Saturday, October 10, 2009

Source bluffing by Dembski and Marks

In their September 2009 article in IEEE SMC-A, Dembski and Marks cite a book outside the literature on evolutionary computation as their source for the (1,2)-EA [evolutionary algorithm] which they do not identify as a (1,2)-EA. I had guessed that they had googled to locate a source with a title that would not telegraph to editors and reviewers, "This is an evolutionary algorithm."

It turns out that my guess was not so bad. DiEb has located the book online. Furthermore, s/he has found that the evolutionary algorithm it presents is quite different from the one that Dembski and Marks analyze. Is it "by chance or by design" that Dembski and Marks neglect to specify where in the book the algorithm is presented?

DiEb slyly points out that Dembski and Marks demonstrate in another book-reference that they know to give the range of pages... though they give the wrong range.


  1. The reference section doesn't show the accuracy I normally expect from mathematical texts. The reviewers of the article should be disappointed by this: usually, they don't check the references, they assume that these are correct - it's just taken for granted.

    Of course, for a brilliant tract one could overlook what seems to be a nitpicking at details. But it's far from being brilliant...

  2. Errors in references are common in the engineering literature. I have read some reference lists when reviewing for journals, but probably not all. I catch many errors going from the citation in the text to the reference, and this is all I do with most conference papers. I'll be checking more closely when reviewing for journals.