a "review" dan rather would love

Corey Pein, writing for the Columbia Journalism Review, helps to burnish that organization's liberals credentials by writing commentary on the debunking of the CBS forged documents that Dan Rather would certainly love.

Evidently annoyed by the success of the blogspere in so quickly bringing down Rather's attack on the Bush reelection effort, Pein criticizes bloggers like Charles Johnson from starting from the assumption that the documents were forgeries.

But the power of Johnson's work was its simplicity. It used Microsoft word with default settings. Johnson merely typed in the text taken from one of the forgeries and came out with a result that was almost the same as the forged document itself. The simplify comparison, he created an animzated gif with the document he created overlaid on the forgery -- the evidence is there for all to see!

Pein's concerns, that it would be hard to make any conclusions about a document, claiming that " copies cannot be authenticated either way with absolute certainty" don't really apply as the match between the modern version and the forgery is so close. Had the forgery been more sophisticated, this argument might applied but the fact is that the forgery was a crude one. This kind of argument is no different from saying that fingerprints can't be successfully recovered from a certain surface in the face of recovery of prints that in fact match some known set. Anyway, for anyone who doubted the elegant simplicity of Johnson's work could reproduce it for themselves quite easily! Perhaps Pein could take some time and do just that.

But what kind of analysis would convince Pein. Evidently that conducted by David Hailey.

"In order to understand “Memogate,” you need to understand “Haileygate.” David Hailey, a Ph.D. who teaches tech writing at Utah State University — not a professional document examiner, but a former Army illustrator — studied the CBS memos. His typographic analysis found that, contrary to widespread assumptions, the document may have been typed. (He points out, meanwhile, that because the documents are typed does not necessarily mean they are genuine.) Someone found a draft of his work on a publicly accessible university Web site, and it wound up on a conservative blog, Wizbang"

What Hailey attempted to do was to determine if the fonts used in the documents could be closely matched to fonts found in an on-line repository searched using the term "typewriter". Finding a font that he liked, he proceeded to cut and paste the characters together (in Photoshop) to produce an approximation of one of the forgeries. He couldn't fully reproduce the character set, however, becuase of the "th" and a couple of other characters. But Hailey's work is not too helpful (except maybe for Rather and Mapes supporters) since it fails to consider the between character spacing as Johnson's work did. To take our fingerprint analogy a step further, what Hailey has done it to say there are so many sworls and so many whorls and you could then put these together in any number of different ways.

So Pein would have us believe that these documents were created in the early 70s, stored away somewhere other than the in the military files they presumably originated from, subject to numerous generations of photocopying and then, some thirty-some years later would somehow match almost exactly a easily-created modern day reproduction using inter-letter spacing technology that didn't even exist when the documents were presumably created? Well, Pein rejects the Johnson/Newcomer analysis because she believes they started out assuming that the documents were forged. Using her logic, no crime could be solved using fingerprints since the police have to start out assuming that they can match the prints against known prints in order to find the culprit.

Having rejected Johnson/Newcomer et al's forensic analysis, Pein turns to the crime-solving equivalent of eyewitness accounts and, sure enough, there are some contradictions. Someone she concludes that, while CBS was wrong to use "unverified" documents, the blogsphere was wrong to attack the network.

So, here we have a body and a knife with fingerprints matching a known criminal. What do we do Ms. Pein? Using the logic in this article, we would throw out the fingerprints, start looking for folking hanging around on the street at the time. One man saw a Hispanic about 5'8" while another saw a Caucasian about 6'. Maybe we could check the photograph file but, using Pein's philosophy, that would be assuming that the perpetrator was a previously identified criminal. And maybe it was a suicide after all -- time to send it to the cold case file.