For instance the Guide falsely asserts:
One easily testable form of CSI ["complex specified information," a.k.a. "specified complexity"] is irreducible complexity (IC)
But Dembski appeals to IC as a justification for assuming pure randomness in calculating the probability of CSI being generated by natural law. (See for instance the section "Calculation of specified complexity" in the wiki entry on "Specified complexity".)
If IC is merely a "form of CSI," then the whole basis of Dembski's probability calculations would be grounded in circular reasoning. Granted, this would only add one more to it's other already debilitating flaws, but it's still pretty stupid (and funny) to find this in a "Guide" to ID published by the Discovery Institute itself!
Cf. The Dover Decision ...
Third, Cleaver voted for the curriculum change despite the teachers' objections, based upon assurances from Bonsell. (32:23-25 (Cleaver)). Cleaver admittedly knew nothing about ID, including the words comprising the phrase, as she consistently referred to ID as "intelligence design" throughout her testimony. ... In addition, Superintendent Nilsen's entire understanding of ID was that "evolution has a design." (26:49-50 (Nilsen)).
In Dembski’s own writings, should you choose to examine them, Dembski calls IC a “special case” of CSI and never uses it as a justification for CSI. Wiki is in general predisposed against ID. Though it is generally accurate and a good resource on many things, this is not one of them.
I think I agree with you. IC's relationship with CSI is indirect at best.
Wiki, though, has a noted bias against ID or any criticism of evolution.