Who will buck up and buy a IBM Composer circa 1972? I am sure the 'sphere would help compenstate the volunteer.

Did you see Donald Sensing's Post on the IBM Composer:

It is 100% mechanical and has no digital electronics. Since it has no memory, the user was required to type everything twice. [boldface added]

"That right there stops the whole defense of the documents because there is no way that such a machine would be used to type pedantic Air Force office documents that were intended only for filing. I know that the military has a reputation for needless bureaucratic makework, but really, that's ridiculous beyond the pale." D. Sensing

Outstanding work, Bill! In his understated way, on purely typographical gounds, Dr. Bouffard has completely undercut whatever explanation Dan Rather will come up with tonight!

New Slogan: "It's Occam's Razor, Stupid!"

guest - See update.

Oh well according to RatherBiased he's going ot prove that superscript could be done on a typewriter as early as 1968. Whether he explains how character kerning hint tables could be conatined on a mechanical device and how it just so happens all the line breaks are consistent with MS Word is another story. I mean these people (Rather and his useful idiots at Kos and Atrios) really take us for fools don't they?

The reference to the IBM composer is interesting, in that one of the commentators there mentioned the IBM Executive (Models A through C -- C in 1959) which were pre-selectric (i.e. traditional strike keys) proportionally spaced typewriters. Expensive and a little tricky to use, they were not unkown and the ANG could easily have gotten some general's or executive's hand-me-down when selectrics were bought.

I used an Executive in 1973 and it did produce nice proportionally spaced text. Especially the C model which used a film ribbon. Those ribbon reels were a cranky and could get tangled easily, though.

What I don't recall (this is more than 30 years ago) is whether it had a superscript function. I don't think it had curved apostrophes or quote marks, and I'm pretty sure it didn't have kerning (in fairness I didn't know what kerning was in 1970).

Someone ought to be able to obtain a Model C IBM Executive and see if it would produce the type on the memo! I've been mentioning this for two days (including on the original Free Republic thread!) and no one seems to be following up on this. I'm sure Bouffard knows the capabilities of the Executive, but you should have him specifically compare the memos to all three models of Executive with all available typefaces.

Keep up the great work!

Not quite over yet. Seems the IBM Executive, not the composer, may be the culprit. This was posted on Washington Monthly - "Kevin, I worked in the IBM Office Products Division field service area fixing typewriters in NYC for over 13 years in the 70s. I can tell you that the Model D can produce those documents, not only did it do proportional spacing, you could order any font that IBM produced AND order keys that had the aftmentioned superscripted "th." Also you could order the platen, thats the roller that grabs the paper, in a 54 tooth configuration that produced space, space and a half and double spacing on the line indexing, this BTW was popular in legal offices. The Model D had to be ordered from a IBM salesmen and was not something that was a off the shelf item, typical delivery time were 4-6 weeks. Also, typewriter keys were changed in the field all the time, its not that hard to do. I wish I had saved my service and parts replacement manuals to backup this claim but I'm guessing a call to IBM with a request for a copy of their font and parts replacement manuals would put this to rest ASAP. Posted by: BillG NYC on September 10, 2004 at 12:24 PM| PERMALINK"

Additionally, since everyone is such an authority on typesetting these days, how do you explain the unevenness of the letters, with some sitting slightly above the baseline? Word processors don't do that. Also, look at the 7's versus Times Roman in Word. There is a little hitch on the top of the Word version that does not seem to be there on the documents.

Until someone produces positive proof that you can produce these memos on a period typewriter, it won't settle the question, but rushing along isn't wise either.

And besides, "expert" does not mean infallible.

And by the way, Word can be used to simulate any number of looks, that is its virtue. That one can mimic an older document is not proof an older document is therefore actually new.

It's Occam's Razor Stupid.

Some Guy -

Dr. Bouffard ran teh IBM Executive, and the possible fonts DO NOT MATCH CBS's memo.

Bucko: I have located a working vintage IBM Selectric Composer. Further details as they become available to me.

Bill, the "type it twice" rule only applied to fully justified type. Type that was set ragged-right, like the type seen in the memos, could be set by just typing it once.

Some Guy, here's a sample of the output of an IBM Executive. Doesn't look anything like the memos in question.

Also, we're not talking about using Word to "simulate" anything. We're talking about firing it up using the factory default settings and typing away.

First, I don't know and am not wedded to the idea it is not a fraud. I just think all the sense of certainty is a tad premature.

Jeff, I don't know how you can read that image clearly. It is small and blowing it up will create pixel smugde which is a critical problem with all of this.

Second, the comparison linked to an Executive on Allah is not definitively different given the poor quality of the .pdfs.

Third, there were four models of Executive, did the good Dr. check the D? A guy who worked on them for years, actual lived experience with the machine, thinks it is capable. That is not insignificant.

Finally, there is, even in its smudged state, irregularities in the level of the letters and I repeat, word processors don't do that. Did the Dr. explain how those subtle ups and downs of letters got there from MS Word?

May well be a hoax, but there are still lots of problems with this.

Some Guy,

And besides, "expert" does not mean infallible.

Good advice, and we all must remember it.

But... 6 named experts on the record plus at least one unnamed expert vs. 0 named experts and an unspecified number of unnamed experts means the burden of proof has shifted to CBS. See my post here: http://www.blogoram.com/000647.php#000647. CBS can clear this up by naming their experts, letting the experts make the case, and letting the people decide who is credible. Until they do that (and they might be doing exactly that as I type, since I don't have access to CBS News feed right now), the evidence is clearly against them. Every blogger with a wild theory both pro and con means nothing. (Sorry, Bill!) 6+ named, credentialed experts that all reach the same conclusion means a lot.

Some Guy -

I saw a recreated docment today (can't remember site) where someone created a forged word processor document with greater (and more realistic) smudging than the CBS docs. Yes. it's possible.

Does the "D" kern? Does it have computer memory? That's the pivotal question, even ignoring the other possibilities.

Is it possible that Dan Rather can fool enough believers to let this fly?

The SMOKING GUN is the 100% EXACT MATCH between the August 1973 memo and the same as typed in default installation of Microsoft Word.

When the two documents are overlayed digitally, there is nary a pixel out of line. Why aren't these idiots going back to their "journalistic" desks and trying the experiment themselves!?

I wrote Dan a memo: http://www.americanlibertyjournal.com/archives/000264.html

If Rather comes out tonight and says this, this better be the end of his LYING ass. THE END, I tell ya! The end!

UML Guy -

"Every blogger with a wild theory both pro and con means nothing. "

That's exactly what I have been avoiding (like many of my peers). I'm conducting the interviews with the experts, not drumming up harebrained theories. I've drawn some conclusions, but please don't insult me by placing me in that category.

I am pretty accurately doing the job of a reporter. Interpreting facts as presented by experts with greater knowledge than myself.

UML Guy, you are dead right, the burden is on CBS. They need to produce a machine that can crank out one of those memos. Only way to be certain.

But my problem with the rush to conspiracy is exactly Occam's Razor: the simplest explanation is probably the best, and a forgery, a bad one if all this is to be believed, floating by CBS? And then to have corroboration from Hodge saying yes they did talk about that stuff? The numbers of variables to explain a forgery are not a few. And if it seems too good to be true, probably is too good.

Last, people have focused on a number of aspects of the docs, but that does not mean there are not other features, like uneven letters, like other numbers such as the 7, and so on.

What would also be nice is for CBS to let a panel of experts look at the originals. I think working from the poor .pdfs, which Dr. B states at least a couple times, makes it tough. Meaning, he could be in error cause the imaging is poor. That is a good reason to hold the horses.

Some Guy -

If the simplest explanation is best, throwing out the CONSENSUS ANALYSIS of forensic document examiners (about 10 at last count), the complaints by Killian's family deeply cloud what the likely/easy analysis would be.

Bill, all of the Executives were pre-selectric standard typebar machines. Even if the D model could be ordered with any font and superscript keys, I doubt they'd have curved apostrophe's or kerning. That needs to be confirmed. I've only worked with the more commercially availabe Executive 'C' model.

Some Guy, don't loose sight of the forest for the trees. Go read Reverend Sensing's analysis from last night. The crucial memo references an Air Force manual that never existed. If you give the benefit of the doubt and grant that it might have been referring to an Air Force regulation instead, then the memo referred to a regulation about special pay for foreign-language proficiency.

Not only are these documents forgeries from the point of view of their physical characteristics. They're also forgeries from the point of view of their content.

Bill,

Sorry if I wasn't clear. You have been precisely the antithesis of a blogger with a wild theory. You have been -- and I know this is a dirty word today, but once it was a compliment -- a journalist. You have printed verifiable facts and indicated the limits of those facts.

But some of the commenters on this and other threads, on this and other blogs, and on both sides of the debate have been into intense wild theory mode, as if their wild theories stack up to ANYTHING alongside people who are the top names in this field.

My field is counting. In that, I claim sufficient expertise in counting to comment: 6+ named experts vs. 0 named experts. Actually, thanks to your post, I'll put Lynn Huber in my table as one named expert supporting the authenticity. And I'll update further as I learn of more named experts supporting Mr. Rather. But right now, I can see how 6 named experts vs. 1 favors the 6.

Bill, granted it is not one way, but doc experts don't bear on how a forgery could have occurred in these circumstances. And as I mentioned, experts have also complained the poor quality of the imaging is a problem.

I really think CBS needs to let people see the originals. Then you could also date the age of the paper and that is a much surer way to end all this typewriter arcana.

The unevenness of the letters could have come from copying the document to age it, and then faxing it, and then scanning it for the PDF.

UML Guy -

re: "But some of the commenters on this and other threads, on this and other blogs, and on both sides of the debate have been into intense wild theory mode"

I agree, and sometimes it gets on my fucking nerves. Thanks for clearing that up.

Some Guy -

The original documents DO NOT EXIST. Or at least, CBS does not have them, only the photocopies that you see in pdf format.

The uneveness of some of the letters in the CBS memos pretty plainly is an artifact of repeated photocopying. What we're seeing is an umpteenth generation version. The forger used this technique as a crude way to make his products look old.

Sherlock Holmes had it right. That the documents are genuine is impossible. That they are forged is improbable but must be the truth.

Thanks for the link. IBMComposer.org says that the first Composer to have memory was released in 1975. It could store 5000 keystrikes. Unlike its predecessor, the typist didn't have to type the text twice. S/he typed it once and the Composer figured out how to do the spacing, then typed the page.

Yes, very good point Bill, we need to stick to verifiable facts, otherwise we join the T-Mac 'Karl Rove is responsible for anything bad that happens to Kerry' crowd.

I think a couple more days of analysis and this will be a slam-dunk. The kerning appears to be the smoking gun.

So now Rather needs to produce a 1972 typewriter capabale of all of the following:

Proportional TNR font
Superscript "th"
Curved apostrophes
13-point vertical spacing
Paired kerning
Perfect centering of proportional fonts

and it must be commonly available enough to have been on a Colonel's desk and all features described above convenient enough to use that a CYA memo would be typed using them.

Feel free to correct me if I missed anything or made any mistakes.

BTW, the last three appear to be impossible.

A little quibble, Bill. In the original post, you said, "...the memo has kerning, something widely believed to only be possible with modern word processing programs that intuitively adapt to the sequence of letters."

Computers don't do anything intuitively. They do it algorithmically. I know, it's supremely anal, but since I'm a "software engineer," I try to fight against the suspicion that computers are actually intelligent.

Great work, pardner.

Thanks Boyd. I was being sort of ... vaguely metaphorical, but I updated with "algorithmically"

GEEK ALERT

Boyd,

Call it "Turing intuition". If it seems to the user like it's intuitive, then it's intuitive in the practical sense... even if that means you or I had to pour two or three thousand hours into that "intuitive" code.

We software geeks are cursed, in this way: the harder we have to work to make the system work, the more the app just makes sense to the user, and the less they understand how hard it was. Give them a painful UI that just barely lets them get the job done (but is better than what they had before), and the users will recognize the effort. Give them something that darn near reads their minds, and they'll say, "Well, of course. How else would you build it?"

Which of the memos is supposed to be kerned? The CYA memo’s character spacing lines up perfectly with the text typed into a Word document in default settings. The default settings include “kerning for fonts” turned off.

When I highlight the text & turn the kerning on, the some of the characters go out of alignment with what you see in the CYA memo. It works just right without automatic kerning turned on.

Wow, did we just get an assist from AP? (!)

Several of the document examiners said one clue that the documents may be forgeries was the presence of superscripts — in this case, a raised, smaller "th" in two references to Guard units.

The network's statement said typewriters were available in the early 1970s which were capable of printing superscripts. CBS pointed to other Texas Air National Guard documents released by the White House that include an example of a "th" superscript.

That superscript, however, is in a different typeface than the one used for the suspect memos. Document examiner Sandra Ramsey Lines of Paradise Valley, Ariz., who examined the documents for the AP, said she was "virtually certain" they were generated by computer.

http://story.news.yahoo.com/news?tmpl=story&cid=514&e=3&u=/ap/20040910/ap_on_go_pr_wh/bush_guard_questions_5

Does anyone have any leads on who "Lynn Huber" is? I did a bit of web searching, but didn't find anything. I did find a Roy A. Huber who did handwriting analysis.

ForNow;

This is a misconception I expected to see: "Kerning for fonts" is not the same thing as "paired kerning." Paired kerning is how the "y" in "my" gets its tail stuck under the "m," and all word processors do this by default.

Kerning for fonts is DISABLED by default, at least on my comp.

Well, that’s a relief. Now if I could just find an example of kerning when I look closely enough at my Word file, I’d feel even better. I have “my” on screen in 500% view & 36 pt size font. The y’s tail doesn’t really go under the m. Is that absence of tucking some sort of monitor-screen thing?

About the kerning, a fella who teaches lay-out, desk-top publishing and all that kind of stuff over at a very anti-Kerry board (Swift vets) is 99% convinced these memos are forgeries, but equally emphatic that there is no kerning in them, for the reasons ForNow cites: it has to be turned on (and most people don't know what it is or how to do that) and if it is, the fakes no longer completely track a Word document using all default settings.

ForNow - look at TallDave's comment and my update. It seems we may have been wrong about the kerning.

Please go to www.bondwine.com\gulag\archives\000183.html. This is a blogger named Jay Random from Canada who has given the most complete analysis that I have found, including an analysis of the superscript TH.

Rather is done for.

By predetermining the results any investigation, he has openlt displayed any sense of journalistic integrity, placing his own hubris over anything else.

Eev if, in the end, he was right-- he's shown himself to be an agenda based hack.

see these:

http://thewanderingmind.blogspot.com/2004/09/rather-would-rather-not-update.html

http://thewanderingmind.blogspot.com/2004/09/cbs-bloggers-and-lawyers.html

OK, based on some quick googling looks like I was wrong. Apparently there is no difference, and the document is not kerned.

Still, that leaves two other impossible tasks for the magic typewriter, which is still two too many.

Make that 3 again, with the typeface change on the superscript

Just watched Rather on CBS. His expert is Marcel Matley, an acknowleged authority on handwriting analysis. His contribution to the piece tonight was to vouch for the authenticity of the signatures. He had zip to say about the issues discussed above...Rather handled that himself.

BTW and IMO, the presence of a superscript "th" on another, undisputed, record of Bush's from that time---even though the font is different---needs to be taken into account during these discussions.

How to explain the perfect match of the CYA character spacings with a Word Document (at least Word 2000) in DEFAULT SETTINGS? That struck me from the start.

The photocopy fuzziness also struck me. That’s why I wondered about the kerning. Things about distinctive font characteristics also are a little tough. If you’ve ever had to read a bad fax & determine a character by patterns of how characters in the fax look, it can be more difficult than reading sloppy handwriting & trying to distinguish a style of script “r”s from a style of script “n”s.

Yes I agree ForNow, that alone flunks the reasonableness test.

But we can't expect Rather or Dems to be reasonable. We need ironclad proof.

I think there some confusion about kerning...

Monoface type, in addition to have letter occupying the same horizontal dimension, also have space between the letters. That spacing is, more or less, the same no matter what the letters.

Letter spacing, used in typography and translated into computer software, varies that space based on the shape of the letter and is more variable than monospaced characters.

Kerning is the adjustment to letter spacing. It can be done easily--if tediously--on typesetting equipment as well as on higher-end word processing (I include MS Word in that category).

The "tucking of the tail of a 'y' under an 'm'" can be achieved through good font design, through letterspacing, or through kerning. You'd really need both the original materials and a contemporaneous machine to tell just how it was done.

I used the IBM Composer, in 1969/70. It was not much bigger than a normal electric typewriter, but it had a lot more doohickies on the face of it. It would fit on an ordinary typing table. You did need to type each line twice to take advantage of its ability to fully justify lines and make best use of its proportional typefaces. But it could be used as a regular typewriter as well, with output that was "near-proportional" in its visual effect. That is, it wasn't as precisely proportional as it would have been with full justification, but looked better than monotyped materials.

The machine I used did not have super- or sub-scripting available. If IBM says such was potentially available, I wouldn't doubt it. That shifts the question to whether the TANG had such machines and whether their machines had super- and subscript typeballs. There should be records--although God only knows where they might be--that could answer that definitively.

The Composer most certainly did not/not have kerning available. I had to learn about kerning when we went to a Compugraphic photo-typsetting machine, one of the very first commercial machines suitable for an office. That machine, though, was the size of a refrigerator.

The machines were not terribly expensive, and they were also available on lease. The USG was not in the habit of leasing office machinery (mainframe computers might be a different case), so there should be purchase orders for the specific machinery the TANG purchased. The TANG and IBM should have records somewhere.

Rather's arguments on the CBS evening news tonight were lame. He notes that Times Roman typeface was available since the 1940s, but elides the fact that it wasn't available on typewriters that early.

He said that in fact "some typewriters o the period had superscripting" but doesn't answer the question: did the TANG typewriters have that facility.

If there's any CYA going on, I think it's at CBS. If Rather gets caught in this one, he's truly toast. And he knows it well enough to be sweating bullets (Full metal jacket? Were those available in 2004? Could CBS get them? Or maybe they were dum-dums??)

Thanks John. Also seems unlikely a memo wuld be written on it by the Colonel.

I think the issue re: kerning comes down to the fact that whatever the kerning is or isnt, what happens in Word in 2004 somehow matches a typewriter from 32 years ago? Extremely improbable.

Is the "superscript" th the one listed on the Kerry Spot? Does it occur more than just one time. Because, it appears to me that the entire segment "111th" is distorted upward. Corrections at that period (done with tape or-later-Liquid Paper) often had a distorted appearance. Or, this could be xerox/fax distortion. I saw at least one other th that was not superscripted. It was possible with Selectrics to superscript at that time. However, a) it was extremely awkward so why do it on a few line memo; and b) it did not reduce the font of the th. Has anyone studied to see whether the obviously authentic, obviously typewritten documents have the superscipt th used repeatedly? Does the font change?

Rather is either falling apart, or is stalling for time while he gets some better arguments or research together. He knows good research takes time, when we go too fast we tend to make more mistakes.

(What’s not clear to me is where the kerning is supposed to be in the CYA memo. One apostrophe is close to a capital I. Other apostrophes are not so well spaced—but are just like in Word.)

The most obvious pairwise kerning example in the CYA memo to me was the "fe" in "feedback". I've also seen people cite "my" and "ny" (from "any"), but the "fe" looked more striking to me.

TallDave: Are you saying there is a phenomenon known as "paired kerning" displayed in the suspected forgeries that a typewriter in the 70s could not do? As I said in a prior post, some guy at the Swift Vets forum got a bit testy with me when I raised kerning as a reason to support that the memos are fakes (but he is convinced that they are bogus).

Bottom line: Does anyone think the rest of the media will put this issue to bed now that Rather has drawn a line in the sand? Or will it only get worse?

OK, the easiest way to address the kerning issue to determine whether there is any kind of kerning (other than proportional font adjustments) that happens by default in MS Word. By default, "kerning for fonts" is turned off in MS Word. Earlier, I thought there was another kind that word processors did by default ("paired kerning"), but Googling now leads me to believe these are the same thing and that we should drop kerning as an issue, given the many other undisputable grounds for the memos to be labelled forgeries.

Thie website has some good pics og the IBM composer...

http://www.ibmcomposer.org/SelComposer/description.htm

I created some snapshots of other word processing programs (Adobe Indesign, Wordpad) using 12 point times roman with 6 inches of margin. The results are here.

Note that each of these is slightly different on word wrap and line spacing, etc, whereas the Killian fakes and the Word docs match exactly.

I’d say the odds of the IBM composer and the Killian docs being the same are about .00001%.

The “fe” combination is built into the font’s proportional spacing for the letter “f”. You can see this when in Word in very large view you type “fe”, “fb”, “fk”, etc. The characters run right into each other & need, if anything, to be kerned apart. I think that the present relevant sense of “kern” is that of special adjustments from whatever monospace or proportional spacing is generally effect—in other words, pairwise kerning, which, done automatically, would use a lot of memory, storring kerning info for the square of the number of characters on the keyboard.

I’m not sure what the heck “kerning for fonts” does in Word, though I was a Word guru & have just checked Help & have googled around. It doesn’t do the things claimed for it on my computer here at home or any other computer I ever worked on — something about not having to repetitively kern the same character combo, some such typically badly explained Microsoft thing. I used to just assume it was some sort of extra-memory-using “sensitive” spacing function, because when you format everything in it, some of the spacing does improve a little.

One of the big differences between typewriters and word processors was the treatment of the space between sentences.

On a typewriter, one space is too small. Everyone's convention was two spaces.

On a word processor, one space is the norm (and some word processors deliberately eat the second space). Regardless, the _SIZE_ of this space is much larger than the normal interword spacing.

I haven't seen this brought up. It isn't "kerning", but it does require the 'spacebar' to know if the space is following a period or not. Not something _I_ know of pre-computer. The match at lgf got me to thinking on this - there's been books focused on the transition between typewriter and computer that pound this whole issue. "The Mac is not a typewriter" comes to mind.

I’m sorry, I worded a little sloppily. The Times New Roman “fe” works beautifully in Word, along with a number of such “f” combos. But this is at the price of “f” getting crashed into by a succeeding “b”, “k”, etc. “f” ’s default spacing is such that the next character comes close to it.

The spaces, yes. The spaces between the sentences in the CYA seem less than two typed spaces on old typewriters but are the same as two spaces after a period in Word. But that might be explained by a proportional-spaced typewriter, & the typist simply typing two spaces after a sentence, which of course is exactly what we all used to do whenever typing. I still do that sometimes, if I know that the double space will be preserved.

I mean, the argument that the CYA letter is real has to be that there was a proportional-spaced typewriter with exactly the same character spacing as Word 2000 with Times New Roman. The argument would be that Word’s character spacings were designed based on the same system which that particular typewriter used. (Also of course that the manual typist chose to make all the same line breaks as the Word program would.)

I think we are getting mired down in the details of typography and not seeing the big picture. There are certain pieces of circumstantial evidence that I think are even more compelling when taken in context.

1 - CBS says it got these documents from an unimpeachable source with presumably altruistic motives for releasing them. Why hasn't that altruistic individual come forward to tell their story and clear up a national controversy?

2 - Why did the Lt Col produce these memos? Did he really have reason to be so afraid about the situation that he had to cover his ass with memos to the file? Remember, at that point, GWB was not a notable figure, he was just some guy in a non-combat job.

3 - Aren't these memos just a little "too good to be true?" Evidence like this is generally found like a needle in a haystack, not laid out in a short and sweet manner. Was there at least some bigger collection that these were culled from?

4 - Why were the memos saved all this time, and evidently photocopied so many times? Somebody must have thought they were significant to do so, yet they never surfaced before now. Why?

If have done some more analysis, and wanted to share. It has been pointed out that some of the pages had centered, typed headers. (111th Fighter Intercepter Squadron,
PO Box... etc) Centering is difficult to do on a typewrite, since you do not know where to start typing for each line. It's easy in a program like word.

For this analysis I did a couple of things. First, to determine: are the headers perfectly centered (like a computer) or approximately centered (like by hand). To do this I just took the bitmap of each header, flipped in on horizontally, and superimposed it on itself. Viola, each line begins and ends exactly with it's mirrored counterpart. The match-up is perfect to the resolution of the image. (about 1 points?) This means that the centering was not just close, but accurate to much less than the width of a single character. A dead ringer for computer control.

Secondly, I matched the headers from seperate pages, and found a likewise perfect match. (Except that one was ~2% smaller in all dimensions than the other. An artifact of copying, presumably?) Such a perfect, repeatable, symmetrical centering could not be the product of a rapidly hand typed memo.

I am conclusively convinced that these documents were produced on a computer.

...and another thing:

5 - The Lt Col apparently could have only produced these on one of two different IBM typewriters (if he could have done so at all). Were there other documents produced by him in this period that look like these? Or are we to believe that he had a special typewriter used only to produce "CYA" memos?

...and another thing:

5 - The Lt Col apparently could have only produced these on one of two different IBM typewriters (if he could have done so at all). Were there other documents produced by him in this period that look like these? Or are we to believe that he had a special typewriter used only to produce "CYA" memos?

Erik,

Good eye, that was covered at LGF and is mentioned in my post above. The auto centering requires a typewriter with computer memory ...

Here's some stuff that goes beyong a discussion of old typewriters (something I honostly haven't seen in the "metal" since I was about 7, and I'm 22 now).

http://qando.net/archives/004070.htm

There's a lot here beyond letters and keyboards.

Two points:

The authenticity of the documents can almost certainly be decided if experts are allowed to examine the source docs even if they are copies (this is not a Godel issue - math nerds know what I mean).I believe they are forgeries, but could be convinced otherwise based upon expert opinion.

True GWB haters cannot be convinced that the documents are bogus no matter how overwhelming is the evidence that they were computer generated (cf. The Protocol of the Elders of Zion)

“Kerning” redux?
In Word, type any letter followed by a lower-case j. Do it in a very large view. The j’s tail tucks under the preceding letter. It will do so to any preceding letter, it’s not discriminate. But I wonder whether any proportional-spacing typewriter did that. Unfortunately, there’s only one j in the CYA memo—after a space in the phrase “my job.”

I found this site. Maybe it will help. Good luck. If a boycott of 60 Minutes' advertisers has not begun, might I suggest that it get started?

Thanks for the good work.

http://www.ibmcomposer.org/

Yes, the j’s tail IS tucked under that preceding space. And the character spacing proceeds perfectly in accord with that of Word 2000 Times New Roman.

So. Could any electric or electronic typewriter have tucked a j’s tail under a preceding character? Because that’s what it would have had to be able to do.

ForNow:

Well, it definitely protrudes into the previous space. But I don't think you can call it "kerning" because it does this for ALL letters; there's no special rule for different positioning beside different 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 "mj gj qj" at 500% zoom, then move the cursor between them, and you'll see what I mean.

A couple of points.

Jeff,

The manual referenced in the memos did exist. A couple of guys on Swiftvets researched it and it's mentioned in several of the legitimate Bush docs.

Now terminology and certain document standards were violated, so there's something there.

John,

I think you meant "monospaced" or "fixed pitch" rather than "monoface", and "proportional spacing" rather than "letter spacing". Letter spacing refers to tightening or spreading out text.

It’s not a matter of whether it’s really kerning or not, I’m not talking about pairwise kerning.

The q-tail does get hit by the “j”-tail. Shrink the q without shrinking the “j” & it becomes even more noticeable.

It’s a matter of whether any proportional-spacing typerwriter would have done allowed a character to have a rule of intruding, so to speak, on a previous character’s space. Or can the rules work only going forward?

Kerning (in the sense of going to Format & expanding & condensing the space between characters in Word) seems usually to be done in terms of what a character on the left will allow the succeeding character to do.

“f” lets any succeeding character intrude on its space, or to put it another way, it has a default narrow space. But every character lets a succeeding “j” intrude on its space. Since kerning seems to be done in terms of what a character on the left will allow the succeeding character to do, I imagine that proportional-spacing typewriters were designed that way, though I don’t know. If they were so designed, it seems doubtful that they would have had a way to let the “j” tuck up under the preceding character. Word 2000 somehow works it so that EVERY letter lets a succeeding “j” come back close to it. And “j” therefore tucks up under whatever is the preceding character, including a proceeding space. That would mean that the CYA memo was done with a word processor.

The IBM Selectric Composer was capable of variable pitch font spacing - not proportional spacing. Here's the main difference.

In variable pitch, unlike regular typewriters such as created all of the other released Bush documents, the inter-character spacing can be varied, but not the individual character spacing. Proportional spacing requires individual spacing adjustments that simply was not available on mechanical typewriters at that time. Important thing to keep in mind. Typewriters of that day usually had a fixed pitch, such as 10 or 12 characters per inch, which lead to characters on each line of typewritten text lining up on a vertical axis. Some more advanced typewriters (such as the IBM Selectric Composer) allowed you to adjust this pitch (thus the variable-pitch) so that you could justify the text both left and right on each line. But they were NOT proportional typewriters.

This, of course, assumes that we are all talking about an IBM model typewriter, which is not a given. Other typewriters, such as and Olivetti, or Sears, or NCR, were not as advanced as the IBM models were in the early 70's. They still relied on the old style arm mechinism to strike the page. It wasn't until the mid-70's the Xerox came out with their "daisy-wheel" printers and electronic systems that allowed the typewriter to "read" the daisy-wheel and determine which typeface family it was, and to allow proportional spacing based on pre-stored character size information programmed inside the typewriter. I had one, a Xerox "daisywriter" model of which escapes me

Kerning does not refer to the proportional spacing. Rather, it refers to the aesthetic way in which, properly kerned, two characters look more natural together when printed. That is why they are referred to as "kerned pairs" because you cannot have kerning data about a character without knowing who the neighbor character is. Some letters may have more than one kerning-pair association, while others may only have one.

And as to the now infamous "th" superscript - the IBM Selectric II did have available a ball called "Symbol" which contained superscript characters as well as basic math symbols. I used that ball extensively in college making money typing papers for people. It was a pain, though, to have to type out for a while, then change the ball to type one character, then change back again. But, go back to the document that CBS cited. Note that the "th" superscript falls exactly into one character space of the characters in the line below. Fixed pitch font. Going from memory here, the "th" superscript came from the T key, the "st" superscript came from the S key, and the "rd" superscript from the R key. But that was a few years ago.

It seems the "paired kerning" thesis is a sound reason to denounce the Rather memos as fakes. Read the series of memos posted at Hugh Hewitt's blog from a Prof. Robert Cartwright who teaches Computer Science at Rice: http://hughhewitt.com/

What I'm gleaning from all this is that there is a sort of automatic kerning in Word, and some other kind you can turn on if you want and know how to. But damn, I could be wrong, and I am surely NO expert. All I know is Prof Cartwright is certain the Rather fakes display kerning.

--Mona--

--Mona--

Definitely check out the site for the IBM Composer. It was not meant to be a typewriter, it was basically a device to help typeset stuff.

No one would ever use it to write memos with. You basically have to type the document twice with it.

Not an expert, and like the good COL (no period) I cannot type. But I do know a smokes creen when I'm lost in one. The whole point is to put enough rounds in the air to get EVERYBODY to duck and find cover. Kerry is using this to try and patch the hull of his campaign, which took a swifty torpedo amidships. Meanwhile the swifties are trying to line up their sights for another torpedo, this one with a bigger warhead( Kerry meting with the VC in Paris to plot the defeat of America). Unless some wants to try and sue CBS for something (? have they broken any Laws?) This will all blow over as soon as something else pops up. CBS will get a ratings boost and Rather will get a raise. Sorry boys, but I would like to see the statute that prohibits forging documents to smear an opponent in an election. The real verdict will be rendered on November 02, by the voters, and then made official after the law suites are finished in December or Janurary.

MS Word uses TrueType fonts, which have proportional spacing (and as the overlays shown on the Powerline Blog, seem to match pretty much dead on with the forgeries). The kerning feature you must activate by going into Format->Font->Character Spacing. It only has an effect on certain character-pairs (such as an A and V next to each other, or a W and A).

The mere fact that the document is done in proportional spacing alone suggest overwhenmingly that it is a forgery. An ANG unit, down at Ellington Field, would not likely have access to the high-end typewriters/typesetters that have been thrown about. In fact, the probably had something like an NCR or standard Selectric, or combinations of many different typewriters handed down from active-duty units. If someone could possibly come up with a scenario why an ANG unit in Houston, Texas, would have a fancy, proportional-spacing typewriter (and, BTW, the IBM Exective Selectric didn't have that capability until 1975), let everyone know, please.

Don't know if this was mentioned yet, but I read somewhere that the Composer had had a crank for manual kerning.

But the odds of a manually formatted typed document lining up so perfectly on overlay with a Word document done in Times New Roman are...well, let's just say I wouldn't make that bet.

No, Jeff, the Composer had two dials that you used to adjust (IIRC) the character spacing and the "space" spacing. Having used one of them many years ago (using one at my church to put out youth newsletters during my high-school years) I can remember that you could adjust the dials anytime you wanted. Here's what I do remember. You had a gauge in the center of the typewriter with two lines on it. You typed until you hit the margin, then you looked at the two colors on the gauge - I think it was blue and yellow. One was for the inner dial (the character spacing, I think) and one for the outer dial (the "space" spacing). You could then hit the MR key to tab beyond the margin and type in the two values (on top it was, IIRC, letters and on bottom, numbers). You then used that the next time you typed your document in to adjust individual line spacing. This was some 28 years ago, so forgive me if I don't remember it all exactly correct.

However, I do remember that the Composer used the same Selectric II Type-Balls as did the normal Selectrics, which meant that they were basically either 10- or 12-pitch (characters per inch) typefaces. Nothing proportional about any of those balls.

Last point, until the "www.selectric.org" site shut down it's link to anything and everything "selectric" - I did look at the catalog of availabe type-balls. Sadly, no Times New Roman. It did remind me that they did have the symbols ball, which I did use (I remember buying it myself because the College library didn't have it) so that I could earn money typing papers for people.

If the IBM Selectric Composer could do only variable pitch, it couldn’t possibly have done the CYA letter.

But it’s confusing. The IBM website talks about having intro’d the first proportional-spacing typewriter in 1941. “The Electromatic Model 04 electric typewriter.”

Well, kerning or no kerning, or proportional spacing or variable pitch or what - could some typewriter, electric or electronic, have been such that the letter “f” reaches forward over the following character (over it or into it if it’s tall like “b” or “k”), & also so that the letter “j” reaches back under the previous character?

I can imagine that the “f” effect could result partly from typeface design. The “j” effect seems more difficult.

To be clear, it’s not just that kerning is usually in terms of how close the preceding character, on the left, lets the following character come to it from the right.

I mean, think of the design choices in proportional spacing the same way.

The rule in force, I mean the amount of spacing, seems to depend on what character is before, to the left, not on what character is to the right. But the “j” is to the right of the character under which it reaches. In Word 2000, in Times New Roman, the “j” gets closer to the previous character than any other character is allowed to. I can’t help thinking that achieving this “j” effect would affect typeset or typewriter design constraints in ways that would have seemd more trouble than it was worth.

Both the “f” effect & the “j” effect happen with the CYA letter. The “j” effect in the CYA letter easily passes unnoticed because, in the CYA letter, the “j” reaches under a preceding space rather than a preceding print character (“this job.”).

Typo, sorry. I meant:

I can’t help thinking that achieving this “j” effect would affect TYPEFACE or typewriter design constraints in ways that would have seemd more trouble than it was worth..

I don’t mean that achieving the “j” effect requires pairwise proportional spacing. I mean that the “j” effect would require shortchanging the “j” of a bit of space to its left, a bit of space which every other character would have. That would mean giving every other character a bit of space to its left, just in order to be able to shortchange the sole character “j” of that space, so that the “j” ’s tail would tuck up under the preceding character. Would an electric or electronic typerwriter, even the most high-end one, have been designed that way in the early 1970s or earlier?

Update:

Just found out that the IBM Executive Typewriter, introduced in the mid-50's, did indeed have true proportional type (intercharacter spacing based on character width). And the font looks very close to TNR. And when typed in Word, using Times New Roman, the vertical alignment of words is very similar. The only difference noted is in the "th" and in the quotation marks. The executive still had the standard ' mark (or double for a quote) but the memo has the more common (these day's) curled quote marks.

Someone should definitely ask IBM to tell us which font face was on that 50's model Executive.

So the f curls over the following e, & the j curls under the preceding space?

So we’re hanging by a th & an apostrophe?

Someone should verify the military standards of that era for documents. They were always big on standards, and likely the guard got hand-me-down typewriters too. My gut says they wanted records to be typed as clearly as possible (i.e. thru multiple carbons and still be readable.) On top of this, many of the forms they typed went directly to key-punching. (put address in columns 23 thru 52, etc.) And those were -all- fixed pitch, not proportional.

Someone needs to go back and look for photos of typical TANG offices during those years and identify the typewriters. And/or old purchasing records.

Wonder what the effect of the swift boat vets would have been if the media had put one hundredth of this effort (and faith) into their story?

/Ari

Shape of Days has been experimenting with an IBM Composer along with Gerry Kaplan, an expert on the machine.

To my satisfaction, the Composer has been ruled out.

Read the linked post.

The centering of proportionally spaced text is tremendously difficult. Gerry got his Composer to closely match the letterheads, but with A LOT of work and tinkering.

What are the chances that Lt. Col. Killian - a man who didn't type - did this much work to center his letterheads?

Low.

And what are the chances that his laboriously centered letterheads of thirty years ago just happen to be a perfect match for the default settings of Microsoft Word Times New Roman?

I'd say zero.

As I said, read the post. IBM Composer is ruled out as far as I'm concerned.

(I always wanted to be a private dick!)

Er... I make truetype fonts. I could load Times New Roman (or any other font) into a program called Fontlab. I could then subtly move individual letters up and down a tiny fraction to give a slight "typewriter" effect. I could then substitute some of the numbers. I could lop off tails on words, and add others. If I was clever I could create a "blended" font by a wholesale auto-combining of two similar fonts. Then I'd save out my "new" font and type my Word document with it. I don't think this is at all likely - because if the forgers knew about this capability then they wouldn't have produced such a blatantly obvious fake - but it's something that hasn't yet been taken into consideration in the debate.

Here is a link to an IBM Composer for sale on e-bay right now.

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=3839419995&fromMakeTrack=true

The item is located in Holyoke, MA. Someone close may want to buy this thing.

Dave - Shape of Days blog has beaten you to it. He has a very detailed analysis of freshly-typed Composer samples that use the relevant font...

http://shapeofdays.typepad.com/the_shape_of_days/2004/09/the_ibm_selectr.html

I think it's just a matter of time before ALL of this is settled and Dan Blather is left wondering what to do.

It's a conspiracy theory for sure, but it wouldn't surprise me if the Clinton clan had something to do with the generation of documents with subtle - yet detectible flaws - to torpedo Karry's election and make way for the Savior (Hillary) to enter the ring in '08. That may be far-fetched, but not any more than the long list of dead people that got in their way over the years!

If you're a note taker but not a typer, you're probably unfamiliar with all the features of any particular typewriter. So if you do use one, you'll use whatever features you're comfortable with and you'll probably repeat it consistently. If the typewriter has superscript capability, you'll either use subscripts consistently or you won't, but the way you use them will probably be consistent. So, it's odd that Killian would put a space between a number and its potential superscript in some places, and then go to the trouble of using an elevated superscript in another. All for a simply "CYA" personal note to oneself.

Even if the typewriter did have superscript capabilities, Killian was familiar enough with TANG reports, memos, etc. that he probably wouldn't have even thought of separating the number from the potential superscript with a space, even if he later suddenly decided to use the superscript feature in the memo. 1st, 2nd, 187th, etc. without superscripts would have looked normal to him, not 1 st, 2 nd, 187th, etc.

Also, anyone raised using typewriters without notes to guide you always looks at what you type to catch errors ASAP. To someone like Killian, the 1 st, 2 nd, etc. would leap out at you and you'd correct it ASAP, or at least try to avoid making the same mistake later on. If he was typing something up from notes, he might take his eye off the print to read the notes, but you usually need to be comfortable typing to do that. My guess is Killian could have been a hunt-and-peck guy. Hunt-and-peck types often are more careful to avoid mistakes and are therefore slower and less prone to add spaces where they aren't needed.

P.S. Was CYA a common term used in the early 70's? I don't recall using it myself back then.

Read many of the posts. Would like to clear up some points:

1) The centered proportional text that lists the address of the squadron was without a doubt letterhead and not typed. In other words, it was typeset and printed on the paper before being put in the typewriter. That was standard practice before computers. You will also note that the "th" is not superscripted in the leterhead but is in the typed memo.
2) The fonts in Word and the fonts in the letters do not match exactly. Specifically, the 8 in Word is composed of two equal o's In the letters, the lower o is larger.
3) If the end of line bell was set to 5.5 inches, that would have produced the words per line in the CYA memo.
4) The memo and a Word document line up very closely. The differences are small but they are there.
5) It is not unreasonable that the Word font is a standard font and could have been the same standard font on a high end typewriter.
6) I look forward to someone finding a typewriter that matches the memo as closely as the Word document does.

I'm just a guy that works for a bank. I had time yesterday to recreate the documents in Word. What is the probability that my XP machine can EXACTLY reproduce a paragraph typed in the early 70's on an unknown typewriter???

Clear evidence of the forgery itself exists in the deliberate spacing between the number and the letters. In this case the forger intentionally puts a space between the number and letter such as in "1 st, 2 nd or 187 th" to deliberately prevent Microsoft Word from triggering the smaller superscript function. However, this forger was apparently rather inept and failed on at least one occasion to insert the space causing the superscript to appear immediately following the numeral. Although, superscipt could be achieved using normal typing by mechanically scrolling up a half line, no IBM typewriters at the time could also shrink the type down to the smaller size as Microsoft Word does in its superscript function. This fact taken together with the forensic analyses of such word processing and typefont experts as Dr. Bouffard and Mr. Bill Flynn cited above, the fact that Lt. Col. Killian did not keep private separate file records and the clear disparities of the known signature exemplar's of Lt. Col. Killian overwhemingly demonstrate to any casual and dispassionate observer that the Memoranda in question are not authentic and that Mr. Rather and CBS can no longer be relied on as reliable or credible sources of news or analysis.

One item for investigation that I have not seen:

IF it is Word, then which printer was used?

I did see a note on LGF or somewhere about having gotten better results once it was "printed" but no one (to my knowledge) has worked on different PRINTERS for matching from Word->Printer->RatherGATE pdfs.

No one seems to remember that people who learned to type on a typewriter, as did all who learned before 1972, used a double space after sentences. An ameteur typist like the Colonel would have continued to double space after sentences out of habit. There are no double spaces after sentences in the Rather forgeries.

Here is a place that specializes in all sorts of old typewriters. Why does not someone go there and see what they have instead of all this "speculative opinion". "I don't think that they had the technology then" please - get facts not opinions.
http://www.mytypewriter.com/aboutus.html

Nick,

You are coming to the party late, and if you were widely read on this blog and others, you'd know that we did this already and you sound like an idiot.

Not to mention, this post represented the evolving "opinion" of one of the foremost forensics experts in the country who WAS COMPARING the memos to his ENCYCLOPEDIA of old typewriter fonts.

Bill, Point taken, you are right. I only became interested in this this morning and I only came across this site a couple of min. before I rushed out the door this morning and had not the chance to see every entry. Went back and read more this pm. Lots of good stuff.
Since this was my first ever entry in one of these discussions, I take this as personal education.

As contribution from personal experience, I will add that when I used typewriters ('60s and '70s) I did not leave space between number & "th" . I also used double space after sentences.