namely, (i) the mutation rate per letter acts with (ii) the size of population per generation and (iii) the proximity to target filter to (iv) strongly select for champions that will preserve currently correct letters and/or add new ones, with sufficient probability that we will see a latched o/p. (This effect has in fact been demonstrated [the link here is the one I have been using all along] through runs of the EIL's recreation of Weasel. [Link is to the development version of EIL Weasel with adjustable parameters]) 16 --> in a slightly weaker manifestation, the implicit mechanism will have more or less infrequent cases of letters that will revert to incorrect status; which has been termed implicit quasi-latching. This too has been demonstrated, and it occurs because an implicit latching mechanism is a probabilistic barrier not an absolute one. So, as the parameters are sufficiently detuned to make reversions occur, we will see quasi-latched cases.  Sometimes, under the same set of parameters, we will see some runs that latch and some that quasi-latch.
Thus, BOTH latching and quasi latching have been defined long isnce in easily accessible materials, and that in terms of how they are achieved. In latching-ratcheting action, generational champions in a Weasel algorithm run have letters that once they go correct are subject to catch and keep [locked up pr latched – think of a panfish stringer with latching snaps here] not catch and release. That may be achieved explicitly or implicitly, as has been demonstrated. 10] Nephmon, 300: in the Dembski algorithm, latching is an explicit and central part of the algorithm itself. Strawman. There is of course no “Dembski algorithm,” other than the cluster hosted by the EIL, which includes the explicitly latched and proximity search with adjustable parameters cases that will show latching implicitly under certain circumstances. (That is, once you show the reasonable reader behaviour that recognises a purely didactic illustration as opposed to presentation of an algorithm and its realistic output. This was already discussed above for today.) As to the idea that latching action is a mere side effect, the gushing comments of CRD as discussed above show that c. 1986, he did not regard this as a mere side effect but a remarkable achievement: “cumulative selection” -- of non-functional mutant “nonsense phrases” on mere proximity to target -- which beats mere “single step selection” -- which requires achievement of reasonably complex function before selection on superiority of function -- by millions and millions and millions of times over. 11] As an aside, if you implement my suggestion of several messages and alter the target phrase slightly in each generation, the WEASEL algorithm would continue to work and converge on the (moving) target, whereas the Demski algorithm would totally fail because it’s already latched into place values that might change in subsequent generations Red herring led away to Strawman. What is to be explained is not something someone makes up for rhetorical purposes today, but he behaviour of Weasel c 1986, as showcased and thus also in absence of credible original code – the actual focus for this thread, onlookers – the behaviour of reasonable replicas thereof. Explicit latching is one reasonable reconstruction [on just the 1986 information], and implicit latching is another which becomes preferred if we reckon with the statements made by CRD and agents in the years since about 2000. _____________ GEM of TKIkairosfocus
September 5, 2009
September
09
Sep
5
05
2009
05:49 AM
5
05
49
AM
PDT
] DJM: Talk about “soaked in oil of ad hominem!” Turnabout, unwarranted [cf the discussion just above], immoral equivalency accusation. Sad. 4] DIEB, 297: I understand your last post that you see no difference in the design of the following ten strings: I have of course focussed my discussion on the Weasel 1986 runs of generational champions as discussed at 1 above again. I have had no good reason to debate resemblances or non-resemblances among the strings listed by Dieb, which at best are distractingly irrelevant. (The strings we need to discuss are reproduced at 1 above, Dieb.) I have also shown why (most recently from 279 on) it is that on p. 1055 of the IEEE paper, it is proper to conclude that Dembski and Marks used a didactic illustration of what ratcheting-latching action is like -- with a link to BW that gives the real-world run in view! -- which then led up to the analysis of probability on the observation of latching. And latching of successful letters in the runs of generational champions, of course, can be achieved explicitly or implicitly; as has been demonstrated. This has been twisted into a strawmannish “algorithm” that would not be feasible of contruction. All, to support the claim that M & D distorted what Dawkins presented in his example and discussion of Weasel c. 1986. Here is the relevant section of the IEEE paper, again: __________________ >> E. Partitioned Search Partitioned search [12] is a “divide and conquer” procedure best introduced by example. Consider the L =28 character phrase METHINKS ? IT ? IS ? LIKE ? A ? WEASEL. (19) Suppose that the result of our first query of L =28 charac-ters is SCITAMROFN ? IYRANOITULOVE ? SAM. (20) Two of the letters {E, S} are in the correct position. They are shown in a bold font. In partitioned search, our search for these letters is finished. For the incorrect letters, we select 26 new letters and obtain OOT ? DENGISEDESEHT ? ERA?NETSIL. (21) Five new letters are found, bringing the cumulative tally of discovered characters to {T, S,E, ?,E, S,L}. All seven char-acters are ratcheted into place. The 19 new letters are chosen, and the process is repeated until the entire target phrase is found. Assuming uniformity, the probability of successfully identi-fying a speci?ed letter with sample replacement at least once in Q queries is 1 - (1 - 1/N)^Q, and the probability of identifying all L characters in Q queries is q = (1 - (1 - (1/N))^Q)^L. (22) >> ______________________ Notice, first, the explicit context of explanation by didactic example, as opposed to presentation of a realistic algorithm and output data from such. Here, it is obviously not practically feasible to have a single member per generation case that varies 26 letters and is sufficiently likely to be observable, to get 5 new letters at one go. Instead we would be looking at a very large population of mutants per generation with an aggressive filter that will pick up far tail of distribution effects such as multiple letters going correct that can see a case of 5 letters going correct. Such would be functionally equivalent to the more realistic case of multiple intervening generations, with the highlighted characteristic: “[once a cluster of letters goes correct in the champions] our search for these letters is finished.” That is, latching of correct letters and ratcheting progress to target. And, credibly, that is what the linked discussion in BW is about, as we may see from 1 above. How can such latching of correct letters in generational champions be accomplished? Ans: in two ways, explicitly or implicitly. Once latching-ratcheting of the run of generational champions is present, and with number of Queries being seen as number of mutants to date, the probability analysis follows trivially. Just, we should note that Q will jump in clusters associated with G generations of size S. Of course, I do not expect those determined to take as suspicious as possible an interpretation of the Marks-Dembski paper will find this convincing. I only aim to show that such are being unreasonable and selectivley hyperskeptical, especially by contrast with the degree of charity they wish us to give to Mr Dawkins, who has said in so many words that his presentation of Weasel and its results is “misleading in important ways.” 5] DNA-J, 298: if you were referring to me, then you are just making stuff up. [This on my remarks that someone was trying to “drive a rhetorical wedge between ratcheting and latching”] In short, I am being accused of making false accusations. However, a scrollup will show that a primary reference is e.g. DNA-J, 285:
Oh dear. We can agree that partitioned search, as accurately described by D&M in eqn22, latches. Unfortunately, you are using this fact to conclude that a (hypothetical) latching search is a partitioned search. All partitioned searches latch. Kf’s imagined Weasel latches. Therefore kf’s imagined Weasel is a partitioned search
But, instead of this strawman and red herring muddles on undistributed middles etc, the defining characteristic of partitioning of search as just highlighted, as used by Marks and Dembski, is that correct letters are latched and ratcheting progress builds on what is achieved by making successive additions of correct letters through searching, until the whole target is cumulatively achieved. And, I am taking “partitioning” in the light of that usage, whatever other uses there may be in other contexts. Worse, as we may see from the mathematical probability analysis above, the factor that is used to deduce the results in Eqn 22 is that such searches latch successful letters so that once achieved the letter stays locked in: catch and keep, not catch and release. That is why we may see that the probability of a given letter being correct in Q tries is (1 - [1 – 1/N]^Q), with N possible letter-states and assumed flat random distribution of guesses:
--> probability of guessing a given letter correctly in any one throw are 1/N. --> probability of missing the letter on that throw is 1 – 1/N --> that of missing it all of Q times become 1 – 1/N]^Q --> So that of catching it on any one of the tries in the Q tries is 1 - [1 – 1/N]^Q) --> this builds int eh implication that once captured a letter does not slip back out again, i.e it is latched. --> Eqn 22 just extends this to a phrase of length L letters, on the assumption of probabilistic independence; and with the “no catch and release” rule in effect, we will see both latching and ratcheting action in the run of generational champions
As long since described and repeatedly explained and demonstrated, this latching-ratcheting result can be achieved explicitly or implicitly. It is also credibly what we see in the results showcased for Weasel 1986. 6] DNA-J: BillB’s first paragraph was a response to Joseph’s Humpty-Dumpty attempts to re-define the word “cumulative” . . . I would be happy if we could agree to use one of the “latching”/”ratcheting” pair to refer to mechanisms, and the other to refer to behaviors, but when I tried to make this behavior/mechanism distinction, you promptly re-defined the word “mechanism”. Turnabout accusation. First, as point 1 above for today shows, latching and ratcheting and cumulartive selection have a very particular context in which it is plain that not only modern recreations of Weasel ratchet and latch the o/ps [through mechanisms that are explicit and implicit, but credibly so did the generational champion runs showcased by CRD in BW and New Scientist in 1986. That is an important datum line. Next, cumulative has a plain meaning and in the context of the showcased runs that is the most natural meaning. If CRD had meant to say that Weasel showed slippages, he sure picked a peculiar way to say so! Nor is it a question-begging after the fact redefinition attempt to point out that
a] AT ISSUE: mechanism has to do with the means by which a result is achieved, here cumulative, no slip-back ratcheting-latching action as may be observed definitively from modern recreations of Weasel, and on the preponderance of evidence for the 1986 runs as enthused over and showcased by CRD; cf. Point 1 above. b] EPISTEMOLOGICAL FOUNDATION: We are discussing an inference to the best explanation on the empirical data of the showcased weasel runs c 1986, and we are also discussing a program that implements a simulation. (Much of science works by such empirically anchored inference to best explanation.) c] DECISIVE POINT: the observed o/p results of the program – runs of generational champions -- are produced by the behaviour of the program, which in turn is determined by its mechanisms and how it therefore takes in inputs [parameter settings, filter characteristics, original random letter seed] and processes such. d] ROLE OF UNDERLYING MECHANISM: It does so by creating mutant generations of a set size on a set per letter mutation rate, filtering on mere proximity to target and picking the nearest to date to become the next seed till the final champion hits home. e] ALTERNATIVE MECHANISMS: In so doing – as demonstrated -- latching and ratcheting o/p results and associated behaviour can turn up based on explicit processing mechanisms, or implicit processing mechanism that rely on the interaction of the input parameters and the filter.
And, in a context of confusion at best driven by misunderstandings leading to misrepresentations, being explicit and precise is far more important than being brief. [ . . . ]kairosfocus
September 5, 2009
September
09
Sep
5
05
2009
05:42 AM
5
05
42
AM
PDT
Onlookers: The idea hitman games, lamentably, continue. I therefore first draw your attention to the always linked, App 7, in which the core issues were addressed step by step dating to April last. Similarly, observe the pattern of corrected Darwinist errors and uncivil tactics in 279 - 81 above. In short, were the objectors above really concerned to be accurate, balanced and fair, they have more than adequate materials to engage on the merits. But, sadly, we are quite plainly dealing with the rhetoric of distraction, misrepresentation and mischaractersation intended to poison the atmosphere of civil discussion, and to enable oppressive misbehaviour; as has been highlighted with all too many cases in point above. Also, while -- sadly -- to many of the commenters above have by indulgence in incivility and/or enabling behaviour for oppressive misconduct, have moved beyond the pale of civil discussion on the merits of a serious matter, it is about time we do another point by point cleanup on the record, addressing some of the grosser fallacies and agit-prop tactics, as a further 101. Of course this is at the risk of dismissive rhetoric about “word salads” etc – which simply proves that there is plainly no intent to engage on the merits cogently to wards a true and fair result, only to distract, distort, demonise and dismiss. But, for the record, starting with the main current objection: 1] DJM, 296: So you have never once seen a complete run of Dawkins Weasel program and consequently have no evidence to think that it latches whatsoever beyond your own inability to comprehend the purpose of the program or the phenomena it illustrates. This is another case of brassy pounding repetition to create the false impression of truth: “Don't believe yer lyin eyez and the dictionary.” For, the relevant showcased 1986 runs of Weasel's generational proximity to target champions – as has been reproduced above and as has been in App 7 all along – are (from BW and NewScientist): ______________ >>  1 WDL*MNLT*DTJBKWIRZREZLMQCO*P 2? WDLTMNLT*DTJBSWIRZREZLMQCO*P 10 MDLDMNLS*ITJISWHRZREZ*MECS*P 20 MELDINLS*IT*ISWPRKE*Z*WECSEL 30 METHINGS*IT*ISWLIKE*B*WECSEL 40 METHINKS*IT*IS*LIKE*I*WEASEL 43 METHINKS*IT*IS*LIKE*A*WEASEL  1 Y*YVMQKZPFJXWVHGLAWFVCHQXYPY 10 Y*YVMQKSPFTXWSHLIKEFV*HQYSPY 20 YETHINKSPITXISHLIKEFA*WQYSEY 30 METHINKS*IT*ISSLIKE*A*WEFSEY 40 METHINKS*IT*ISBLIKE*A*WEASES 50 METHINKS*IT*ISJLIKE*A*WEASEO 60 METHINKS*IT*IS*LIKE*A*WEASEP 64 METHINKS*IT*IS*LIKE*A*WEASEL >> ______________ It is easy to directly observe that for over 200+ letters that could easily revert to incorrect status, were the “there is no credibly observable latching effect” claim true, WE NEVER SEE A SINGLE CASE OF REVERSION. That, from a sample of 300+ relevant letters, and in a context where there are many cases of incorrect letters persisting for decades of generations of champions. So, on good runs of Weasel, c 1986, such reversions are at least rare enough to not be easily observable – indeed, they are not observed at all in the actual evidence to be accounted for; that is, in the sense to be further described, showcased Weasel runs of generational proximity to target champions, c. 1986, evidently and observably ratcheted towards the target and latched correct letters in so doing. (All that is required to see that is to accept that one normally showcases typical but “good” behaviour of results in scientific and similar work.) And, the latching effect is precisely the observation that when a letter goes correct, it stays that way all the way home to the target, even as in successive generations, Weasel finds further letters to lock in as on target. Creating the ratcheting action with cumulative – AmhD: “Increasing or enlarging by successive addition” -- progress to the target. Nor is this just a matter of our observation of the published results in two different venues. For, here is how Mr Dawkins enthusiastically described Weasel's action, in BW: ____________ >> It . . . begins by choosing a random sequence of 28 letters ... it duplicates it repeatedly, but with a certain chance of random error – 'mutation' – in the copying. ,i>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 . . . . What matters is the difference between the time taken by cumulative selection, and the time which the same computer, working flat out at the same rate, would take to reach the target phrase if it were forced to use the other procedure of single-step selection: about a million million million million million years. This is more than a million million million times as long as the universe has so far existed . . . . >> _______________ CRD here speaks explicitly about cumulative selection, rewarding slightest increment to target, and contrasts the progress of non-functional nonsense phrases through targetted search with the effects that a random walk that would have to first meet a threshold of reasonable functionality before being rewarded with generational champion status, would have. In short, by words and examples, CRD makes it plain that latching, ratcheted action is the simplest and clearest sense of his language; consistent with the showcased outputs. (And, it has been demonstrated that such ratcheting, latched action can be achieved by two basic mechanisms – AmHD: 3. An instrument or a process, physical or mental, by which something is done or comes into being -- explicit latching and implicit. There has been an attempt to try to define that the former is the only correct sense of the term so that references to latching or ratcheting distort what Dawkins showcased and gushed over in 1986. But of course a term that describes an observable effect is not something one can game around rhetorically like that, on pain of the absurdity of “Don't believe yer lyin eyez.” [Latching is a description of an output pattern in the letters of runs of generational champions, and it is demonstrated that this effect occurs, as well as that it can be achieved by explicit code or implicit matching of parameters: generation size and mutation rate coupled to proximity to target filter characteristics.]) The real reason this has subsequently become “controversial” is that the latching ratcheted action as nonsense phrases creep in to the target vividly reveals the fundamental dis-analogy between Weasel and the claimed power and mechanism of random variation and natural selection based on improved functional fitness. Worse, as Marks and Dembski demonstrated, once we see ratcheting-latching so that a correctly guessed letter does not in practice revert, for a good run, a fairly simple probability analysis obtains, which leads to the simple deduction of the advantage conferred by the active information embedded in the targetting and artificial selection on mere proximity of non-functional phrases confers. 2] DJM, you slag Dawkins and call him dishonest! Actually, I have not called Mr Dawkins “dishonest.” I have instead repeatedly drawn attention to his writings in BW, and in particular the implications of the further words in his comments on Weasel's action in BW, as has again been immediately accessible all along in App 7 the always linked and as has been pointed out step by step many times in various UD comment threads: ______________ >> Although the monkey/Shakespeare model is useful for explaining the distinction between single-step selection and cumulative selection, it is misleading in important ways. One of these is that, in each generation of selective 'breeding', the mutant 'progeny' phrases were judged according to the criterion of resemblance to a distant ideal target, the phrase METHINKS IT IS LIKE A WEASEL. Life isn't like that. Evolution has no long-term goal. There is no long-distance target, no final perfection to serve as a criterion for selection, although human vanity cherishes the absurd notion that our species is the final goal of evolution. In real life, the criterion for selection is always short-term, either simple survival or, more generally, reproductive success. [TBW, Ch 3, as cited by Wikipedia, various emphases added.] >> ______________ These words make it plain that at the time of publication, CRD knew that Weasel's artificial cumulative selection of non-functional “mutant nonsense phrases” on resemblance to a distant ideal target was fundamentally dis-analogous to the claimed ability of random variation and natural selection on superiority of biological function. That is why he had to acknowledge that it was misleading in important ways.” And it is a plain duty of care of the educator – public or otherwise – not to mislead. More specifically, Weasel works to gain an advantage over a random walk reference base because it pre-loads a target and then rewards hooter-colder guesses, without reference to reasonable thresholds of complex function. And, it was precisely the need to achieve bio-function and improvements in bio-function, with all the implied complexity, that was at the heart of the background to Weasel: the challenge raised by Sir Fred Hoyle and by others going back to Schutzenberger and others at the famous Wistar consultation at the top level, of 1966, on the challenges to get to complex function in initial and novel biological systems. In short, Weasel begged the question at stake in a misleading way. And, even the qualifying – aka “weasel” [this might have something to do with the very phrase chosen as target . . . ] -- words cited above exploit a yet subtler rhetorical effect: the headlined, spectacular simulation grabs and focusses attention, but the qualifications will be easily and generally – predictably – overlooked. (Just as a misleading headline often smears a person, but few ever read the qualifications and corrections buried deep in the story or on a back page a few days later.) Currently as well, the Marks-Dembski analysis on active information and its role explains the performance gain of Weasel over random walks amidst large configuration spaces with islands of code-based complex functionality. Again, it is the duty of educators to not mislead. So, this tactic by DJM, regrettably, is plainly a case of turnabout (“he hit back first . . .”)accusation. [ . . . ]kairosfocus
September 5, 2009
September
09
Sep
5
05
2009
05:38 AM
5
05
38
AM
PDT
Sorry, I ducked out for a while because, well, I have a life. Yes, Gaz, I used read Private Eye many moons ago. I have one question for kairofocus that requires a yes/no answer. Let's see if he can answer it in fewer than 1000 words. As we know, the implicit "latching" in WEASEL is dependent on the parameters of the algorithm. It's very easy to make it not latch simply by choosing a suitably high p(mutation), or a low n(offspring). Thus "latching" isn't central to the algorithm, but a side-effect of its operation when run with certain parameters. On the other hand, in the Dembski algorithm, latching is an explicit and central part of the algorithm itself. No change to input parameters to the program will cause it not to exhibit latching, because it's hard-coded to do so. So my question is: in the light of this, do you think that there is any substantial semantic equivalency between the two algorithms? (As an aside, if you implement my suggestion of several messages and alter the target phrase slightly in each generation, the WEASEL algorithm would continue to work and converge on the (moving) target, whereas the Demski algorithm would totally fail because it's already latched into place values that might change in subsequent generations. I wonder which one more closely models the process of natural selection?)nephmon
September 4, 2009
September
09
Sep
4
04
2009
02:46 PM
2
02
46
PM
PDT
Hark! I have been accused of being a "Darwinist"; a baseless and unfounded slur for which not a shred of evidence has been provided. Shall an apology be forthcoming? Moving on. KF offers examples of latching and claims this as implicit latching, but is unable to offer a clear definition of implicit latching that I can detect (1). A demonstration of implicit latching is not a definition, and a claim that it is would be circular logic. I suggest this definition of implicit latching: Implicit latching occurs when the probability of advancing up the search gradient is strictly greater than the probability of regressing down the gradient. This definition should enable the behavior which you call implicit latching. It is also a fundamental property of all search algorithms (2). As such, implicit latching is a trivial property of all searches (2), and a useless definition. KF: What is your definition of implicit latching? Please explain how latching (explicit, implicit, quasi, any form) is not a property of all search algorithms (2). (1) I could have missed it, as there was an awful lot to sort through. (2) All search algorithms other than blind random search. I'm tired of typing this bit over and over.Tomato Addict
September 4, 2009
September
09
Sep
4
04
2009
05:52 AM
5
05
52
AM
PDT
You are heading a bit off-topic, Kairosfocus, but thanks for some great links. Two minor quibbles: You said:
(Contrast that with someone above who tried to drive a rhetorical wedge between ratcheting and latching.
Now, if you were referring to me, then you are just making stuff up. So I am guessing that you were referring to BillB at 270 (citations should be specific and appropriate, btw)
Joseph: Latching individual letters is a partitioned search, ratcheting towards a target is only a partitioned search if individual letters are locked out of the search when they reach their target. Partitioning means to divide up the search into separate, independent units. WEASEL is not an algorithm that searches for individual letters, it is not partitioned.
But BillB's first paragraph was a response to Joseph's Humpty-Dumpty attempts to re-define the word "cumulative", so he is trying to employ Joseph's lexicon. I would be happy if we could agree to use one of the "latching"/"ratcheting" pair to refer to mechanisms, and the other to refer to behaviors, but when I tried to make this behavior/mechanism distinction, you promptly re-defined the word "mechanism". Rhetorical wedge, indeed. BillB's second paragraph sums up the entire issue at hand in only 25 words. Brevity is the soul of wit. Ironically D&M discuss a Weasel algorithm (in a rather hand-waving manner) in Section F of their paper. Ooops. Second quibble: I have a better FSCI link for you. Tons of stuff there, actually.DNA_Jock
September 4, 2009
September
09
Sep
4
04
2009
05:21 AM
5
05
21
AM
PDT
--kf, I understand your last post that you see no difference in the design of the following ten strings: 1. SCITAMROFN*IYRANOIEULOVE*SAM 2. SCITAMROFN*IYRANOITULOGE*SAM 3. ECITAMRI*N*IYZANOITULOVE*SAM 4. SCITAMROFN*IYRANOITUL*VE*SAM 5. SCITAMROFN*IYRANOITULOVE*SEM 6. SCITAMOOLNOIYRAMOITULOVE*SEM 7. SCITANROFN*IYYANOITULOVE*SAM 8. SCITIMROFN*JYRANOITULOVE*SAM 9. SCITAMROFN*ICRHNOITSLOWE*SAV 10. OOT*DENGISEDESEHT*ERA*NETSILDiEb
September 3, 2009
September
09
Sep
3
03
2009
11:06 PM
11
11
06
PM
PDT
kairosfocus @ 290: "6] DJM, 289: Question for KF: Have you EVER seen a complete printout of a run of Dawkin’s Weasel program? KF: Have you? " I'll take that as a "No" and ignore the 400 following words which don't answer my question. So you have never once seen a complete run of Dawkins Weasel program and consequently have no evidence to think that it lataches whatsoever beyond your own inability to comprehend the purpose of the program or the phenomena it illustrates. Yet you slag Dawkins and call him dishonest! Talk about "soaked in oil of ad hominem!"djmullen
September 3, 2009
September
09
Sep
3
03
2009
10:53 PM
10
10
53
PM
PDT
PPS: Implicit latching, as noted and liked ever so many times in recent days, has long since been DEMONSTRATED, in the context of ratcheting, cumulative progress to target of nonsense phrases. And so we have a viable mechanism fro explaining the evident cumulative progressing, ratcheting latching action from the Weasel 1986 showcased o/p's. Therefore we can nail the stake through the heart of the Darwinist attempt to dismiss latching, as a distraction from the root problem of fundamental dis-analogy between targeted search on mere proximity and the need to actually originate complex functional information by chance plus blind necessity; in the teeth of the massively substantiated -- and Internet full of FSCI is exhibit no 1 -- observation that FSCI is only credibly produced by intelligence. Which Darwinists, plainly, would rather not discuss.kairosfocus
September 3, 2009
September
09
Sep
3
03
2009
10:52 PM
10
10
52
PM
PDT
Onlookers: Look at the above. Do we really want to cede more and more institutional and policy power in science, education the courtroom and the parliament to men who behave like the above, or who tolerate and defend behaviour like this -- an excerpt from Mr Dawkins' latest book?
(For those who need a 101: planets orbiting the sun is a direct, current observation. The holocaust is copiously documented [cf the records of the trials of the major war criminals which should still be accessible at major libraries] eyewitness lifetime [my parents are still alive] history, denied only by Neo-Nazis and Islamist radicals or the like. In Europe people have been gaoled for holocaust denial. The -- unobserved, unrecorded by us -- alleged deep past origins of the cosmos, of life in it and of major body plans then man by claimed spontaneous forces tracing only to chance plus blind mechanical necessity is by contrast all too reminiscent of the tall tales of Baron Munchhausen pulling himself out of the primordial swamp by pulling on his bootstraps. When it is not playing at tyrannical [cf also here] neo-magisterium by the materialist high priesthood, as we have documented by Lewontin and now implemented by the US National Academy of Sciences and others.)
To be forewarned is to be forearmed. G'day. GEM of TKI PS: On matters of substance I have just one supplemental footnote. It so happens that the Dembski-Marks calculation of probabilities in the IEEE paper p 1055 is based on the pattern of latching of letters once they go correct. (Contrast that with someone above who tried to drive a rhetorical wedge between ratcheting and latching. Instead, we may easily see that implicit latching is demonstrated and that once such latching of generational champions exists as the weasel "nonsense phrases" ratchet their way through targetted, proximity -- not functionality -- based "cumulative selection" filtering, then the analysis applies.) With double force, we see that Weasel's headlined and trumpeted "success" of getting to target much faster than a random walk would, is based on the built in active information manifested in the targetting and the proximity-based hotter-colder filtering and promotion. As Dawkins admitted in 1986, Weasel builds in the required complex coded functional information to begin with [as the target], so it cannot reasonably be used as a good analogy of a process that is claimed to originate de novo complex functionally specific coded information. And to pretend otherwise is simply and plainly to reduce one's reasoning to question-begging absurdity.kairosfocus
September 3, 2009
September
09
Sep
3
03
2009
10:42 PM
10
10
42
PM
PDT
DNA_Jock, good observation @ 286. It also highlights the inaccuracy of referring to active info as a measure of problem-specific information. A search without replacement is based on the same problem-specific information as a search with replacement, and yet the active information measure of the former is higher. The active information measure is a function not only of the information available, but also of the way that information is used.R0b
September 3, 2009
September
09
Sep
3
03
2009
09:17 AM
9
09
17
AM
PDT
Kairosfocus, you are as enjoyable as ever.
2] DNA-J, 286: Page 1055 describes the mathematics of a partitioned search with sample replacement, but the authors then slip (without indicating to the reader that they have switched algorithms in midstream) into describing in words the behavior of a partitioned search without sample replacement on p 1056, the page I cited. The analysis relevant to Weasel is that on 1055 leading up to Eqn 22, as was pointed out correctively above. What is discussed later on 1056 is another case, which has been dragged across the discussion as a red herring.
Interesting. Please show me where D&M indicate that they are now talking about a different algorithm. Why do they include it in the paper, if it is such a red herring? Let's take a quick look at the original post that kf was "refuting":
DNA Jock at 269: Joseph, I understand that you are going for the second door: “A latching process is (necessarily) a partitioned search (as modeled in eqn22)” which is a much better choice than the third door: “Weasel is a partitioned search (as modeled in eqn22)”, which is blatantly wrong. The key aspect of the partitioned search is that it is a “divide and conquer” procedure(p1055), in which the search for each character is independent of the search for the other characters. Thus for partitioned search without sample replacement, the target can be found in N-1 queries (p1056) irrespective of the target length. There are whole categories of search algorithms that work by step-wise comparing a short search string (one or two letters) with the target (starting at one end and moving along until a match is found.) This information is then used to infer the FOO (p1054) for the target, and subsequent short search strings use the derived FOO. Useful if the FOO is unknown. It latches, but it ain’t a partitioned search
I am merely referring to a particular type of partitioned search, as described by D&M. I am not claiming that they think it describes Weasel (although they may well leave unsophisticated readers with that impression). I certainly don't think it describes Weasel. So where's the beef? Furthermore, given that kf had clearly read this post, and therefore been introduced to the idea of a algorithm that latches but is not a partitioned search (as described by eqn22), I find his continued efforts to assert "latches, therefore partitioned" hilarious. I point to his undistributed middle, thus:
All partitioned searches latch. Kf’s imagined Weasel latches. Therefore kf’s imagined Weasel is a partitioned search All ostriches have two legs Kf has two legs Therefore Kf is an ostrich. I don’t think so.
And his response is to merely repeat his fallacy:
Partitioned searches ratchet to the target and part of that ratcheting is latching of already successful letters – as has been plain from the very beginning of the discussion. Latching – implicit or explicit –is a reliable sign of ratcheting action and so of partitioning. (But then, Darwinists seem to have basic problems with inferring based on empirical signs.)
Maybe I was wrong about the bird.
As to the further reiterations of misdirecting, misrepresenting and mischaracterising, polarising arguments above, they have all been long since answered repeatedly and cogently in details above and in onward linked materials. (Darwinists often seem to think that by trumpeting specious claims loudly enough, long enough and long enough, they will prevail. All they succeed in doing here at UD is demonstrating the puerility of both rhetoric and attitude. But, sadly, in the wider public, such self-discrediting and destructive behaviour may not be recognised for what it is and where it too often leads if uncorrected and unchecked.)
Say no more. In closing, I would like to thank kf for demonstrating that Weasel performs much better than a random search, and that mutation and selection can, cumulatively, produce results that appear directed/designed. That was, after all, CRD’s only point. I would also like to thank him for highlighting the sloppy writing seen in section E of the D&M paper. If someone as smart as kf did not notice the elision from Partitioned to Deterministic searches on pages 1055 – 1056, there is little hope that typical readers will be able to spot that eqn22 CANNOT describe a search that has generational champions, and therefore it CANNOT describe Weasel (irrespective of pseudo-quasi-implicit-latching).DNA_Jock
September 3, 2009
September
09
Sep
3
03
2009
06:32 AM
6
06
32
AM
PDT
--kf you made me smile: Darwinists often seem to think that by trumpeting specious claims loudly enough, long enough and long enough, they will prevail. Long enough and long enough, that's us Darwinists. But seriously, I asked: Can you spot the difference in design in the following ten strings? Can you explain the difference? 1. SCITAMROFN*IYRANOIEULOVE*SAM 2. SCITAMROFN*IYRANOITULOGE*SAM 3. ECITAMRI*N*IYZANOITULOVE*SAM 4. SCITAMROFN*IYRANOITUL*VE*SAM 5. SCITAMROFN*IYRANOITULOVE*SEM 6. SCITAMOOLNOIYRAMOITULOVE*SEM 7. SCITANROFN*IYYANOITULOVE*SAM 8. SCITIMROFN*JYRANOITULOVE*SAM 9. SCITAMROFN*ICRHNOITSLOWE*SAV 10. OOT*DENGISEDESEHT*ERA*NETSILDiEb
September 3, 2009
September
09
Sep
3
03
2009
05:52 AM
5
05
52
AM
PDT
Onlookers: It is sadly plain that the assorted Darwinists above simply do not understand the lines they have now collectively crossed, Nor do they seem to understand that the real issue is now whether they are willing to acknowledge serious error and return to civil conduct. As to the further reiterations of misdirecting, misrepresenting and mischaracterising, polarising arguments above, they have all been long since answered repeatedly and cogently in details above and in onward linked materials. (Darwinists often seem to think that by trumpeting specious claims loudly enough, long enough and long enough, they will prevail. All they succeed in doing here at UD is demonstrating the puerility of both rhetoric and attitude. But, sadly, in the wider public, such self-discrediting and destructive behaviour may not be recognised for what it is and where it too often leads if uncorrected and unchecked.) I will simply briefly note on points for the sake of reminder and record, for those who came in late: 1] DNA-J, 285: We can agree that partitioned search, as accurately described by D&M in eqn22, latches. Unfortunately, you are using this fact to conclude that a (hypothetical) latching search is a partitioned search. Partitioned searches ratchet to the target and part of that ratcheting is latching of already successful letters – as has been plain from the very beginning of the discussion. Latching – implicit or explicit --is a reliable sign of ratcheting action and so of partitioning. (But then, Darwinists seem to have basic problems with inferring based on empirical signs.) 2] DNA-J, 286: Page 1055 describes the mathematics of a partitioned search with sample replacement, but the authors then slip (without indicating to the reader that they have switched algorithms in midstream) into describing in words the behavior of a partitioned search without sample replacement on p 1056, the page I cited. The analysis relevant to Weasel is that on 1055 leading up to Eqn 22, as was pointed out correctively above. What is discussed later on 1056 is another case, which has been dragged across the discussion as a red herring. 3] DNA-J, 287: kf whines . . . . kf is correct that I omitted the text that he bolded. But since I did not quote anything AFTER the Nobel Prize line, ellipsis seems unneccessary. As can be seen from what is now point 7 in 280 above, by omitting what followed, D-J materially misrepresented what I said, twisting my words into a blind appeal to authority, when in fact – as the suppressed example of the Steady State theory showed – I was pointing out that even the errors of someone like Hoyle are instructive. (One cannot simultaneously be talking about learning from the errors of great thinkers while appealing to blind adherence to them!) 4] I would like to thank kf for demonstrating that Weasel performs much better than a random search, and that mutation and selection can, cumulatively, produce results that appear directed/designed. That was, after all, CRD’s only point. D-J here neatly omits WHY Weasel outperforms random search: it is targetted search with an already built in target that uses warmer-colder proximity to target metrics to reward “nonsense” -- i.e. non-functional – phrases. In short, Weasel is an example of intelligent design and the power of active information. And, CRD's intent was plainly to give the impression that his 'cumulative selection” was a good analogy for random variation and natural selection based on differences in functional performance. Dawkins had to admit that Weasel is fundamentally dis-analogous, but seems to have relied on the difference in rhetorical impact of a spectacular simulation and qualifying weasel words that point out that there is more to the simulation than meets the eye. This is manipulation, not education. 5] Dieb on runs, generations and latching Implicit latching is of course demonstrated, and the abstract possibility of slippage – which has been discussed, is irrelevant to the fact that good “tuned” runs do show latching of the generational champions. That is all that is required to explain the apparent latching effect in the SHOWCASED 1986 runs. 6] DJM, 289: Question for KF: Have you EVER seen a complete printout of a run of Dawkin’s Weasel program? Have you? Have you got the code to demonstrate what the program actually did as opposed to inference to best explanation on its behaviour? More directly, the following is what I had to say, without clipping off conveniently:
3] Weasel also shows cumulative selection, which was showcased by a the “famous” 1986 runs that showcased how the progress to target in the sampled runs is without reversion. (This is consistent with the basic, plain garden untwisted meaning of “cumulative” [i.e. I here cite Dawkins -- KF]: Increasing or enlarging by successive addition. A sense that we have every reason to see from his descriptions and examples as the one intended by Mr Dawkins c 1986 in BW; latterday “revisionising” notwithstanding.) 4] two mechanisms have been shown — yes, SHOWN — capable of producing runs of generational champions that produce the same effect: (i) ratcheting action based on explicit latching of successful letters to date in Weasel, and (ii) ratcheting action produced by implicit latching due to interaction between selected population sizer per generation, mutation rate per letter and selection filter characteristics. 5] Mr Dawkins has more or less has said that the actual code for Weasel c 1986 is not forthcoming, but the various algorithms and programs currently on the web are good enough to replicate the essential action, whch seems to have been implicit not explicit. Now, it is Mr Dawkins who described Weasel as showing “cumulative selection,” showcasing runs in 1986 that show that for over 200 places where such reversion could presumably occur, it never happens once. (And that in a context where we easily see how incorrect letters often persist for decades of generations.) So, we have good reason to infer that for good runs, Weasel c 1986 did ratchet to target, and in so doing effectively locked up already correct letters in genrational champions. The only question on this secondary point is whether implicitly or explicitly. And BOTH mechanisms have been demonstrated. On Mr Dawkins statements as reported since about 2000, it is believed that Weasel 1986 latched letters in the generational champions implicitly. +++++++++ Again, we see the same saddening pattern of misdirection, misrepresentation and mischaracterisation. We have been warned on what is happening and what its consequences are likely to be if we let Darwinists get away with such uncivil and misleading or outright deceptive tactics. GEM of TKI
kairosfocus
September 3, 2009
September
09
Sep
3
03
2009
05:32 AM
5
05
32
AM
PDT
KF @ 279: "3] Weasel also shows cumulative selection, which was showcased by a the “famous” 1986 runs that showcased how the progress to target in the sampled runs is without reversion." Question for KF: Have you EVER seen a complete printout of a run of Dawkin's Weasel program?djmullen
September 2, 2009
September
09
Sep
2
02
2009
10:26 PM
10
10
26
PM
PDT
--kf Dembski and Marks start in their paper with the phrase: SCITAMROFN*IYRANOITULOVE*SAM I calculated a next generation using Dawkins's algorithms with populations of 10,50 and 100 - and mutation rates of .04, .05 and .1. The tenth string in the list is the second generation given in the paper of Mark and Dembski. The differences with the first generation are in bold face: 1. SCITAMROFN*IYRANOIEULOVE*SAM 2. SCITAMROFN*IYRANOITULOGE*SAM 3. ECITAMRI*N*IYZANOITULOVE*SAM 4. SCITAMROFN*IYRANOITUL*VE*SAM 5. SCITAMROFN*IYRANOITULOVE*SEM 6. SCITAMOOLNOIYRAMOITULOVE*SEM 7. SCITANROFN*IYYANOITULOVE*SAM 8. SCITIMROFN*JYRANOITULOVE*SAM 9. SCITAMROFN*ICRHNOITSLOWE*SAV 10. OOT*DENGISEDESEHT*ERA*NETSIL Can you spot a difference in the design of the strings? And for latching...DiEb
September 2, 2009
September
09
Sep
2
02
2009
11:22 AM
11
11
22
AM
PDT
1 2 3 4 5 13

Leave a Reply