Skip to main content

I need help, and right away.  (Ain't that the truth, tho...)

Seriously, I could use some help from any interested parties.  This site has gotten the much-needed help of someone with an actual working IBM Selectric Composer, and that person as sent back a type sample.

As you can see from their site, there are serious differences between the type sample and the actual memo, which they are using to argue that the Composer could not have produced these documents.  However, they may have proven the opposite.

There are several big differences between the memo and sample.  First, the line spacing is very, very off; the Composer author admittedly did not set the linespacing the same (it is adjustable to the point, on the Composer.)  So that's not really an issue either way.

The superscripted 'th' is incorrect; that's not really an issue, because we have already heard that custom typeheads were available with special characters for large clients, and certainly the United States Military would count as a large client, and one that would want the 'th', 'nd', and 'rd' superscript glyphs.

Most significantly, however, the space allocated to spaces is quite off... the Composer sample has space characters that are roughly twice the size of the memo sample.

That would be a deal-breaker, except for one detail.  As the Composer manual, page 18, says:

The Space Bar on your Composer can be used just like the one on a typewriter -- to space between words.  Three units is the standard space created by a single depression of the Composer's Space Bar.  This, however, can be adjusted by turning the inside dial at the lower right of the machine.  Turn the inner knob so that the white line points to one of the colored wedges.  The numbers on these wedges represent the unit value of the Space Bar in that adjustment.  You can, for example, set this dial at fiave and create a five-unit space with each depression of the Space Bar.  (For rapid movement of the Element across the page you can set the dial on nine, the largest number of units which the dial permits you to space.)

So it would appear that the Composer can indeed also create the spacing of words in this document.  And note that the spacing doesn't (seem to) vary within the memo; it was just set on one setting, and left there, as you normally would when typing.  But that setting is much smaller -- about half -- on the memo sample than on the sample sent by this Composer user, who seems to have the setting on his machine set to perhaps 7 or 8.

We have established that both vertical and horizontal spacing can be duplicated by the Composer, so now there is something much more interesting.  The font style of the Composer sample is 11pt Press Roman; Press Roman is the version of Times New Roman developed for the IBM Selectric and Executive typewriters.  Look at the sample.  Can you see any differences between that typeface, and the memo typeface?

Seriously.  This is very important.

As we have discussed, the "Times New Roman" font used to compare the memos by the LittleGreenFootballs crowd was, upon examination, very different from the memo typeface.  People have pointed out differences in perhaps a dozen different letters and numbers.

What about this type sample?  Can anyone see any differences there?  The 9 is a little lower on one vs. the other, but that could very well just be an artifact of the individual machine.  The 1's and 3's, as well as the other letters, seem absolutely identical upon my own examination.

What is even more interesting, the proportionality between letters -- take a look at the "F.L.S.", for example, which has rather awkward spacing -- also looks identical in both letters.

I know people are tired of this, as am I, but take a look.  Can anyone find differences in typeface, aside from the 'th' glyph?  Seriously, don't f--- with me here, look honestly and closely.

Have they just inadvertently proven the typewriter used to produce the Killian documents, down to the typeface, point size, and settings?

UPDATE:  And note that the IBM Executive, a much cheaper machine, was available with this exact same typeface (though we don't yet know the spacing characteristics.)

Originally posted to Hunter on Sun Sep 12, 2004 at 10:19 AM PDT.

Your Email has been sent.
You must add at least one tag to this diary before publishing it.

Add keywords that describe this diary. Separate multiple keywords with commas.
Tagging tips - Search For Tags - Browse For Tags


More Tagging tips:

A tag is a way to search for this diary. If someone is searching for "Barack Obama," is this a diary they'd be trying to find?

Use a person's full name, without any title. Senator Obama may become President Obama, and Michelle Obama might run for office.

If your diary covers an election or elected official, use election tags, which are generally the state abbreviation followed by the office. CA-01 is the first district House seat. CA-Sen covers both senate races. NY-GOV covers the New York governor's race.

Tags do not compound: that is, "education reform" is a completely different tag from "education". A tag like "reform" alone is probably not meaningful.

Consider if one or more of these tags fits your diary: Civil Rights, Community, Congress, Culture, Economy, Education, Elections, Energy, Environment, Health Care, International, Labor, Law, Media, Meta, National Security, Science, Transportation, or White House. If your diary is specific to a state, consider adding the state (California, Texas, etc). Keep in mind, though, that there are many wonderful and important diaries that don't fit in any of these tags. Don't worry if yours doesn't.

You can add a private note to this diary when hotlisting it:
Are you sure you want to remove this diary from your hotlist?
Are you sure you want to remove your recommendation? You can only recommend a diary once, so you will not be able to re-recommend it afterwards.
Rescue this diary, and add a note:
Are you sure you want to remove this diary from Rescue?
Choose where to republish this diary. The diary will be added to the queue for that group. Publish it from the queue to make it appear.

You must be a member of a group to use this feature.

Add a quick update to your diary without changing the diary itself:
Are you sure you want to remove this diary?
(The diary will be removed from the site and returned to your drafts for further editing.)
(The diary will be removed.)
Are you sure you want to save these changes to the published diary?

Comment Preferences

  •  have you sent this (none)
    to dan rather/cbs?  

    they may be interested.

    "..The question was, who was going to attack us, when and where, and with what." ..."now, watch this drive."

    by x on Sun Sep 12, 2004 at 10:26:55 AM PDT

    •  No, (none)
      I'd like some verification first.  This stuff is tricky; the more sets of eyes, the better.
      •  Any concern that we're doing the wingnuts' work? (none)
        We've demonstrated pretty clearly that there's no way you can definitively say MS Word produced the CBS memos. And CBS has stated the same thing. At this point, I'm not sure that it's beneficial to emphasize that we're devoting a lot of time and energy trying to PROVE that a typewriter from that time could produce an exact match. The documents are old photocopies, so the match won't be exact in any case.

        It is, as you have pointed out, a needle-in-a-haystack effort, and I have serious doubts that there is time before Nov. 2 to really settle it anyway. But this could lead to claims such as: "They couldn't find an exact match! It's a forgery!"

        I say, let them prove that no typewriter could have done it. The burden of proof should be on the accusers.

        I'm going to go watch some football now :)

        •  Not really worried. (none)
          If we can prove the typewriter, that would be definitive.  We've already demonstrated that their own Word demonstration was, for lack of a better word, brick stupid.  We don't need to find the exact typewriter, but wouldn't it be nice if we did?

          If we can't get this settled in the next day or two, it'll be meaningless, and we should let it drop.  If there are any real journalists left in the world, they're descending on IBM headquarters like locusts right now, and we should know soon.

  •  Did you see the character width chart? (none)
    Hunter- did you see the update on  The guy posted the character widths for every letter on the composer.

    Q: What is the character spacing on the IBM Composer?

    A: The IBM Composer measured characters in "units" - a character occupied between 3 and 9 units. A "unit" size is dependent on the font's size. There were three preset horizontal spacing settings. A larger font size, like 11pt, would require the widest setting, while a smaller font such as 6pt would require the tightest spacing. That being said, here are the three unit sizes:

    Widest: 12 units per pica, (72 units per inch)
    Middle: 14 units per pica, (88 units per inch)
    Narrowest: 16 units per pica, (96 units per inch)

    Here is the table that shows the widths of all of the characters on the Composer keyboard:

    With this information we should 100% absolutely beyond a doubt be able to say whether the composer did these memos.  From this chart we know what the spacing would have been using all 3 different spacing patterns.  We can total up every line and say how long they would have been if the composer had written them.  Then measure how long they actually were.  This is a dead on perfect way of verifying whether the composer did this.  I am sure a similar table would be available for the executive.  It wouldn't matter which font ball was used because the spacing would be the same for all of them on the machine.

    You know fonts so much better than me, but it seems to me this is going to decide it one way or the other.

    •  I've seen it, (none)
      but it doesn't tell us much, because other similar typefaces would have had identical character spacings, as we can see from the Word Times New Roman comparison: slightly different font, but identical proportional spacing.

      The spacing setting of the "space bar", however, is a big deal.  As you can see from their samples, it makes a great deal of difference.  If you set the space on the Composer to a smaller dialed width, would these lines indeed be identical?

      •  character spacing (none)
        Hunter- are the character spacings that word uses well known for times new roman or other variable width fonts?  I find it rather hard to believe that word would have limited itself to just 6 character widths.  But if we have that table and can compare it to what the composer used that would help a ton.
        •  It would be defined (none)
 a standard part of the typeface; that is, the "proportionality" between letters is a part of the typeface's detailed design.  It'd be in the design handbooks, somewhere, if anyone has those around.

          Since Microsoft's Times New Roman was very explicitly designed to mimic the original Times New Roman as closely as possible, I think it's safe to assume that they did indeed use the exact same character widths, and so the character spacings are identical in each.

          •  standard widths? (none)
            Hunter- I am REALLY confused now.  Online I have read that the executive could only do 4 widths and every letter had to fall into one of those categories.  Now we learn that the composer can do 6th widths (and we know that all are used).  Times New Roman on the executive would be distintuishable from Times New Roman on the composer because of this right?  Or am I missing something important here?
            •  It's possible... (none)
              ...either way.  So far I don't necessarily know whether they were the same or not.  We also don't know if the differences would be enough to pick up on in such a blurry photocopy -- there's a lot of 'bleeding' going on at each letter.  If you compared the Executive and Composer typeface to each other, side by side, you might be able to see very tiny differences, but comparing either to this memo won't tell much.

              The differences between 4 proportional widths and 6 proportional widths, when talking about fonts this small, are very subtle, but I suppose it's possible.

              •  See below... (none)
                Someone appears to have made an off by one error in reporting the of "only 4 widths" thing that I see all over. I think the known values are 5 (for the Executive Model A) and 7 (for the Composer). Although the values for the Executive B, C and D are not known they must be at least 5.
      •  one more thing (none)
        The word spacing would ahve nothing do do with the start of the 2nd line.  May,1972  is the "word" that starts out the line.  But by my eyes, the character spacing has already thrown off the sample compared to the memo.  To me, this would indicate they are using different character spacings and the word spacing doesn't matter at all.  Is there another explanation that I am missing?
        •  That's a good point, (none)
          and one that I've been trying to figure out either way.  First, in his example, he isn't quite lining the two samples on top of each other to begin with -- he's put the 'red' sample noticably to the left of the 'black' sample, and while he's tried to scale the samples so that the memo matches the size of the Composer sample, since the spacing between words is so completely different in the two samples, his efforts at scaling are pretty badly off too.

          Plus, of course, the Composer author used true '1's for his sample, and the memo quite clearly uses lowercase ells in place of ones, which screws up the spacing both of the "1." and of '1972' below it.

          Added up, that's why the only useful formulation here, at the moment, is whether the glyphs themselves are or are not identical.  I am looking closely at the capital "M"s, however... are they different, or is the scaling just off?

          So if someone here is willing to fight with photoshop and do a more credible job of lining and scaling, that would tell us something, but in order to do the true match we need the Composer author to fix his typos and settings.

          •  Hate to ask... (none)
            The M on the Memo certainly does look wider, even with the scaling.  Assuming (from information lower in the thread) that the widest characters on the executive would be 33% narrower than that of the Composer that would pretty much eliminate the executive as a candidate.  Do you tend to agree with this?  
            •  The M is wider in the memo (none)
              because the memo is not Times Roman

              "The military and the monetary... Get together whenever it's necessary" - Gil Scott Heron

              by zane on Sun Sep 12, 2004 at 04:39:05 PM PDT

              [ Parent ]

              •  ummmm... (none)
                If it was typed by the composer the maximum the width could be is 9 units.  If it was typed by the executive the maximum the width could be is 6 units.   The point is that the Composer sample provided didn't seem to be wide enough to match the M of the memo.  The font doesn't matter, there is a maximum width based on the machine.  It could be a scaling issue so the Composer M is wide enough to match the memo but I don't think a 6 unit letter would be enough.  If you are looking at the shape of days site, it would be the equivalent of the letter o in to on the same line eventually scaling to be as wide as the M of the original.  I just don't think that is possible.  But for this discussion the font doesn't matter because the width limits are set by the typewriter.
            •  The M looks wider to me, too (none)
              the more I look at it, as does the 'W'... see the text "IAW AFM" (which, as a side note, shows absolutely no kerning on either copy.)  Everything else looks like a perfect match, Composer-to-memo, but those two characters still seem off to me, and I can't rationalize it with photocopy bleeding or other distortions.

              So it's much better than Word -- at least the letters all have the serifs in the right places -- but not good enough, yet, IMHO.

              Hmm.  We think that the Executive had fewer 'units' of proportionality than the Composer, but we don't know if those 'units' were differently sized than the Composer.  So I'm not sure we can assume the characters would be narrower; in fact, I would really doubt it.  An 'M' 33% narrower than that of the Composer sample would be very, very awkwardly shaped, after all -- it'd look goofy.

              Sigh.  We simply have to find an Executive from that era with Press Roman font.  That's still our best bet.  Supposedly, the things were dumped into the surplus market by the bucketful in the mid/late seventies, so there should be some around.  And damn it, IBM should have one.

  •  Have you seen this? (none)
    It's another example of Selectric Composer vs. MS Word.  The two are virtually indistinguishable:

    Granted, the paragraph is pretty simple.  But the word placement, line and character spacing and fonts are virtually indistinguishable, using default MS Word settings.

    Visit the Diary of the Lying Socialist Weasels, for commentary from the Original Progressive Web Warriors!

    by Jonathan on Sun Sep 12, 2004 at 10:32:26 AM PDT

    •  Oops (4.00)
      Here are the two samples side-by-side:

      Visit the Diary of the Lying Socialist Weasels, for commentary from the Original Progressive Web Warriors!

      by Jonathan on Sun Sep 12, 2004 at 10:35:02 AM PDT

      [ Parent ]

      •  Yep, (none)
        It's from PC Magazine.  And they even link to us:

        A great deal has been made of the fact that some documents that are claimed to have been typed in the early 1970s look very much like documents prepared in Microsoft Word in 2004. This fact proves nothing, because (1) a document may well have been typed on a typewriter in the 1970s and (2) virtually the same document can be prepared on a computer in 2004. (Some other comments on this issue, from a notably better-informed perspective, may be found here.)

        The part about "a notably better-informed perspective"?  It links right to the original dKos article!

        •  fun fact (none)
          The guy who wrote that article is a famous and distinguished literary critic at Columbia--Edward Mendelson. He's the world authority on W.H. Auden, and the poet's literary executor. As a sideline, he's also a computer and typeface freak.
      •  superimpose them (none)
        Somone want to go ahead and superimpose them like the right has done with theirs.  Maybe someone who is really good can make it alternate back and forth in an animated GIF?
    •  Price of Selectric Composer (none)
      I've heard this model was very expensive in 72 and it isn't likely the TANG would have had one. Anybody have pricing information?
      •  The IBM Composer was expensive and large (none)
        When I joined a newsletter division of McGraw-Hill in the early '80s, they had IBM Composers for formatting the newsletters. They were larger, more cumbersome versions of a Selectric, not easy to use, and after heavy usage they would start to show little differences in letter impressions and spacing. Needless to say, they were expensive, specialized machines for publishing. They were definitely NOT the kind of machine one would use to type up an internal memo.

        I must agree with another diary entry by a typesetting expert who says that a document photocopied multiple times and then rendered into a PDF and displayed at 72ppi is impossible to evaluate.

        Beyond that, has anybody observed the office trappings of a typical government bureaucracy? Can anyone imagine state-of-the-art office typesetting technology being employed by the TANG in the early '70s? While such a thing is possible, is it probable? My experience with government offices is that their technology, whether in '70s typewriters or 21st century computers, is several generations behind the times.

        I'm certainly prepared to believe that the memos are genuine. Maybe the TANG just got a shipment of spanking new IBM whatevers with a load of different type-balls. I do agree with Steve Gilliard that keeping the story alive hurts Bush, and that's all to the good.

  •  IBM Executive Model D (none)
    Was there any functional difference left between a Composer and an Executive and the memos?

    It seems that since Microsoft Word 3.0 had the goal to mimic the output of the IBM Executive (going on memory here) that it's the better candidate.  It had the split spacebar that could be used to fine tune interword spacing too.

    I'm tempted to email this guy to see if we could pay to borrow his Model D (although we still need to get the typeface).  I wouldn't even bother with the superscript 'th' since that was obviously a common military customization back then.

    •  The only problem with the Executive (none)
      ... at this point, presuming these fonts do indeed match, is that we don't know the exact spacing between words or between lines that was available on the Executive.  Indeed, however, the Executive had the split spacebar.

      So we'd still need to find a working model with this exact typeface to be able to prove it was an Executive, but if these fonts match, then we know for certain that at least the Composer could do this.

      •  I don't have anywhere else to put this (none)
        But I stumbled across a little factoid while browsing docs at IBM Research. The first IBM Executive Model A had 5 different possible character widths (which also doesn't mean that Models B, C or D couldn't have had more either).
        Ever since the invention of movable type, each letter in the alphabet had been given a unique width to make its appearance pleasing to the eye. The conventional fixedescapement typewriter required that all characters be of equal width. This squeezed large characters such as M and W and provided more than ample space for the thinner characters such as i and 1. To provide variable spacing for different letters, IBM's first proportional spacing typewriter used a rotary type of escapement mechanism of three separate escapement wheels designed to provide 2, 3, or 4 units of carriage motion [Fig. 2(a)]. Used in combinations, it was possible to obtain 2,3, 4, 5, or 6 units of carriage travel, which provided a wide range of possible character widths and a more pleasing quality of print. A subsequent design employing a multiple- pawl linear escapement [Fig. 2(b)] was first used in the Model A Executive and has remained basically unchanged for more than thirty years.

        Later the same document mentions that the Select Composer has 7 different character width possibilities

        Because providing for an unlimited variety of type widths was impractical, a compromise was established at seven different, selectable widths. Thus, all fonts were designed to fit a proportional system that provided escapement values of three through nine units with three pitch sizes of 1/72, 1/84, and 1/96 inch.
        Again, I'm just dumping this info here in case someone else can make use of it.
        •  Great work (none)
          thats really nice work. So do I understand it that on the executive the widest characters could only be 6 units and on the composer the widest characters could be 9 units?  Does that mean the widest characters on the composer are wider than on the executive?  Just want to make sure I am comparing apples to apples here.
          •  All signs point to maybe? (none)
            It really seems like the units would correspond linearly, but I can't be sure yet.  It's possible that it's one of: linear, constant + linear, or something else.

            Given the contents of the second quote I'm pretty sure it's one of the first two (probably first).  Which means for a pitch size of 1/72 (inch?) you had a range of 1/24 - 1/8 (inch?).

            I'm really stretching on this however.  I'm open to better interpretations.

            •  must be it (none)
              On the composer at least thats how it works.  There are 3 unit settings.  On the narrowest there are 16 units per pica which is 96 units per inch.  Middle is 88 units per inch.  Widest is 72 units per inch.  

              On the exexutive there were no removable fonts (it was all a typepad) so I am guessing there was no ay to vary this setting at all.  And I certainly would guess that the executive's unit size was one of the composers but this could be wrong.  But to me the point is that we need to focus on getting the details for the composer then backtrack and see of the executive could also do it.  Basically anything the executive could do the composer can do as well.

              We have the character chart for the composer so we should be able to say exactly how long any given line of text would be if written on the composer right?  look at the may 4th memo in the line that starts out "1.  You are ordered" there are 453 units (assuming the default 3 units per space- there are 17 spaces so it could be off by a multiple of that)  453/72 = 6.29 inches.  The l vs 1 issue could also throw this off some.  

              The line below that "MAY,1972" should be 4.98 inches long using the same caveats.  

              Is there any way to determine from the PDF's how long those text lines are on the originals?

  •  I think we need to be careful (none)
    Just because someone is typing a sample up now on a Selectric Composer to compare with MS Word doesn't mean they will match.  

    For one thing, the person could have an intent to make them look different.  

    For another, the Composer is likely old, and may not be functioning correctly.  

    For a third, it's not enough to have the typewriter.  You presumably have to be very good at using it, as a secretary from 1972 would have been.

    Visit the Diary of the Lying Socialist Weasels, for commentary from the Original Progressive Web Warriors!

    by Jonathan on Sun Sep 12, 2004 at 10:54:45 AM PDT

    •  Correct. (none)
      ... that's why I'm explicitly asking about the typeface when compared to the memo, not Word.  We already know Word is different.  I'm asking whether, if you leave aside all the spacing issues, the Composer typeface from the sample they have obtained is identical to the memo typeface they're comparing it to.

      If the typeface matches the memo, even independent of spacing between words, it's a huge deal.  Because nothing else found so far does match that memo typeface, letter for letter.

  •  Executive did proportional (none)
    And note that the IBM Executive, a much cheaper machine, was available with this exact same typeface (though we don't yet know the spacing characteristics.)

    The Executive did proportional.

    It was a typebar machine, but individual, special typebars could be ordered for the "th" glyph.

    Case closed.

    •  Right, (none)
      but we don't know how many 'points' the spacebar would advance between words.  That's the only thing we're missing.

      The odds are looking increasingly good that it was an Executive.  Proving a font matchup with Press Roman would almost clinch it, since it and the Composer both had this font.

      •  why? (none)

        Chances are, that MS Word does exactly the same, by design.

        Heck, CBS or any newsorg with pull should contact IBM.

      •  I'm with you on this (none)
        My gut says that it's an Executive. We just need to find one with the right typeface.

        On the PCMag samples I can't get them to match without streatching one out.

        I agree that the fonts don't seem to match completely, but I think it's because of artifacting on the copies.

        I had after diary here thought that it was Aldine Roman, but the capital J doesn't match.

        I don't know if you saw my post that the Times New Roman font was redesigned in the 80's to match the Times Roman font. So I think we could be likely looking for Times Roman.

        typepad site though has a good section on brands of fonts.

        Have you been able to reproduce the LGF test with the Word doc lining up with the Memos? Or are you working on the assumption that they are legit.

        typepad site did some tweaking to make the pdfs match for his centering test.

      •  Two spacebars... (none)
        My fater had an Executive and the spacebar was split - IIRC one half did two unit space, the other half three.  I don't recall if it was a model A, B, C or D.
  •  Hunter, I don't know if you've seen this yet (none)
    more Killian memos have been uncovered.

    The PDF quality isn't as good as the CBS News versions, however.

  •  Re: Centering, Etc. (none)
    The blog linked to in the original diary (above) makes a big deal out of centering and how it is precisely the same between the two memos.

    First of all, precise typewriter centering is not as big a deal as claimed, especially for an experienced senior secretary, and especially for the equipment we're talking about (the IBM Selectric Composer). (I'd also put the IBM Executive in that category.) Check the Selectric Composer's manual, available at, and read through the section on centering.

    Second, that information (the unit's return address) would have either been typed over and over again (and, thus, done repetitively by the same secretary, the same way) or been standard pre-typed (or possibly preprinted) letterhead. (Remember, the Selectric Composer was routinely used in creating camera-ready copy for job printing, such as letterhead.)

    The Microsoft Word "test" is, once again, not particularly illuminating. The comment that Word can position text more precisely is academic when we're talking about a low resolution original (the Bush memo). It's difficult to get even a full point of detail out of that degraded PDF everybody has been using.

    Finally, I'm still betting on the IBM Executive Model D rather than the IBM Selectric Composer, so I'd like to look at a sample from a military-issue Executive D from that era first.

    Support our troops. Let's not repeat Vietnam in every detail.

    by sipples on Sun Sep 12, 2004 at 11:06:58 AM PDT

  •  Look especially: (none)
    at the 3's, 4's, 5's, 7's of the Composer.  They seem to match the document exactly, but Word's 3's, 4,'s and 5's, don't match.

    Note also that you can tell the memo uses lowercase 'L's for ones, whereas the Composer sample is using true '1's.

    Seriously, the proportions of everything look right, here.  What am i missing?

  •  Former expert for wingers retracts (none)
    I'm not much for amateur analysis of these things.  If anyone could really be an expert on it in two days, then forgeries would never be discovered, cause it'd be so simple to do.

    I direct you to the Boston GLobe article where an expert who formerly believed the docs could not be generated on computer now believes they could have been:

    Bouffard, the Ohio document specialist, said that he had dismissed the Bush documents in an interview with The New York Times because the letters and formatting of the Bush memos did not match any of the 4,000 samples in his database. But Bouffard yesterday said that he had not considered one of the machines whose type is not logged in his database: the IBM Selectric Composer. Once he compared the Bush memos to Selectric Composer samples obtained from Interpol, the international police agency, Bouffard said his view shifted.

    In the Times interview, Bouffard had also questioned whether the military would have used the Composer, a large machine. But Bouffard yesterday provided a document indicating that as early as April 1969 -- three years before the dates of the CBS memos -- the Air Force had completed service testing for the Composer, possibly in preparation for purchasing the typewriters.

    As for the raised ''th" that appears in the Bush memos -- to refer, for example, to units such as the 111th Fighter Interceptor Squadron -- Bouffard said that custom characters on the Composer's metal typehead ball were available in the 1970s, and that the military could have ordered such custom balls from IBM.

    ''You can't just say that this is definitively the mark of a computer," Bouffard said.

    That's from this article

    So, if an expert has changed his mind based on new evidence, I figure some amateurs maybe should study for at least a few weeks before starting to photoshop text and make definitive declarations from their "expertise."


    •  actually (none)
      actually bouffard feels he has been grossly misrepresented by the globe (and now the times) and he is "sort of pissed off" about it.

      He says that he is more convinved now that they are fakes

      So everyone can now go back to thinking the guy is a total whackjob.

      •  I don't think he's a whackjob (none)
        ... I think so far, he honestly doesn't know either way, and both reporters are perhaps putting words in his mouth.  The other two commonly quoted "experts", I think are whackjobs.

        But I am disappointed, to say the least, that here's a "leading expert" in this stuff who didn't have any type samples from an IBM Executive or Composer, and had to get one from Interpol after a colleague suggested it as a possibility.  WTF?  Doesn't fill me with a lot of confidence.  He also didn't know, apparently, about IBM custom keysets.

        So any expert who I can eclipse in knowledge about a given issue with a few days of Googling... well, let's just say maybe we should trade jobs.

        •  faith in the legal system (none)
          If someone ever tries to convict you on the basis of the testimony from a "typography expert" remember to contact Hunter.
        •  Hunter, Thank You! (none)
          ... for keeping the debate on the typography above the obvious, knee-jerk preconceptions.  The evidence is not clear cut, but with careful research such as yours and others seeking the facts, we should soon learn if the memos could have originated from a government typewriter in the early '70s.  Carefully defining valid questions, seeking the best data and seeing where these will lead distinquishes our quest from the wingnuts.

          ~~~~~~~~~~~~~~~~~~~ / -LONO- ® \ | Disjecta Membra | | Belligerati R.C. | \ San Francisco / \

          by lono on Sun Sep 12, 2004 at 12:28:02 PM PDT

          [ Parent ]

  •  And don't forget (none)
    we are ALL looking at digital copies.

    I could bring a word doc into photoshop and shift the baseline of all the "e"'s in about 2 seconds.

    Just because something is on your computer doesn't make it comparable to the original.

    •  Worse still ... (none)
      We are all looking at really crappy digital copies: 100dpi JPEGs.

      The fact that they are JPEGs is probably worse than the low resolution. Even at a low resolution, if a pixel is dark, you could normally assume that there was a significant amount of ink/toner at that point on the page. But with a JPEG, it might just be an artifact of the lossy compression.

      Really, unless CBS are going to release detailed scans (300/600dpi, uncompressed or lossless compression), we're all just guessing.

  •  Breaking? Staudt Had Retired? (none)
    The Seattle Times is reporting that one of the memos dated Aug. 1973, which says Killian is getting pressure from Staudt, may not be geniune because Staudt had retired from the TANG many months earlier.  Geez, what a roller coaster.
    •  Naw (none)
      I saw this a couple of days back.  It's not new, and as several people have pointed out, there is more than one way to pressure someone.

      Visit the Diary of the Lying Socialist Weasels, for commentary from the Original Progressive Web Warriors!

      by Jonathan on Sun Sep 12, 2004 at 12:05:38 PM PDT

      [ Parent ]

    •  Staudt (none)
      The memo doesn't say Staudt is in TANG.  It just says he is applying pressure.  This is a very short note with a lot of data.  If they don't bother to put someones first name in -- I wouldn't expect them to put his full title in.  

      We already know that oil men and politicians had significant power within TANG.  Why not ex-generals?

      What did Staudt do after he retired?  If you can tell me he didn't become a texas lobbyist or oil company vice president and instead went to the South Pacific and stayed on an island without telephone service -- I'll agree that wingnuts are on to something.  

    •  All that means (none)
      is that the TANG was run by an old boy network that valued its connections to the Bushes more than it valued its duty to the country.  That should be no surprise--in fact, we knew it already, since that's how GWB got into TANG in the first place.
  •  Nice work on this Diary, BUT (none)
    Why do we care.  Many documents say bush wasn't their and eye witness accounts say he used Daddy's influence to get him into the guard and get him a job in Alabama when he should have been getting a physical and going to drills on a campaign where he showed up late and hung over all the time!

    Show me a document that proved he did his service.  Until then I don't care if this one memo was chiseled in stone.

    /Kerry-Edwards 2004/ - Because America Can Do Better, Must Do Better, and Deserves Better

    by ETinKC on Sun Sep 12, 2004 at 12:14:42 PM PDT

    •  We care only because... (none)
      ...they're making the mainstream press care.  This is how they get all of their fake stories spread -- through blind and overwhelming repetition, which we do not counter because we don't want to "stoop", or whatever.

      Take a look at the front page of any of the right-wing sources, and it is full of links to various briefly-famous sites.  On the right, this is an utter sh-tstorm.  Hugh Hewitt, Michelle Malkin -- all the usual nutcases are riding the horse for all it's worth.

      Normally, I do not care what these people think, and I never care about what Instapundit or Drudge thinks.  (Actually, Instapundit has yet to post anything presenting evidence of actual independent thought; I remain convinced he is merely an Eliza program coupled with a blog trackback script.)  But in this case, the mainstream media is lapping it up with a spoon, thanks to some serious behind-the-scenes pushing from right-wing and Republican affiliated consultants.

      The next few days will provide a media consensus, one way or the other; I want to have the best possible evidence before the media loses interest, not after.  'After' don't help squat.

      We are about one day away from very, very damning evidence, evidence that will provide proof to the mainstream media that these documents were indeed created by a typewriter available at the time in question.  We only have a little way to go.

  •  Another interesting tidbit (none)
    ... from oliverwillis, a poster called Marbux:

    My outfit in Nam used IBM Executives (8th Psyop Bn). I later owned one myself after I got back to the world (I got it because I had seen it in action in Nam). I was a commercial typesetter for some 20 years both before and after Nam, so I know my typefaces. The typeface in the signature line for the 1 August 72 memo looks very much like the typeface on my old IBM Executive Model C. "

  •  Another; (none)
    this time from Atrios comments, from cervantes:

    Hate to break it to the trolls, but I used an IBM selectric composer in the 1970s myself. I'm ashamed to say where -- I was working for Ralph Nader, it was in his Congress Watch office. It was not "incredibly expensive," the office was a dump and we also used a mimeograph machine. It was very easy to use. Basically, all you had to do was type. There were a few things to learn if you wanted to do real fancy stuff, but no special training was needed. The killian memos aren't right justified and don't have any other elaborate formatting, so anybody could have sat down at a composer and just typed them. Note that there are no evident corrections in the memos, which would be consistent with use of a composer.
  •  And Another: (none)
    From MSNBC:

    A principal source for CBS's story was Bill Burkett, a disgruntled former Guard officer who lives in Baird, Texas, who says he was present at Guard headquarters in Austin in 1997, when a top aide to the then Governor Bush ordered records sanitized to protect the Boss. Other Guard officials disputed Burkett's account, and the Bush aide involved, Joe Allbaugh, called it "absolute garbage." Burkett may have a motive to make trouble for the powers that be. In 1998, he grew gravely ill on a Guard mission to Panama, causing him to be hospitalized, and he suffered two nervous breakdowns. He unsuccessfully sued for medical expenses.

    Still, in theory, Burkett may have had access to any Guard records that, in a friend's words, "didn't make it to the shredder." Fellow officers say he wasn't a crank, but rather a stickler for proper procedure--a classic whistle-blower type. Burkett was impressive enough to cause CBS producer Mary Mapes to fly to Texas to interview him. "There are only a couple of guys I would trust to be as perfectly honest and upfront as Bill," says Dennis Adams, a former Guard colleague. The White House, through Communications Director Dan Bartlett, called Burkett a "discredited source." Indeed, Bush strategists are convinced--or have convinced themselves--that the issue will backfire on its purveyors.

    •  I have a feeling (none)
      things are about to get pretty hot for Mr. Burkett.
    •  allbaugh (none)
      While we're talking about "discredited sources," let's not forget who Joe Allbaugh is:

      Chief of Staff of Gov. George W. Bush in Texas

      National Campaign Manager of Bush-Cheney 2000

      Short-lived Director of FEMA in Bush Administration

      Now Chairman and Director of New Bridge Strategies,

      A unique company that was created specifically with the aim of assisting clients to evaluate and take advantage of business opportunities in the Middle East following the conclusion of the U.S.-led war in Iraq. Its activities will seek to expedite the creation of free and fair markets and new economic growth in Iraq, consistent with the policies of United States Government.

      Talk about "absolute garbage"

  •  ebay (none)

    ebay auction ...

    Current high bidder is "john_kerry_blew"

  •  It's almost certainly not the Composer (none)
    A lot has been made about the technical possibility, or lack of it, that those memos could have been done on a Composer.  But it's just very unlikely even if it's possible.  Remember the story about Rosemary Woods erasing 18 minutes of the Watergate tape by pressing 3 buttons and a foot pedal at the same time by accident?  Yes, technically possible, but bloody unlikely.

    The memos were almost certainly done on an IBM Executive, not a Composer.  The Executive was a very common proportional-space office typewriter in the 60's before it got displaced by the Selectric.  There were so many of them made that there are still some around somewhere.

    I myself used a Composer a few times many years back, at a small company where I worked at the time.  Anyone who's used one of those things would understand that it's silly to think that it would have been used for batting out casual memos.  It was just not that kind of machine.  I guess the one I used might still be around.  I might call the owner and ask if he still has it.

    But I think you're far better off pursuing the Executive angle rather than the Composer.

  •  You know what pisses me off? (none)
    It's obvious that if you definitively prove that you can reproduce the Killian memos exactly, using 1970s equipment that survives today, all you'll have done is give the Republicans the ammunition to "prove" the memos could have been forged using 1970s equipment that survives today.  

    Like Hunter said, they'll just shift from "it's a stupid forgery!" to "it's a diabolically clever forgery!"

    They. just. lie.  

    •  But its their main argument... (none)
      They have put a lot of weight on the equipment. I think if it can be proven that a typewritter in 72/73 can be used to do this their ballon will be deflated and they will move on to something else.
  •  Selectric Composer (none)
    Ok, I found this link. I think this is some wingnut but it produces something that initially looks close to the memo...

    •  Have you even read the original post? (none)
      Hunter's very first link is to this site you 'just found'.

      WARNING: When not being directly observed, this post may cease to exist or exist only in a vague and undetermined state.

      by Democarp on Sun Sep 12, 2004 at 06:34:42 PM PDT

      [ Parent ]

  •  Info re executive; probs with wingnut comparisons (none)
    1.  The IBM executive typewriter had a split spacebar.  One side spaced 2 units, one side spaced 3 units.  The backspace went back 1 unit.

    2.  Xerography magnifies and distorts.  A Xeroxed copy is no good as a basis for measuring distances between characters, words or lines.  Try Xeroxing a ruler, if you don't believe this.
  •  "optical centering" (none)
    Thought I had seen all the arguments that the documents were faked but this is a new one.  

    This is a synopsis of the arguments but there is a new one down the page I hadn't seen anywhere else.

    In this document the author says that MS word does not center its text in the center of the page.  Word centers to the "optical center" of the page which is slightly to the left of where the absolute center of the page is.  Perhaps this a relic of how the default settings on word are set, but this would go a very long way towards saying whether word produced these.  Hunter- do you have any information on this optical centering thing?  I can't find squat about it on the net.

    •  "Optical Center" (none)
      basically refers to the artificial ability to make the document best "look centered", which isn't necessarily the same as "true center".  (The human eye is a goofy but accurate thing.)  Optical centering as an algorithm to put about half the "weight" of the letters on each side of center, so that it looks centered.  Oddly enough, "true" centering often doesn't look quite right.

      So it's an attempt to make computer-generated centering look as good as human-made 'eyeballed' centering that a good typesetter could do.  It's mainly useful when centering more than one line relative to each other, of course.

      Now, he's using that to show us that "see, MS Word puts it in the 'optical' center, and this memo basically puts it in the 'optical' center, therefore it's Word."  So he thinks that the fact they're both well-centered means it's Word, but ignores the fact that the header is very badly skewed compared to the rest of the text, and so is obviously a letterhead.  Sheesh, talk about putting the cart before the horse.

      (He's also claiming to measure millimeters to the edge of the paper to judge centering.  We don't really know what the edge of the original paper is, however.  And optical centering really only moves things a pixel or two.)

      So he gets points for an original argument, but again, he's proving Word behaves like old typewriters/typesetters, not the other way around.

  •  More info: (none)
    I'm just copying these over from The Talent Show.  Great info from Don Munsil.

    I worked in type for several years, for what it's worth, and am co-inventor of a patented font-generation system (pat 5586241, FWIW). Kerning is a red herring, as while it's true that typewriters don't kern, neither does Word, as you can verify yourself by typing 'To' and setting it to 72 point. If Word kerned, the o would tuck under the bar of the T. Try the same thing in Pagemaker or QuarkXPress, and the kerning will be obvious.

    The overlap you see with a 'v' next to an 'o' is just that some characters are intentionally designed to stick slightly outside their box, including 'v' and 'w'. This makes for more visually pleasing type.

    The font in the document is clearly not Times New Roman, though it is darned similar. Look at the numbers. Look at the lower-case 'g'. It's worth noting that Monotype designed both Times New Roman (in the 30's) and "IBM Typewriter Executive," which I would assume is the typeface used on the IBM Executive typewriter.

    The Executive was a fairly popular typewriter for IBM - I remember seeing them. They had a distinctive split space bar because they could do both 2- and 3-unit spaces. They weren't as popular as the lower-end monospaced models, but they certainly weren't some weird model only used by typesetters. And there were definitely typewriters with superscript numbers, 'rd', 'th', and 'st.' I don't know if the IBM Executive had them, but given that it was a fancy top-of-the-line model, it seems likely.

    As for the spacing and line breaks matching in Word, if in fact the Executive typeface was created by Monotype and modeled after Times New Roman, that wouldn't be too surprising.

    Overall, I think it unlikely that any document expert worth his salt would fail to notice that a document was typed using a completely unusual typeface and typewriter. I have a feeling that if you looked at type from an Executive, it would look like these memos.

    I found an example of the IBM Executive typeface, and it looks darned close to both Times New Roman and the memos in question:

    Look at the ad on the right. According to the text, it was created on an IBM Executive. To me, it's clearly in the Times family. Given that IBM made the Executive from 1947 to sometime in the 70's, it doesn't seem completely unlikely that the memos were just typed normally on an IBM Executive. The text on that web page refers to the labor involved in setting the page, but that's because the guy that made it justified the type. The memos in question have ragged-right margins, which would be easy on an Executive.

    This stuff about the special 'th' seems silly. There are certainly examples of typewriters with special 'th' keys, and it would very much surprise me if the Executive didn't have them on at least some models. Apparently some models had replaceable typebars so you could customize some of the symbols you wanted to use.

    Here's a link with the list of Monotype faces, including Times New Roman and Typewriter IBM Executive:

    And this is interesting. Here are some reminiscences of a WordPerfect old-timer talking about how important it was that word processors turned out documents that looked like they came from an IBM Executive:

    So it's not a given that they are authentic. But the idea that there wasn't any way to make proportional fonts and superscript 'th' on a typewriter of that vintage is not supported, IMO.

    The overhang on the 'f' is not kerning. The f has an advance width which is narrower than the letter itself and thus has "negative sidebearing." This means that the f sticks outside the space allotted for it, no matter what character is next to it. Without that, the f would look like it had too much space on the right-hard side.

    Kerning involves adjusting the amount of space between specific pairs of letters. Word does not have that feature and never has. If it had that feature, you would see it on the 'To' example (and on 'Te,' 'Ta,' 'Yo,' 'Ya,' etc.) Similarly, look at, say, 'fh' in Word and note how the f crashes into the h. Kerning would move those letters a bit further apart.

    You can see the sidebearings and advance width if you open up Times New Roman with a font editor and look. Many letters have negative sidebearings, but lower-case 'f' has the most of the standard alphabetic characters. The same feature can be seen on the IBM Executive, though not as strongly. This is a perfectly normal thing, and something a typewriter can do just as well as Word.

    I really don't know if the memos were typed on an IBM Executive. I don't think the font in the CBS documents matches the font on that Nibble ad exactly. At the same time, I don't think the CBS documents' font matches Times New Roman exactly. It's darned difficult to tell, though, because of the blurriness of the original documents. It might match Times Roman more closely that Times New Roman. They're different faces, though obviously related and very similar. And they have the same advance widths, which makes them break the same in Word.

    I also spent some time trying to "recreate" some of the other documents in MS Word and got nowhere. I could not get one of the other documents to break the same way no matter how I fiddled with the margin or font size.

    So are we to assume that one or more of the documents was created in Word and the others weren't?

    I'm just saying that it's not immediately obvious that the documents are forgeries. Not yet. If they are determined to be forgeries, fine. I just am annoyed by people with no experience in type rattling on about 'kerning' and the like.

    You're right and I was wrong about kerning in Word - sometime in the last few years, Word added kerning, though it's not turned on by default, and it's not turned on in the document that Little Green Footballs generated. And I don't see any kerning in the documents posted on CBS.

    Do the 'To' thing and then turn on kerning. See the o tuck under the T? That's kerning. Now look at 'fe'. Even with kerning turned off, the e tucks under the f. Turn on kerning and it doesn't change.

    Kerning has a specific meaning, and it's not that the characters tuck under each other. Kerning when the standard spacing is changed for a specific pair of characters, like 'To' but not 'Th'. This is a high-end feature, one pretty recently added to Word, I would guess, as it wasn't there when I worked in typography. And a typewriter would not be able to kern.

    But a typewriter is more than capable of having the top of the 'f' overlap the next character, just by having the advance width of the character be smaller than the character width. That's absolutely common, especially for the 'f.' Many fonts also have the tail on the lower case y stick a little into the character before it. That's normal, and happens no matter what the previous character is.

    Here's some terminology. The 'advance width' is the amount of space on the paper that character takes up - the amount the carriage advances when it is typed. The 'character width' is the actual width of the character, measured from its furthest left extent to its furthest right extent. The sidebearings are the distance from the sides of the character extent to the beginning/end of the advance width. Characters almost always have different character widths and advance widths. Usually the character width is smaller - the character fits neatly into its "box." But some characters, like the lower-case 'f', have smaller advance width than character width because they visually look thinner than their actual character width. Most of the character is thin, but there's this big overhang on top. To correct for it, in proportional type, the 'f' has a negative right-hand sidebearing.

    The powerline blog has a whole post where they rattle on and on about 'kerning' and they don't have a clue what kerning is.

    Look, like a lot of people, I saw the memos used proportional type and thought "that can't be from 1972" because I was used to monospaced typewriters like the Selectric. But having now seen proportional type from an IBM Executive, I don't think it's such an open-and-shut case.

    The fact that CBS has reasserted that they had the documents vetted and is not backing down strengthens, to me, the idea that their document experts are confident that that type could have been produced by a 1972 typewriter. The quotes in the news from the supposed "document experts" saying things like proportional type wasn't readily available on a typewriter at that time are just complete hogwash.

    Moreover, why in heaven's name would somebody forge documents that aren't really all that damning? Why not write, "Lt. Bush was unable to serve because of his ongoing drinking problem" or something really damning? As it is, there's nothing much in these documents that hasn't already been confirmed, and the implied criticisms of Bush are pretty mild.

    Here's a link to a scan from the IBM Executive manual, showing how to remove the typebars and substitute different characters.

    •  Trying to match Executive Text (none)
      Hey Hunter,

      I should have an executive in my hands this afternoon to test this, but at first glance, this sample text has some of the same problems as the Aldine.

      While the "7" is a better match to the memo than Word's Times New Roman, compare the "5".

      The "5" matches TNR more closely than the Executive sample we see.

  •  another backfired attempt to prove forgery (none)
    Look at these enlargements of the memos and 12 point Times New Roman.

    They're trying to make a point about kerning, but I don't buy it--a one or two point difference when the letters do not even look like each other?

    But this is the first view of the text that makes me go, in a glance, "yep, typewriter."

    The MS Word letters look like they are made of pixels, and the memo letters do not. The memo letters are noticeable thicker, noticeably more rounded, and there are noticeably more variations between letters. Look how the seraphs bleed over, and there is excess ink filling in the lower case e. Some of that's reproduction error, but I doubt all of it is.

    Are word processed documents that jagged and sharp-edged and uniform when you print them out, or is he working off a screen shot? If the former--I would now put money on it: these were typed. And CBS could prove it by offering some higher resolution scans & enlarging them.

    My bet is on the Executive in Press Roman. For one thing it's more common. For another the Selectric Composer Manual seems to have a lot less baseline drift than the memos.

  •  I posted this on Kevin Drum's: (4.00)
    put here just for archival completeness:

    This pro-forgery site has been referred to several times in this thread.  Let me say that I have combed the internet for evidence on this issue, which I have presented on DailyKos, and this link is quite possibly the biggest piece of crap I have yet seen.  Possibly even worse than the original LGF comparison that started it all.

    First, he uses the relative height of the 'th' glyph to demonstrate that Word puts 'th' at a certain height, and goes to great lengths to show how different it is from another CBS memo in an entirely different f-ing font.  A clearly monospaced font, to be precise, though in his brilliant deductions he doesn't quite seem to notice that.  Wow, really?  Different fonts look different?

    Then he compares the proportional font with the entirely different monospaced font and discovers -- I shit you not -- that the spacing is different.  Wow, that is convincing!

    Then he goes on to say that this proves only a computer could have created the proportional font, and lists the places with the vast computing resources necessary in 1972 to create such a wonder of science.

    Then, he goes on about how good the resolution of the (fax of the) typewritten copy is -- it's better than 180dpi!  That proves that it could only have been done on highly advanced equipment!  Why, only the five richest kings of Europe could afford such an advanced machine!  (Hey, poindexter?  What's the dpi on a typewriter?)

    He then finally mentions the existence of proportional-font typewriters, but says they could never do Times New Roman!  And besides, um, Don Knuth only printed his book on computer fonts in 1979!

    And typewriters couldn't tuck letters close together!  (Hint: Google "negative sidebearing", genius, which proportional-font machines were designed with, or just go here.  That link is especially funny because that guy says the memos don't have kerning, and the lack of kerning proves they're fake!)  And curly quotes were impossible, because, um, 7-bit ASCII fonts (used on typewriters? WTF?) didn't have them!

    I don't mean to be rude, but you have got to be f---ing kidding me.  Is this guy for real, or is this a hoax site?

    Do me a favor.  If you have good sites that have pro-forgery evidence to support them -- I would consider the shapeofdays site to be by far the best, on the pro-forgery side -- by all means post them, and I'll link to them from the DailyKos reports.  But if you can't even tell this fancily-worded shitdump for what it is, do not waste my f--ing time.  Gawd.  You owe me fifteen minutes of my life back.

    •  Response to Hunter regarding Flounder Site (none)
      H"First, he uses the relative height of the 'th' glyph to demonstrate that Word puts 'th' at a certain height, and goes to great lengths to show how different it is from another CBS memo in an entirely different f-ing font.  A clearly monospaced font, to be precise, though in his brilliant deductions he doesn't quite seem to notice that.  Wow, really?  Different fonts look different?"

      I think that you missed the point.  For days now the defenders of the memos argued that not only was superscripting available, but that it existed in one of the other Bush documents.  The document that he exposed was the one used as proof of superscripting.  And that document was clearly not a case of superscripting being used in another Bush memo.  The th in that script never rises above the 111.

      H"Then he goes on to say that this proves only a computer could have created the proportional font, and lists the places with the vast computing resources necessary in 1972 to create such a wonder of science.

      Then, he goes on about how good the resolution of the (fax of the) typewritten copy is -- it's better than 180dpi!  That proves that it could only have been done on highly advanced equipment!  Why, only the five richest kings of Europe could afford such an advanced machine!  (Hey, poindexter?  What's the dpi on a typewriter?)"

      Again you miss the point.  He never said that only a computer could produce proportional fonts.  What he said is that if we want to argue that the document was produced by a 1972 computer then we could not do so because the computer printers of that time only printed to 180 DPI.  He is not using this argument to dismiss proportional font typewriters.  In fact, he acknowledges their existence.

      "and although there were some proportional-spaced typewriters (such as the IBM Executive) and print production technologies (such as the VariTyper),"

      H"He then finally mentions the existence of proportional-font typewriters, but says they could never do Times New Roman!"

      Now you are just plain lying.  He didn't say that they could not do Times New Roman.  What he said was:

      "none of these would have produced something that was a near-perfect match for Times New Roman under Microsoft Word."

      H"And typewriters couldn't tuck letters close together!"

      The issue is not if typewriters could tuck letters close together, but if they could tuck one letter under the other.  Clearly they could not do this.  Both the author of the flounder site and the linked site that you give agree that kerning is not enabled by default under Microsoft word. And they both agree that the CBS memos do not have kerning enabled.  But the flounder site author argues that the MS Word Times New Roman font uses a characteristic of MS True Type fonts which allows for a kind of pseudo kerning.  You can observe it for yourself if you go to the 1 August 1972 memo, look at the "from" in the first item and see how the r is tucked under the f.  The flounder author shows a couple of examples from the memos which illustrate this pseudo kerning.  Basically this is irrefutable and the memos have been shown to be forgeries.

      H"Is this guy for real, or is this a hoax site?"

      Looking at some of your explanations caused me to ask the same questions about dailykos.

    •  Fifteen Minutes (none)
      You owe me fifteen minutes of my life back

      Great Minds. When someone posted that same link on TalkLeft, I clicked over and checked a couple of the characters that we have all seen were different in most of the purported Word versions — and damned if they weren't different in this one, too.

      So I hit the "Back" button on my browser, and complained: "That's fifteen seconds of my life that I'm never getting back."

      See — you should have done like I did, and simply stopped at the first incompatability you found, rather than going around trying to catch 'em all.

  •  Nail in the coffin (none)
    Well the nail is almost certainly in the coffin.

    Read email three here.

  •  1st in War, Further Down the List in Peace (none)
    I apologize if this has already been mentioned elsewhere, but I haven't seen anyone take note (pro or con) of the fact that the earlier memos never have the superscripted "st" whenever "1st" appears.  In fact, in at least one memo an unsuperscripted "st" appears along with a superscripted "th" (111th).  Needless to say, Word superscripts "1st", and in fact refuses NOT to unless you type "1-s-t-space-ctrl-Z".  I can envision three scenarios that would result in this particular set of circumstances (that don't involve Roswell and Black Helicopters anyway :)

    • The memos are fake, and our unscrupulous forger remembered to type ctrl-Z on each occurrence of "1st" but forgot all other times.
    • The memos are real, and the typewriter in use had a superscript key for "th" but not "st" (and probably not "nd", or "rd").  If you could only pick one of the four, "th" would seem to be far and away the most useful.
    • The memos are real, and the typist didn't know what key to use to produce superscripted "st".  If the typewriter used a replacable ball, it's quite possible that the keyboard would show incorrect (or nothing) characters for "specialty" characters.

    The first could be true, but IMHO it fails the sniff test...

  •  For the archives: (none)
    The military indeed had proportional-spaced typewriters during the period in question:

    1970 101st Airborne doc

    Not proof; not the same machine; not the same unit.  But still more evidence, if any more is needed, that they did have machines like the ones we're talking about.

  •  For the archives: (none)
    From the shapeofdays site, a commenter:

    DeurphMonkey -- My typesetter demonstrated the memory to me. Maybe we are talking about different models. I remember all of this since my typesetter and I were both extreme enthusiasts of the machine. Some people still have Atari computers. The typesetter enthusiasts were even more intense. Part of the reason they were so intense was that knowing how the machines operated immediately doubled your income. My involvement with the machines tripled my income, but I was also writing the manuals, not just having them typeset. Not only was my income tripled, but I didn't have to go in to work; I could work at home.

    I'm unsure what the machines were called, exactly. Selectric Composer sounds close, but I think there was a higher level machine than that.

    Satin Hinge -- "(a good assumption if it's being TYPED and not TYPESET; it's supposed to be a memo for crying out loud)"

    It's a typesetter. There were only a few of them, but they were far more influential then their numbers. That's because people were enthusiasts and because typeset documents back then were VERY influential psychologically. The woman who owned the one used on my manuals had at least a hundred customers, maybe many more.

    Many people now spend huge amounts of time customizing and exploring their computers. There were no personal computers then. Those who had access to the IBM Composer often spent huge amounts of time tinkering.

    The problem with the machines was that IBM sold them as the top of their line of typewriters. The brochures were printed on clay-coated paper with high resolution photographs. The marketing made them look like very fancy typewriters, and their output was truly excellent. They used one-use carbon ribbons, and their output was clearer than today's laser printers.

    Typesetting cost $30 per hour, and the typesetters would always make small mistakes so that you would have to pay a second or third time. If you had a complicated page, you could easily pay $90 or more having it typeset. If you used typesetting back then it began to look easy to justify buying one of those very expensive machines. (They cost about half as much as a car, as I remember.)

    On one job, a one-page advertisement for my services as a technical advertising writer, I spent 11 hours calculating the instructions to the typesetter, to try to reduce the typesetting charges. It was much easier to have someone who would just try it on an IBM Composer, and who would not charge so much.

    But, the machines were not IBM's top of the line typewriters. They were typesetting machines, and they had hundreds of little adjustments. It was difficult then to get TVs repaired, although there were a lot of TV repair shops. But TV repair was easier than becoming skilled on the IBM Composer.

    So, people with money would buy them and find that the cost of retraining someone to use them was exorbitant. Mostly, even if you had the money to pay for training, it was impossible, because anyone who knew detailed technical things was considered a freak. Women felt if they knew things like that it would destroy their chances of finding a man who would pay for them to stay home and have babies.

    The clash between the reality of the machine and the U.S. culture meant that machines would be bought and then not used. Sometimes they would be sold relatively cheaply; that's how my typesetter got hers. Sometimes one would just sit in a closet, and some enthusiast would eventually dust it off and begin trying to use it.

    When I saw the Bush documents, I laughed, because it was so obviously a case of something that we laughed about back then. Looking at the documents, it is clear that the machine was not under maintenance contract. That's because the baseline varies in the characteristic way of an out-of-adjustment IBM machine of that era. Varying the baseline in that way can be done with Ventura Publisher and Quark Express, but not with Microsoft Office 2000. I don't know about later versions of Office.

    There were reasons IBM machines were not under maintenance contract. One could be that the machine was stolen. IBM had a number you could call to ask if a particular serial number was stolen. Maintenance contracts were very expensive. If someone called for one-time maintenance on a stolen machine, the machine would be confiscated by police.

    Of the other reasons someone would have one of those machines, the more common one was that it was rescued from a closet. Those who had decided not to use one would be glad that someone would take it out of their sight because it was an embarrassment to make such a big purchasing mistake. The rescuer would be an enthusiast, of course, and would soon become very popular with those who wanted some document to look nice after it was printed. "Church Bake Sale" for example.

    But, the rescuer would certainly not have the maintenance contract. If I remember correctly, the contract cost as much as the rent on a cheap apartment. So, the machine would slowly get worse and worse, until it was so bad that people who specified typsetting ("spec'd type") like me would reject the output from it. When you paid someone to do the work, and you had spent hours specifying the typesetting dimensions, you were extremely sensitive to everything. That's what's so funny about the machine used to do the Bush documents. It is way out of adjustment in that characteristic way.

    I was never in the ANG, I was in the regular Air Force. USAF offices were dreary, usually. They all had the same grey desks and the same office equipment. But sometimes you would walk into an office and see some equipment that didn't belong. There were often circumstances where people would have things that were not standard issue. I mostly repaired the wasp-wasted F-106s (Delta Dagger), but sometimes helped repair straight-waisted F-102s (Delta Dart), the plane George W. Bush flew. When I discovered a more efficient way to do a particular repair, I had no trouble getting a professional photographer to take photos. Other people who wanted to build electronic projects would just requisition the parts. If you could find some equipment, and no one else was particularly wanting it, you could have it, and you wouldn't have to pay, of course.

    It's not impossible that some general's office requisitioned an IBM Composer and then found that no one would in the office would learn to use it. It would be just an everyday occurrence if a clerk in the local ANG office visited a clerk in the general's office, and asked if he could "borrow" the dusty IBM Composer sitting in a closet. The movement of goods and services sometimes flows very freely in a military culture where the taxpayer is paying, and the participants don't pay. Read the novel by Joseph Heller, "Catch-22". Someone told me to read it, and I was 2/3 of the way through it before I realized it was supposed to be funny.

    Culture, #1 -- It's really important, I think to visualize the culture at the time. George W. Bush admitted he was an alcoholic back then. But back then, there was very little understanding of alcoholism. There was little negative feeling toward someone habitually getting drunk. It was, to a lot of people, just what men did. If you know active alcoholics, you know that they often don't meet their commitments. If you know active alcoholics, you know that what is covered up is ten or a hundred times more ugly than what is talked about. George W. Bush could be be "famously obnoxious" at parties. ( Laura Bush told her husband she was thinking of leaving him. Underneath that are a lot of ugly stories, it is easy to guess.

    Culture, #2 -- Back then the average level of education was less. There was much more acceptance of some kinds of government corruption (and less acceptance of others). It was understood that those with connections would get special treatment.

    Culture, #3 -- Someone I knew in the USAF told me he was called out on a weekend to repair the automatic flight control system of an aircraft. They used gyro boxes then called SCRG, Stable Coordinate Reference Group. The entire SCRG gyro assembly cost $22,000, if I remember correctly. Everyone knew that; it was the most expensive single part in the flight control system. The airman did not want to work on the weekend. He called base supply and asked them how many SCRG's were available. They said one. He got it and placed it on the edge of the wing. Then, he told me, he looked around and made sure no one was looking. With a small movement, he pushed the SCRG off the wing. If fell to the concrete. He told the crew chief that the repair would need to wait until Monday, when another SCRG would be available.

    What is even more remarkable about this story is that the airman would tell me the story. I repaired the electronic boxes; I didn't work on the flight line. The fact that he would tell me shows the casual disregard that sometimes occurred. The point of this is that it is easy to overextend logic when trying to think about cultures with which you are not familiar. Yes, things were done by the book, except when they weren't.

    Sure, some really, really smart forger who had the same experiences as I did could have forged the documents. But why? Just to impress me? Why would a smart forger not choose the Courier font? One thing this discussion has taught us is that few people know much about typesetting and many people are willing to make conclusions on too little evidence and with too little logic.

    My typesetter would type everything on her machine, even notes to me, like "Michael, I changed the spelling of a word on page two. I hope I was right." It was a shock to see myself seemingly immortalized by having a note to me typeset. Typesetting was a big deal back then, and VERY psychologically powerful.

    The only way to test the Bush documents is by trying different combinations on an IBM Composer, with the same type ball. Remember, they were typesetting machines. Yes, they followed the extremely rigid 1932 design of Times New Roman. However, they also had ways to change letter and line spacing.

    Nothing I have said here proves anything definitively. However, when I saw the documents, I laughed and immediately thought that the documents were genuine because what forger would know all that I've said here?

    In the interest of full disclosure: I don't think George W. Bush should be president. I once asked three recovered alcoholics if a recovered alcoholic could be a good president. One said no. Another said something like, "Of course not, that's crazy." The third said, "I'm very competent at my job, but no one like me should ever be president."

    Whether George W. Bush should be re-elected does not depend very much on old documents. There are literally thousands of incidents of more serious recent corruption. For example, see the article "Unprecedented Corruption: A guide to conflict of interest in the U.S. government" ( More disclosure: I wrote the article.

    Posted by: Michael Jennings | September 13, 2004 04:44 PM

    •  hunter? (none)
      Haven't seen you post on this in a while.  Do you still think the documents are authentic after all your research?
      •  Sorry, I'm here... (none)
        I do think they are authentic; or at least I'm convinced that there is still no evidence of forgery.  As for actual proof either way, we're stuck in a holding pattern at the moment.

        I should follow up with another diary that explains the current state of things.  Look for it maybe tonight, or tomorrow at the latest; right now I have a crushing work deadline that has taken me out of things for the moment.  Aack!

  •  Hunter (none)
    I've just reviewed the documents for the past hour, and extra time didn't really help that much, as some glaring anomalies seemed to stand out after just a few minutes.

    1.  The 'th' specialized superscript (when used) rides much higher than Word.  The documents were not typed in Word (though perhaps some other word processor or software was used to forge).

    2.  There is some unusual spacing going on with the non-specialized superscripts as a whole, some having a gap between number and superscript, others not having a gap at all, within the same document.  I'm not sure how to explain this.

    3.  One sentence in particular has two different space/gap characteristics with an apostrophe.  Word does not do this.

    4.  Several words like MEMORANDUM seem to have drastically different spacing characteristics than Word.  

    Now, I'm not an expert, but I've seen a lot of Word documents, and I am confident there were definitely not typed in Word.  The 2 different apostrophe spacings in the same sentence would tell you that alone, as would the 'th' specialized superscript riding too high.

    Keep in mind also that to explain the 'th', one would have to suggest that any automatic formatting would have to be alternately turned on and off to explain the presence of specialized and non-specialized superscript th's in the same short document, and this is highly unlikely for a forger (at least a forget who doesn't want to get caught).

    For instance, one of the memo's has the specialized 'th' superscript in the header, while a different memo doesn't have it in the header, but does have it down in the body.  In every case, the 'th' is definitely not from Word, as it's not in the right spot.

    Since Hodges seems to have verified the substance and content of the memos, as sounding about right, his contention now that he assumed they were hand written doesn't impact that.  If anything, you could say he was tricked into confirming what the memos purport to say (that Bush got special treatment).

    Still, the memos just seem too inconsistent, in the varying superscripts, the varying spaces between numbers and characters in superscripts (at least 3 different gaps), in an error where there is not a space for the next word after a parenthesis.

    "I'll backdate but won't rate" seems particularly off, from comparison to my retyping in Word, because Word seems to put a 1/4 space between an apostrophe and the preceding character.  If you look at memo #6, "won't" conforms to that, but "I'll" seems way off.  

    The varying spacing with these 2 apostrophes in the same sentence tells us much.  Word was not used to type the document, and that this obviously was not something that either a forger or a typist would manually seek to do.  

    So, look at the superscripts and the apostrophes, all of them, and if you can find a device that would explain them, then you have the right device.

    Otherwise, the only other option seems to be that there is a forger who wants to get caught, by being wildly inconsistent, but that doesn't seem to play well for anyone, as the content of the memos was confirmed, for the most part, by Hodges.

    free the information

    by freelixir on Mon Sep 13, 2004 at 10:49:21 PM PDT

    •  a slight correction (none)
      I'm standing behind the points made above, except for #3.  When examining the Word overlay document the freepers are touting, I bumped up to a font size 14 and the apostrophes shifted a bit.

      Though they were still not exact, they were close enough for me to back off the claim, though that whole sentence seems weird, in that the letters seem blocked off rather than true to the Times Roman font.

      All the rest I'm sticking with, including failure to put spaces after an end parenthesis (or proportional spacing glitch), the wildly varying 'st' and 'th' designations and spacing used with them, and especially the 'th' superscript which consistently, on different memos, seems to be riding too high (MUCH too high) than the default Word behavior.

      I don't know if this can be modified in Word, so as to alter settings that dictate where a superscript will lie, but if it can then that would take some of the steam out of this charge too (in some ways).

      Overall, these documents were definitely not typed using default Word settings, though it's always possible that someone tweaked and manually manipulated the document in Word so as to make it a more credible forgery.  

      Just as it's possible to take an old typewritten document in a standard font and reproduce it in Word almost exactly.  

      Finally, just eyeballing the documents, it seems there are spacing anomalies that are clearly less clean than Word, and seem more akin to a typewriter.  I've worked with Word for over a decade, and I'm not sure I've ever noticed lines that appeared to have characters at slightly different vertical positions.  The memos seem to have this, but it could be due to the poor quality of the PDF's I've been looking at.

      We must also keep in mind that even if Word, or Word with default settings, turns into a red herring, the document still could have been forged as it should have been if forged - with a typewriter.

      And the forgers would have to a lot of inside information on the Texas ANG to know the names, the circumstances, the dates, and even the particular terms (if even the abbreviation is claimed to be off by a character), in order to trick Colonel Hodges into confirming that this is consistent with Killian and his beliefs at the time.

      Hodges did say that, and he has only later said that he thinks the memos are forged, since they're typed, and has specifically tried to retract confirming the bit about 'sugercoating'.

      Meanwhile, the personal secretary for Killian seems to have confirmed that Killian did have his memos typed up, but not in the typeface of the CBS memos, and that their content is consistent with what was going on and Killian's dealings with Bush.

      Unfortunately, I'm pulling the information about Killian's secretary from an article I don't have registration to, so if someone who posted a blockquote earlier at Atrios is messing with me, then I may have the personal secretary information in the article all wrong.

      free the information

      by freelixir on Tue Sep 14, 2004 at 04:03:42 PM PDT

      [ Parent ]

      •  There's a diary (none)
        here about the secretary.  Very curious, indeed.
        •  The secretary thing is cerainly interesting, but (none)
          she doesn't know if the memos are forged or not.  What, she never took a day off, followed Killian everywhere he went, typed every single thing he ever had typed?  The only think she can know is whether or not she typed them.
          •  methinks it would be more likely (none)
            That Killian really covered his ass and, either around the same time, or later, exercised true CYA due diligence and had his memos reproduced outside the unit.

            Personally, I would have done the same.  That way, if the shit goes down, I have 2 copies, and can even if necessary play hardball if my private copy no longer exists in the official files.

            A different secretary is reasonable, but not if there was no machine on site that could format the way the memos did.  I'm sure Hunter will look into that, and we'd have to be able to find other memos with that formatting.  If I know the cronies, there are no such memos or documents available, if they ever did, because they were 'accidentally destroyed' any number of years ago.

            free the information

            by freelixir on Tue Sep 14, 2004 at 08:49:19 PM PDT

            [ Parent ]

    •  The th superscript (none)
      "The 'th' specialized superscript (when used) rides much higher than Word.  The documents were not typed in Word (though perhaps some other word processor or software was used to forge)."

      False.  Type something using the th superscript and then print it.  Notice that the printed th is higher than the screen th.  The printed version will match the height of the CBS memos exactly.

      •  I'll have to check that out (none)
        I don't get that on my print preview.  But the 111th superscripts are much higher than the single 187th.  So unless that's a reproduction error, that seems odd.

        free the information

        by freelixir on Wed Sep 15, 2004 at 01:20:34 AM PDT

        [ Parent ]

      •  looks like I may change my mind on this (none)
        You're right.  The print seems much different.  Interesting.

        So, I'm down to thinking it's not Word because of all the inconsistencies, and why some stuff is superscripted, and other stuff isn't.  Also, why the 'th' is only superscripted off and on.

        This would have to be toggled in Word to accomplish.  Thus, it would have to be intended, and not done accidentally.  Why would someone intend to do this?

        Still, I would not be surprised if it turned out to be a Word document, or from similar software, or a word processor, rather than typed, if I had a better copy and could reduce my concerns of a wandering baseline and other spacing anomalies (reducing noise).

        I tend myself to take Occam's razor the other direction on this matter - I've yet to imagine how someone would have chosen to forge this on Word or a modern word processor (rather than typewriter), and why they would alternately superscript 'th', and not any others like 'st' and 'nd'.

        That's a hard hurdle for me to jump, and has played into my doubts.  If anything, such a hurdle has at times suggested to me that the document is a plant and intended to be discovered as a smear.

        But I don't believe that anymore.  Mark me down as not knowing what to believe, but I'm no longer convinced that it couldn't be in Word, though I'm still open to any number of other possibilities.

        free the information

        by freelixir on Wed Sep 15, 2004 at 01:27:59 AM PDT

        [ Parent ]

        •  The positioning of the 'th' (none)
          isn't too much of an issue either way.  Word's onscreen implementation of the font puts it way up, but nearly all other (Postscript printer) implementations put it farther down.  I'm inclined to believe all the non-Word versions are the more correct implementation of the original 1930's font.

          It's being used as a red herring by both sides; we know superscripted 'th' existed back then, but we haven't found the right typewriter font, so we don't know if it actually would match.

  •  Probably not needed for anything, (none)
    but I downloaded the USA Today .pdfs on Friday when they were 485K instead of the 88K there now. They are still fuzzier than CBS's but lack the obvious compression artifacts around the text. If they would help in any examination you are doing and you have a place I could upload them to, I would be more than willing to do so.

    I swear the extra 2 memos were most likely "hidden in plain sight" on USA Today's web since Wednesday evening. And what putzes USA Today has been on this - just hanging out with the memos (their original story was a bit ambiguous on exactly where they came from), letting CBS take the heat, not adding to the story, testing the wind and then chiming in on Sunday with a fairly negative story. If I understood the CBS piece tonight, CBS did not have the extra 2, so if USA Today had had any guts they could have advanced the story some on their own. Too bad since I think relative to many others USA Today has been pretty good on this election in general... but that is faint praise indeed.

    Good luck if you are still working on this. Your info has been the most consistent I have seen.

    Save for the utter focus on the documents in the media - I think they would now rightly become a secondary story - even without clear provenance established. After all they are just another expression of the motives and actions of the principals - and CBS has effectively used them to confirm via Hodges, Strong, and by lack of denial, the WH, that the contents are basically valid. Imagine for a second that they were indeed forged but based on good info - CBS would now have used them to effectively leverage more/confirming info from primary sources. An old trick, and if performed by the other side it would have been hailed/reviled as a stroke of Rovian genius (and gotten very different press treatment.) So I relished CBS reiterating the direct content-based questions at the White House tonight - the WH knows this stuff is true, but to me clearly don't know what CBS has as a hole card (maybe nothing, maybe something on why it all happened) - not a very comfortable position for the WH.

    Of course this is a fascinating poker game, but the tragedy of it is is that there is going to be a real freaking election pretty soon (... well, in 49 states at least) and the outcome matters.


    •  Thanks, (none)
      I have the better 488k copies too, but thank you for pointing that out.  The original CBS pdfs seem to be better still -- but still not very good, compared to what we need.

      I'll post something in the next few days that summarizes where we're at.  As freelixer says above, the closer you look at the docs, the more obvious it is that they didn't come from Word, in particular.

  •  For the archive: (none)
    Howard Kurtz, what a surprise, weighs in at the Washington Post in an article that is resoundingly critical of the documents.  No new information in it, other than to note that the author of the 'flounder' site, mentioned above, has been elevated to the mainstream.  They also talk to someone from Adobe who doesn't think it's an IBM.
    •  what's funny about this (none)
      Is the constant moving target.  In the beginning, the proof of forgery was proportional spacing, Times New Roman, and the 'th' superscript (whether any of these existed at the time).  All of these have been shot down.  They did exist.

      Without hesitation, we then got kerning and the likelihood that Killian would have had access to a typewriter with all of the above.  Competing explanations have come across, and there has been no confirmation of this by either side of the debate.

      So, did Kurtz notify his readers that the vast preponderance of charges of forgery by the celebrated Right wing blogs have been shot down and shown to be silly?

      No.  He acts like every charge is still ongoing, at least the momentum ongoing, as freepers and Bush cronies take AK-47's and spray machine gun fire on the many moving targets they throw up.

      And finally we're left with stuff like the abbreviations are off by a character, which hardly would explain how someone would know the right terminology and lingo in the first place, and know to abbreviate them, and only then get it wrong!

      But there's something to the new charges, as they purport to be comparisons of past Killian documents, so we'll have to wait and see.  And the personal secretary apparently is claiming to have typed up his memos, including CYA memos, and claiming the type face is wrong, even though the content is right (I'm going on a sketchy quote from the article, however, as I'm not registered, so buyer beware on this until confirmation with Dallas Morning News).

      It seems almost as if there's a high stakes game going on right now, and Rather is very confident in his hand.  What's he not showing?  

      free the information

      by freelixir on Tue Sep 14, 2004 at 04:14:00 PM PDT

      [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site