The Barton letters have occasioned an active debate in blogworld. Here are some links to some of the more active discussions. I’ll add others as they come to my attention as well as some blog comments without chats. Some of the links have a list of outbound topical links as well.
Many posters do not distinguish between the PC codes for tree ring which are on Mann’s FTP site and the code for the rest of the calculation, which Mann has refused to provide. We are obviously aware of the code on the site, since we published an article discussing it and specifically cited the URL. I’ve made this distinction on several occasions in very specific terms, but many blog posters have failed to make this distinction. As far as I’m concerned, the problems with the code which we’ve been able to examine demonstrate that the rest of the code needs to be examined as well. Tim Lambert has written in to say that, even though Mann has not provided his source code, Mann has provided his "algorithm".
Why should anyone be content with Mann’s verbal summary of his methods, when the verbal summary has been demonstrated to be unreliable when compared to actual code in the case of the tree ring PC calculations, when series said to have been used have not been used and when undisclosed extrapolations material to 15th century results have been misrepresented? Lambert’s position is contradicted by Mann’s own explicit statement to the Wall Street Journal that he would not be "intimidated" into releasing his algorithm.
Lambert’s explanation for this (see posting below) is that the Wall Street Journal either misquoted Mann or that Mann got mixed up. I will bet Lambert that Regalado did not misquote Mann on such a sensitive point. To give an example of the degree of Regalado’s fact-checking, Regalado cross-checked with McKitrick that the picture that I sent him was actually a picture of me. (Update: It may amuse readers to contrast a comment of Lambert’s at this blog with his posts here. At this blog, he says:
I have never said that Mann has released all of his code. He has, however, released the data, the algorithm, and some of the code. Perhaps McIntyre is unable to fathom the disticntion between “code” and “algorithm”.
When asked about this in post #53 below, his exact words were:
“code” and “algorithm” mean different things. An algorithm is a method for doing a computation, while code is an implementation of an algorithm in some programming language. The WSJ either misquoted Mann, took something out of context, or Mann mixed the two things up.
Myself, I think that Lambert is trying to make a distinction without a difference. But did anyone see Lambert say:
"Perhaps Mann is unable to fathom the distinction between “code” and “algorithm.
I didn’t think so.
Also one more time, Wahl and Ammann have not replicated anything that we had not already done. They have emulated Mann and their emulation of the RPCs using same assumptions was identical to mine to 10 decimal places, as I reported here previously. Wahl and Ammann do NOT replicate many aspects of MBH98 calculations and do NOT report on any of the specific issues noted in the Barton letters. Most posters don’t seem to understand the difference between publishing an article in a journal and highlighting your own journal articles in your capacity as an IPCC author. Vranes gets close to this point, but doesn’t follow it through to its logical conclusion. Anyway here’s some discussions – some friendly, some unfriendly.
JREF: Science forum JREF: Politics forum
UK Weatherworld discussion group
Debunkers Kevin Vranes at Prometheus Outside the Beltway
David Appell 1
David Appell 2
Tim Lambert 1
Tim Lambert 2
escribe discussion group
[note: edited slightly on July 6, 2005 and again on July 24, 2005]