It’s disappointing that the Bush National Guard memos appear definitively to be fake. However, on the plus side, it’s good to know that all the criticisms raised by lunatics like Charles Johnson have also been definitely proven false (also here).
As a bit of a typography geek, I was underwhelmed by Johnson’s notorious animated GIF. It showed that two documents typed in the same typeface using the same spacing and then shrunk to a very low resolution looked quite similar; this barely even *implies* anything about authenticity, never mind proving it.
Based on a similar gut feeling, Hunter at Kos (linked above) investigates all the blogosphere criticisms of the documents: it’s clear that they could reasonably have been created in Lt Col Killian’s office in 1972. The key evidence that they weren’t comes from the WaPo interview with his secretary: she says that in real life, she didn’t use the kind of typewriter that would produce such a document. Even though they existed, and not just as enormous print-shop typewriters.
As Matt Yglesias says, the rightybloggers have no right to be at all triumphalist about their achievements here. They’re just as wrong as Dan Rather was (just as incorrect in this case, I mean; obviously the ‘we must slay all the Muslims’ crowd are far more *wrong* than Mr Rather…)
Update: the Johnson document doesn’t work when blown up big at all – there are some really obvious differences (from Juliusblog)
Update 2: heard an interesting rumour that Microsoft deliberately modeled Word’s type spacing on an IBM Executive, so that major corporates who switched to word processing could keep a consistent corporate image. Can’t find any credible sources for this as yet – any confirmation or denial would be appreciated. Incidentally, I really don’t care whether or not the documents are real; the issue is more one of not letting liars like Johnson pretend they were right.
Uh, there is a another website (lgf links to it) where the documents are not shrunk really small, but on the contrary, enlarged, and they match very well.
Also, why has not someone been able to demonstrate such a close matching for the Killian memos and a document produced on the kind of typewriter people suggest produced these documents? I have not seen that anywhere. There is something special about the Microsoft Word match-up that has not been duplicated.
This whole thing is idiotic. About the only thing we’ve learned is that many web loggers aren’t very good at using Microsoft Word. Good for them, but they oughtn’t to think that they’re making a serious forensic point.
John,
You really should have looked at LGF’s subsequent articles.
The Kos article mentions that a font consists of more than just the typeface; it also includes rules about spacing, etc.
Unfortunately for Kos, these rules vary in their implementation – not just between modern day word processors and 1970’s typewriters, but also between the same font on the same word processor, printed on different printers (trust me, this has on occasion been the bane of my life!)
Presumably, LGF and the unknown forger use a similar type of printer.
Many of his other points (eg. apparent differences in font, letters appearing to be out of alignment) are artifacts of multi-generational copying (I’ve seen worse).
The point about superscripting – a fair number of typewriters would have allowed sub/superscripting, but it would have been at normal typesize, not a reduced typesize. It could have been achieved by a special character (as he suggests), but would they really have gone to all the trouble of specially ordering a customised <sup>th</sup> character? Even more improbable, what are the chances of it exactly matching the same effect in MSWord’s default settings?
In summary, Kos’s criticisms definitively prove nothing.
Definitively proven was a bit strong.
Still, if I were ordering typewriters for the US armed forces, customised superscript "th", "st" and "nd" characters would certainly have been on my customisation list, for relatively obvious reasons.
If you look at the Juliusblog GIF above, it’s clear that LGF is using a different *typeface* variant from the letter – not only the spacing, but actual characters, are different.
Finally, here’s a third Kos story, which looks directly at font samples.
Customised superscript "th", "st" and "nd" characters wouldn’t be on my list at all; it would be a lot quicker for them to type (eg) "st" than to fiddle about finding the special key for it. Also a lot cheaper.
The Juliusblog image is a blown up view of the online image. The differences could be caused by different typefaces, or by "noise" creeping in through multiple copying. Viewing the originals (or 1st generation copies) should be enough to settle the argument, one way or another.
The Kos story didn’t look directly at font samples, but linked to another site which did – and drew exactly opposite conclusions!
Uh, geez, you guys. If you think something other than Word could have produced these documents, then let’s see it. Let’s see a document that matches up as well as the little green footballs examples when overlayed.
It has not been done!
The issue is not so much the superscripts specifically, or even the actual font typeface, but the perfect match (accounting for multiple copying and other degrading) with the Microsoft word. Even the last letters on each line and the last letters on the document are still in alignment.
If something else could have produced these documents with such as good match, then why is there not one single example of this anywhere?