FYI, Bill, I'm going to redo the update as a separate post. I also forwarded the information to Charles Johnson and he says that it's correct. MS Word does "semi-kern" its text.

I think that the only way to lock this down is to get a sworn analysis by a panel of forensic experts. Seriously.

And then I'd like to lock down the originator of those documents. What?--I just thought we could ask 'em a few questions, is all . . . .

I agree, get some analysts the actual document, not a worked over scan, and check it for every kind of difference. Clearly the line of certainties keeps shortening (spacing, superscript, etc.).

Then there is the question of a verified signature and Hodge's comment.

If CBS did not present the comments of the family, that is pretty lame and insensitive. But from a forensic point of view, workplace colleagues directly connected are credible. Family opinions are not.

But if there are folks out there claiming to have seen Bush, and angry they are not heard, why on earth have they been silent? Why hasn't someone contacted them or they come forward? I am not saying they don't exist, I am saying that is weird isn't it? As many times as this has come up?

So it is awful for Killian's family to be dragged around in this, I would absolutely hate it, but evidence-wise, I am not so sure their opinion has decisive weight.

Dammit, you mean I was right and then I retracted my correct statement?

I think we need a link or (harkening back to the Stone Age) a bibliographical reference to that kerning info if it's true, or at least an informed opinion from an acknowledged expert. Otherwise the issue is too murky.

Now I know how the White House must have felt when Joe Wilson was debunked re his challenge of those 16 words in the State of the Union regarding uranium in Africa: D'oh.

One more thought, what seems the quickest way to the barn on this one, is that CBS seems to be claiming they have originals, not copies. That was always my impression , but they never said so until today.

Anyhow, not a specialist here, but typewriters leave imprints when the key strikes and jet printers don't. The original, if it is the original, will have impressions only a typewriter could make. And the ink would be verifiable .

So why not end all the typographimania and just have some forensic types look at the paper and the ink.

Doesn't that just seem easier and if CBS is right, the fastest way to have their journalistic name cleared?

I'm not certain that CBS has the originals, actually, which makes the process of verification more difficult. I don't think the lack of an original in itself is pro-forgery evidence, though, since (if the documents are in fact real) whoever leaked them would probably not want to remove the originals from wherever they found them.

If CBS does have the originals, then, yeah, it should be easy to determine if they were created with a typewriter.

Someguy -

As far as I am aware, CBS does NOT have the originals.

And what documents they DO have, they have not agreed to release to independent analysts.

The first step is to get a panel of experts to agree on a verdict based on the visual characteristics of the copies. It may not be 100%, but the resulting statement would pressure CBS to release the docs.

Is it too late to trademark the phrase "semi-automatic kerning?"

A more animated version than Charles (but I emailed it to him and have it in high res)

http://img47.exs.cx/img47/56/animated_aug18.gif

Can you do a larger, more distinct version and state your methodology?

OK, I hate to argue against this but I don't think there is semi-automatic kerning in Word, at least for the letter "j".

Lowercase j definitely protrudes into the previous space. But I don't think you can call it "kerning" because it does this for ALL prior letters; there's no special rule for different positioning beside different prior letters based on how they fit together. It looks like they solve this without kerning by not allowing any other letters to have ink in that portion of the cell-block.

Type "To mj gj qj" at 500% zoom, then move the cursor between them, and you'll see what I mean. The last two are letters that would have ink at the farthest bottom right positions where kerning would not be allowed because they might run together. I don't see a spacing difference when I move the cursor though -- and I do see one in "To" when I turn "kerning for font" on and off and move the cursor there, or put an "L" next to the "T".

Sure...I'll email you offline but my methodology was the same as Charles' latest. Took his word doc, loaded into WordXP, printed through Acrobat so it rendered straight to PDF. Loaded those suckers into Adobe After Effects, mapped one to blue the other to red. Made only horizontal and vertical scalings of the "new" document layer (as in manipulating the whole page at one). Set the key frames and hit render. I have it as a master document in the same resolution as the original CBS PDF.

Whatever. I'm canning the amateur research in favor of an attempt to get a consensus statement from experts.

I look at the animated GIF posted. My question is if the resizing of the document to make an overlay somehow messes up the integrity of the document. Seems that someone (for arguments sake) could just say "of course they fit perfectly if you resize it and make it so". Just curious.

LoydRight -

I'm not basing the definitive argument on the overlay, but even if the sizing angle that you mention were true, the idea that the spaces between words, letters and lines of text would match up to scale is almost ludicrous.

Actually, I thought Dan Rather said tonight that CBS had copies of the originals.

The way to settle this is to let everyone see the originals.

Let's see the originals!!!

I think I understand why Rather stood behind the story - he REALLY BELIEVES he has genuine documents that incriminate Bush and that all this is just flak from wishful-thinking Bush supporters.

I actually feel kind of sorry for the guy. This is going to be a real gut-twister for him when they are definitively proved to be forgeries by the experts.

My boyfriend thought of an explanation that I don't believe I've seen yet.

If you scanned a document into text recognition software, and then printed it out, it would be formatted using whatever the default is for the software (e.g. Word).

Now, I have no clue why you would scan the document like this, rather than as just an image, unless you had many such scanned documents, and you weren't sure which ones you wanted, and so had to have them all searchable by text content.

Then, you printed it out and realized no one would ever take this nice, sharp document for a thirty-year-old one, and somebody said, "Hey, just Xerox it a thousand times. No one will ever know the difference."

I think what happened here is clear: CBS was given access to an archive of Presidential material, collected by a time-travelling civilization of the far future (who, being time-travellers, can reach into private files and copy documents without anyone knowing). This civilization stores their information as ascii text rather than images (kind of remiss of them, in my opinion), and one of them traded the information for something unavailable in the grim but highly advanced world of the future (my guess: Moon Pies).

It's Occams Razor Stupid!

(I trust there'll be no sniffing from you, Bill, about "random speculation".)

http://img20.exs.cx/img20/5/animated_aug18bigger2.gif

Here's bigger. You are correct, it would be nearly impossible for the proportions to match when you scale something as a whole image. That's what I was trying to show, that no one is messing with word tracking or letter kerning. The date is warped out from a lot of copying, as it is the fatherest point from the center of the imaging device.

Some Guy,

I am not a lawyer. Take the following with a HUGE grain of salt. But...

A lot of CBS's case rests on people saying, "Yeah, that sounds like something he would say." Other people saying, "No, that doesn't," sounds important to me. So that then opens the question: who would know the man's mind better? People who worked with him 30 years ago? Or the family he left behind 20 years ago? I don't think there's a clear answer there.

More important, to my mind, is the fact that the family explicitly AND implicitly cast doubt on the authenticity of the documents. Explicitly, because they say he never typed, and because they say the phrasing and message aren't consistent with the man they knew. Implicit, because they say the memos never came from them, so that raises the question of where the memos DID come from. CBS isn't saying, so that creates a vacuum. And that vacuum makes room for suspicion that the memos came from -- a forger!

CBS has the power to settle this: identify their source and let a panel of experts examine their documents. But they DO NOT have the power to declare it settled when it isn't.

Thanks Brad.

And Angie, I think you're joking, but I'm so bleary-eyed, I'm not quite sure anymore.

Even though Microsoft Word doesn't turn on kerning by default, there are still some kerning pairs evident in the CBS documents, all having to do with the lower-case letter "f." Look at the words "from," "flight," and "officer." In all cases the top part of the "f" overlaps the next letter.

This isn't quite "kerning" as such; it's more of a pseudo-ligature. Nonetheless, the essential point remains the same: there is NO WAY that a 1970s typewriter would be able to do this.

Dark they were, and bleary-eyed.

I'm not entirely joking about the text-recognition software. Assuming the documents were genuine, this would explain their formatting.

But, time-travellers aside, who would have such an archive? My boyfriend thinks it's impossible to over-estimate the desire of the US military to collect, classify, file, track, and just generally snoop. I think he's wrong.

In other words, I think they look forged, too, but this was technically a possibility, so I thought I'd mention it.

I got carried away about the Moon Pies. I'm sorry.

Angie, why would they stil be signed?

Just saw Jonathan Klein debate Stephen Hayes on Fox news. Sadly, the broadcast still lags well behind the blogs on the details of why they may be forgeries.

The Fox guy asked Hayes if bloggers were driving the story development and Hayes began to say, essentially, yes there are raising important questions based on deep knowledge -- when Klein interrupted and said that bloggers are a bunch of guys in their pajamas making stuff up!!! So funny, but so sad.

Bill - We NEED typographic experts to evaluate this...

To expand on Christopher Hlatky's post above: Go to the August 18th PDF from CBS. Zoom in on the word "feedback". It sure looks to me like the top of the f overhangs the e, but I would want an expert to speak to that. Now, look at the word "CYA". Not kerned at all. But come on people, "YA" is a really easy kern-pair. And look: it's NOT kerned! So whatever printed this document must kern "fe" and NOT kern "YA".

Does that narrow down the options enough?

Angie, why would they stil be signed?

Well, assuming the time-tra--er, the military--did not keep the images, but only the ascii, the signatures would have to be forged by someone who realized belatedly that they'd look a lot more convincing with signatures.

It's like one o' them murder mysteries where the murderer tries to make it look like suicide and puts the gun in the (left-handed) victim's right hand.

Look guys, it's going to be as simple as this: A pair of FBI agents are going to knock on Dan Rather's door with a suopena for all documents related to this matter so their own forensic experts can examine them. Y? Simple: IT IS A FEDERAL CRIME TO FORGE OFFICIAL US GOVERNMENT DOCUMENTS.

There is sufficient evidence that documents SUPPOSEDLY sent to Lt. Bush in an official capacity were in fact forged. That's all the Justice Department needs to begin an investigation. And if they don't, don't think CONGRESS won't, as early as next week.

At this point a good old fashioned letter writing campaign might be in order, to your congresscritters to demand that they look into it. Some of you must have Republicans representing your district; they'll pay attention.

Here's a question: Is there any way to find out the budgetary outlays for the TANG during 72/73 ? Like what kind of office equipment they purchased ? What did they have on hand ? There must be a paper trail somewhere ...

Chief, you're getting all wrapped around the axle here. Kerning is kerning, semi- or not. Typewriters at Air National Guard offices in Pasadena, TX in 1973 did not do it. Someone's gonna have to come up with the superest government typewriter ever, in the history of ever, to prove that it could do those damn dumb memos.

Never mind that poor old Buck Staudt had retired from this onerous duty a year before he pressured Hodges. Yeesh. Big red flag, anyone?

In Times, the lowercase 'f' overhangs into the next letter all the time. That's not kerning, and it is, in fact, possible on a 1970's typewriter. Look at the type samples in this 1953 ad for an IBM Executive typewriter:

1953 ad

Note in the word "straightforward" how the f overhangs the o.

Or look at this old "Nibble" ad done with an IBM Executive:

Nibble ad

The Nibble ad is clearly not the same typeface as the Killian memos, and the f's don't overhang as much, but note that it has the bumpy baseline of the Killian memos. Note the way the tail on the y's stick a little way into the previous character's space, as do the tail on the j. Look specifically at the word 'enjoy' at the beginning of the 10th line from the bottom.

I'm not at all saying that the memos were definitely done on an IBM Executive. I just haven't seen anything in the memos that an Executive couldn't have done, including the superscript.

The Executive wasn't a wacky expensive typesetting machine. It was a business typewriter. According to the ad, you could buy it with 12 different type styles. And to my eye "IBM Bold Face Type" is broadly similar to Times Roman Bold.

Although i'm all the way at the bottom of this page Your Gif is a joke. The number two starts on bottom right and moves at a nearly 45 degree up and to the left angle the middle of the document goes straight up and the right side moves up and two the left slightly.... Fakeries all around?

Um, it moves cause I sperated the two layers in the x and y dimension to show them land on top of each other. The top layer as a whole scales around 92% in the horizontal and 90% in the vertical, no individual characters move on their own, the whole layer moves as a unit. This would be consistent with the scaling a copy machine inadvertenly does (go to the library, run some pages through a couple times and compare with the original, you'll see the reproduction isn't a 1:1 perfecly scaled version, in fact you'll start to see warping at the periphery too, oh and you'll see serifs start to blur). I'd love to know how all the character and font proportions stay the same if the original wasn't done in Word as well. I'm going to do all 4 and try to post the examples and source documents (including After Effect project files) somewhere.

Word kerns TrueType fonts automatically, because pair kerning information is built into the font. What pairs are kerned depends on the font and what the font designer puts in, obviously. That option doesn't affect it.

PostScript fonts have sophisticated user-editable tables as separate files for use in kerning. It's not built directly into the file. These effects can be turned on and off, and that's what the Word option is more, the more sophisticated kerning used in professional situations.

Simply put:

1 - Kerning-related technology was nothing new at all in printing, and was old news in typewriters as well.

2 - By the 1950's business typewriters by IBM were already using it, and bragging about it as a big selling point. IBM sold a ridiculous number of typewriters, and a lot of people have an old kerning IBM in their basement someplace to this day.

3 - During the 20 some odd years that followed IBM's introduction of kerning tech, I must suppose that other companies began to offer their own versions of IBM's "trick". That's the beauty of competition.

4 - To say that Killian "couldn't have had" access to a typewriter, from the most WILDLY popular maker is crazy talk. Bets are pretty good that he probably had an IBM, though he might've had another model that utilized the "trick" of kerning.

5 - To say that his typewriter in the 1970s was older than a 1950s model is not out of the question, but come on now. Lets not run around presupposing that he used a 40 or 50 year old typewriter, just so we can try so hard to "smell a rat". That is a case of trying to fit your facts to a desired conclusion. Something called biased investigating.

We can see a popular model of typewriter, from probably the biggest maker of that time. That manufacturer introduced this technology 20 years prior to the letters. One must suppose that competitors "kept up", and knowing IBM, they sold like hotcakes, and they and their competitors likely improved it steadily and incorporated it more and more into future models and series'. The odds that Killian had a kerning typewriter are VERY VERY good odds.

Folks are starting to sound kind of "politics driven" here.

That, or completely losing perspective of time, like people who think the internet "started" in the 1980s (more accurate to say the 1960s, though its a little debatable), or that color movies started in the 1950s (try around the 1930s for feature films).

Kerning was mainstream for decades.

This, to me, is now a dead issue.