Goffr
Posts: 14
Joined: Feb. 2007
|
As per the UD thread, from the Demski paper:
| Quote | 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 characters 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 characters are ratcheted into place. The 19 new letters are chosen, and the process is repeated until the entire target phrase is found. |
So Demski's program jumps from:
SCITAMROFN?IYRANOITULOVE?SAM
to
OOT?DENGISEDESEHT?ERA?NETSIL
in 1 generation. There is only 1 child, and it in no way represents the parent sentence.
In my understanding of Dawkins' program, there is no ratcheting of each correct charater, nor does it randomly select new characters for every incorrect character. In Dawkins' program there is a only a low chance of each letter mutating (determinded by the search parameters), so nearly all stay the same and the children represent the parent almost exactly. Each and every letter may mutate, there is no latching in this model. There are multiple children each generation (each with their own mutations) and the one that matches most closely the target phrase is selected.
For example. The parent phrase:
SCITAMROFN?IYRANOITULOVE?SAM
will have 'x' number of children (determined by the search parameters, here 8):
SCITAMREFN?IYRANOITULOVE?CAM
SCITFMROFN?IYRANOITULOVE?SVM
SCITAMROFN?IYRANOITULOIE?SAM
SCITAMROFN?IYRANOITGGOVW?SAM
SEITAMROFN?IYRANOITULOVE?SAM
SCITAWEOFN?IYRANOITULOVE?SCM
SRITAMROFN?IYRANOITUROVQ?SAM
SCITAMROFN?IYRANOITULOQE?SAP
The sentence that most resembles the target sentence will be chosen for the next generation. Here it would be the 5th child, as it has 3 matching letters.
SEITAMROFN?IYRANOITULOVE?SAM
Compare Demski's 1st & 2nd generation vs the Dawkins model:
Dembski:
SCITAMROFN?IYRANOITULOVE?SAM
OOT?DENGISEDESEHT?ERA?NETSIL
Dawkins:
SCITAMROFN?IYRANOITULOVE?SAM
SEITAMROFN?IYRANOITULOVE?SAM
In Dawkins model the Parent and Child remain almost the same, apart from minor mutations. Demski's changes completely. Obviously there's been a major screw up in Dembski's understanding of how the Weasel program is supposed to work, and for some reason no-one on the ID side can see this.
I don't get how they don't it. How can they be so completely wrong on this, and remain wrong on it for so many years?
|
