I wasn't paying attention to type when I was four, but I do know that that blackout is working about as well as US foreign policy.

hhp

It seems that ABC has picked uup on this now. And the NY Times.

I'm with Allan - when I saw the doc on the little green footballs site the superscript 'th' just didn't seem right for a document produced in the 70's

Cheers, Si

Some experts defending CBS are pointing out,
correctly, that while most typewriters in general
use in the early '70s did not use proportional
spacing or superscript ordinals (raised "th"),
typewriters were produced for business and
government use with these very features since
1968, including machines set with proportional
Times New Roman characters.

CBS further reports that these same typographic
features also match some of the documents
that the Bush campaign itself released earlier
in the week.

David

As david mentions, there's some technical debate.

The bigger issue is, if they are fake, was it the dems or reps that seeded them and why?

The biggest issue is why is this a major issue of this campaign to begin with? ;o)

It has less to do with the campaign and more to do with the idea that CBS could be fooled by such obviously fake documents.

There goes the Republican's Media Conspiracy official line again. Little people with little concerns. As if there's a huge difference between the two parties to begin with. Freedom of choice my finial.

hhp

"It has less to do with the campaign and more to do with the idea that CBS could be fooled by such obviously fake documents."

Well, as reports are showing, it's not necessarily obvious. They may not even be fake. Either way, CBS should have done more background on ANY 'evidence' regardless of whether it looks fake or not.

The documents look like pretty blatant forgeries, but that has not yet been proven absolutely. What *is* absolutely certain is that they were not produced on the IBM Selectric Composer or the IBM Executive, the two machines most often touted as possibly producing the documents. I have done experiments that prove this beyond any reasonable doubt (while simultaneously supporting but not proving forgery).

See this morning's Washington Post:
http://www.washingtonpost.com/wp-dyn/articles/A18982-2004Sep13.html


Clarifications I sorted through last night to my bit of the article:

1) Adobe did the digital version of Times Roman from which the Microsoft/Monotype Times New Roman borrowed its widths to use with the Monotype design. Adobe got said widths from Linotype. I did mention the various other players, but what can you expect when an entire interview is boiled down to a couple of sentences?

2) Although certainly a run-of-the-mill Selectric could not have produced the memo, more importantly I eliminated the IBM "Selectric Composer" (often just called an "IBM Composer"). This was an actual typesetting device based on typewriter-like technology, which has frequently been touted as the likely source of the memos.

3) The IBM machines used fonts of different widths than the font used for the memos. The font used for the memos is a perfect match for Times. The IBM machines were not capable of producing a perfect match to those widths. Times is based on a system of widths of 5-17 units (at 18-to-the-em). The Composer uses a system of 3-9 units, and the Executive a system of 2-5 units.

4) The reporter didn't tell me of Glennon's statement, so it's funny to say that I disputed it. Indeed, I won't dispute that the IBM machines could produce results "similar" to the disputed memos. But they cannot reproduce the exact relative line lengths, which is what I measured. Can they do something "similar"? Sure. Could they produce those actual memos? Impossible.

5) Of course Adobe's HQ is in San Jose. I am at our Seattle offices.

6) I've done this solely on my own time. Adobe as a company takes no position on the memos (or the election).

That's about it for now. Here's the first visual aid I produced (attached), along with the original memo for comparison.

Cheers,

T

1. Where is the 'visual aid'?

2. Some of the discussion of this issue has mentioned the problem that the several generation-removed copy has introduced distortions. Also that some of the glyph proportions, such as the 8 are different from TNR. Is your interpretations of the measure of the lines affected by the repeated copying? What is your view on the issue of the proportions. (see the discussion at typographi.com)

1) As a courtesy, I'm going to give ABC News and my blogging writer friend Jules Siegel first shot at showing the "visual aid." I'll post it within another day or so after they've had their chance.

2) No, all I care is when the lines end, what letter is above/below what. The cumulative effect of the advance widths. This is practically immune to the issues of distortion from repeated copying.

It's possible that the font is something other than Times or Times New Roman. If so, it exactly matches the advance widths of Times/TNR, and it's an interesting coincidence that virtually all the letterforms are close enough that differences could be attributed to degradation from repeated scanning/copying (like being faxed, then being scanned and having the scan dropped to ~120 dpi).

Here's the short version of my analysis:

SUMMARY

The incredibly bad reproduction of the memos makes it hard to state many things definitively. But one thing that is not degraded by the reproduction is the simple question of relative line lengths. Where does each line end, relative to the lines above and below it?

Given proportionally spaced fonts, and a large enough sample (as these four memos are), these line breakings offer a sort of digital fingerprint of the widths of the font used. The memos precisely match current digital versions of Times (and previous phototype and hot metal typesetting versions), but they do not match the IBM Composer fonts (1), or do they match any version of the IBM Executive (2).

DETAIL:

This is not overly surprising, since the number of possible character widths was much more limited for the Composer (all characters were 3-9 units wide) and the Executive (all characters were 2-5 or perhaps 2-6 units wide).

It is worth noting that the digital versions of Times available today are all based on an earlier 18-units-relative-to-height ("to the em" in font-geek-speak) system, with common characters being 5-17 units wide. So while they have relatively discrete widths for common characters, these widths are at a "finer grain" than early typewriters or low-end typesetters of the 70s such as the Executive and the Composer.

Today's digital versions of Times have widths that descend from those used in Linotype's phototypesetting and earlier hot metal versions. Monotype had previously had a (more original) version with different widths, but when Microsoft licensed Monotype's version, they wanted it to be compatible with Adobe's, so the widths were changed to match the Adobe/Lino versions. Thus all the main digital versions of Times used today have the same widths.

(1) The IBM Composer proportional fonts all had the same relative character widths, regardless of font design. Thus there is in essence only one "fingerprint" for the Composer fonts. Times matches the memo fingerprint, but not the Composer fingerprint.
Contrariwise, I made a digital version of a Composer font (since I have the widths info). This allowed me to do "virtual Composer" simulations and prove in the reverse direction, that the relative line lengths set with the virtual Composer are quite different. (Note: My Composer simulation font was accurate to the nearest 1/1000 of the point size for each letter.)

(2) The IBM Executive did not offer switchable fonts, so you literally had to buy a different typewriter to get a different proportional font. None of them is particularly close to Times.

Special Thanks to all the people who have helped me in my research:
- Gerry Kaplan on the IBM Composer
- Norm Aleks on the IBM Executive and the Composer
- Frank Romano and Jack Powers on other possible devices
- a retired Adobe person who verified the origins of the Times metrics
- Jules Siegel for pointing me at many blogs and arguing with me enough that I realized how visual I needed to make things

And thanks also to Carol Wahler for putting Newsweek in touch with me and Adobe PR for putting the Washington Post onto me.

I am of course solely responsible for my opinions and conclusions, which are not necessarily the same as anybody else's, and do not reflect any particular views of my employer.

Although I don't feel any need to promote my personal political views, those who know me well will be well aware that my politics obviously had no influence on this analysis.

Are there no circa-1970-era typewriters with proportinal fonts that someone could use to just type the damn memo and compare it to? ;o)

Darrel, good question. The secretary who typed the memos for the person in question - she is now 84 - is quoted in the Washington Post today as saying that she didn't have a proportional spacing typewriter. She does think that the faked memos accurately reflect her late boss's views, and apparently she despises Bush.

It only took me a couple of hours to make my "Virtual Composer" font. This was quicker than tracking down somebody locally who owned one.

The only real *typewriter* that anybody has identified with proportional spacing is the IBM Executive, which as I said previously isn't even close on several levels.

There are three typewriter-like typesetting machines that were around at the time: the IBM Selectric Composer (NOT a regular Selectric), the AM Varityper 1010 and the Friden Justowriter. None of them make much sense for typing internal memos, but the folks who want the memos to be authentic seem to be immune to probabilistic arguments. So I thought I'd use pure logic and spacing info.

Cheers,

T

Glad that the line length argument makes sense to somebody besides me.

Yes, we did do some stuff with font substitution matching metrics of fonts. Of course, the limitation is that the incoming font metrics have to be already be digital.

T

> we did do some stuff with font substitution matching metrics of fonts.

Not past tense. It's built into the PDF reader. But I wouldn't call that "parasitic".

hhp

Christian you'll hafta finish your thought... Email reply chops the ends off of posts, unfortunately.

Hee hee, I was just watching 60 minutes, where they were interviewing some woman somehow connected to Bush's Guard service... At one point they quoted from one of these memos, and had the quoted material on the screen in big letters. What font do you think they used?

(Hint: rhymes with "American Typewriter")

And just following 60 Minutes tonight I had the great misfortune
of catching the first few minutes of Hannity and Colmes. Who
the F' is Joseph Newcomer?

Thank you, Thomas, for doing your analysis. I got caught up in the discussion over at Typographica and posted some graphics showing that the memos matched the metrics of Times/Times New Roman, but with the small amount of information about the old IBM typewriters I had, I couldn't take it any further.

Incidently, in light of the fact that a lot of us here, I'm sure, are not big fans of Bush, I'm reminded of this joke:

Three people were going to the guillotine, and the first was a lawyer who was led to the platform, blindfolded and his head put on the block. The executioner pulled the lanyard and nothing happened. To avoid a messy lawsuit, the authorities allowed the lawyer to go free.

The next man to the guillotine was a priest. They put his head on the block and pulled the lanyard, and nothing happened. The blade didn't come down. They thought it must have been divine intervention, so they let the priest go.

The third man to the guillotine was an engineer. He waived his right to a blindfold, so they led him to the guillotine and put his head on the block. As he lay there, he said, "Hey, wait. I think I see your problem."

Joe, I think the point is that many people who want to see Bush out have a natural reaction to this story which is:
'Who gives a flying •••• through a ring donut whether the documents are fake or not? Its not as if they are the deciding factor between whether he's a fit President or as much use as a sh1t flavoured custard cream. The whole thing is irrelevant to the real issue, which is his record as a president'.

On the other hand, as typophiles, they can't help being curious and wanting to get to the bottom of the interesting technical issues surrounding the attempt to validate historical documents through typographic knowledge.

Its not the sussing out of the truth that's the thing that the anti-Bush lobby want to avoid, its treating this red herring as if it were relevant.

Mark, apologies if that wasn't your point.

While we're on the subject of lawyers in jokes, and as I'm in the middle of protracted legal proceedings, here's another:

A guy goes to see a lawyer and asks 'What are your rates?'
'$500 dollars for three questions' replies the lawyer.
'Wow, that's expensive isn't it' says the guy.
'Yes' says the lawyer, 'What's your third question?'

I laughed at Mark's joke. ;o)

"Its not the sussing out of the truth that's the thing that the anti-Bush lobby want to avoid, its treating this red herring as if it were relevant."

The current GOP 'marketing' team is very shrewd. They understand that you don't have to be 'right' to win the debate. As long as you muddy up the topic enough, the debate becomes irrelevant to any real issues. I don't like what the GOP is doing, but have to give them credit for knowing how to market themselves via the system.

Hey don't cha think there are already too many variations called "Times New Roman"...

Let's call this one "Karl Rove".

That's a great joke, Mark. But are you saying that people shouldn't try to suss out the truth if it might favor someone you disagree with?

No. It's just the irony of the situation that made me remember the joke. I don't necessarily endorse the implications of it.

I don't know why everyone assumes if it is a fake that it was done by a Democrat. It seems more likely to be a ploy by the Right Wing to discredit the "Media" and take the focus away from the real issues. It seems to me that it is too obvious a fake. It had to be unintended for discovery. The "fake" story came out REALLY quickly as if planned.

ChrisL

The entire issue is, as Steve said above, a red herring. (Although the typophilic aspect has a grain of interest--I don't remember proportional typewriters at all.) It's perfectly clear that Bush avoided combat and got preferential treatment, but it doesn't make a whit of difference compared to things that matter. The real issues are economy, environment, health care, terrorism, war, energy, corruption, equality, civil liberties, etc., not what these guys were doing more than 30 years ago. There has been a terrible decline in the quality of journalism, and what passes for news and commentary these days is a joke and a scandal.

I'll be the first to agree that this is not actually "important" and is a distraction from real issues in the election.

That being said, it's kinda fun to see typography making front page headlines.

The Washington Post published a front page article Sunday. I'm quoted again, near the end.
<http://www.washingtonpost.com/wp-dyn/articles/a31727-2004sep18.html>

with accompanying graphic <http://www.washingtonpost.com/wp-srv/politics/daily/graphics/cbsdocs_091804.html>

CBS News is finally really starting to back down on their story. <http://www.cbsnews.com/stories/2004/09/06/politics/main641481.shtml>

Cheers,

T

Big interview on CreativePro.com discussing my approach in depth, with more graphics (don't miss the linked PDF if you're looking for visuals).

http://www.creativepro.com/story/feature/21939.html

Never before have so many fonts mattered so much to so many...

Font Politics:
http://www.sfgate.com/columnists/fiore/

Aha! Now it's all making sense:
The memo was forged through a secret agreement between ATypI and SoTA, in order to raise font consciousness among the rabble.

hhp

Not to beat a dead horse, but I thought this was an interesting article on "memo-gate". It links to a research project report by the Director of the Interactive Media Research Laboratory at Utah State University. Unfortunately, the images for that page don't load.

http://www.bluelemur.com/index.php?p=324#postcomment

You can download the PDF of the document from the IMRL, and that works fine.

Unfortunately, Hailey's artice reflects a poor understanding of fonts. He identifies the font as one dating to 1905 called "Typewriter," and then shows an example--of a monospaced font! It's not "the same" font if it has been adapted to proportional spacing and the letter widths have been dramatically altered. Hailey dismisses my arguments with attacks on my starting with Times New Roman--but admits he isn't an expert on widths, and indeed it's apparent to me that he doesn't understand my argument. He makes arguments about how certain kinds of inconsistencies couldn't possibly be caused by scanning degradation--but obviously he didn't run the right kinds of tests, because when I ran tests using Times and involving faxing, then scanning back in, using the resolutions in the posted CBS docs, I got results that looked very much like the CBS docs, including things like the cupped serifs.

Finally, his close matches to the actual memos were achieved by taking his (unclear what version, unclear on what basis he chose widths) typewriter font and condensing it (unclear what method). Guess what: proportional-spacing typewriters in 1972 didn't have a special button for "Condensed" like you can do in Word. He'd have to find an actual font that matches, not invent one from whole cloth.

Gotta run, must do some real work.

T

It is startling how much satire seems to get to the heart of the matter so much better than "news".

http://www.theonion.com/news/index.php?issue=4039&n=0&id=3839

Actually, to be fair to Hailey, it's not at all clear that he has read a full version of my analysis, or even that it was my analysis in particular that he was criticizing. If he's just going by what he read in the Washington Post, then he doesn't have a very detailed analysis to respond to--by anybody.

That being said, his belief that (for example) the "t" shows considerable wear is based on it initially matching his typewriter font. The supposed "wear" alters it so it looks remarkably like... the "t" in Times. It's a much simpler explanation.

All that being said, the best defense anybody has mounted so far is the possibility that these documents were reproductions, created by the TexANG's own scanning and OCR of the records, then printing out again. Of course, it would require considerable manual intervention to do that and retain signatures that overlap the text. And no other records that have been produced match the dubious records in terms of the font used.

Cheers,

T

I'm glad to see someone articulately describe the line length argument. This was one of the first things that stood out to me. Obviously, the letterforms themselves are mangled beyond recognition, but the cumulative effect of the metrics is much more apparent. Even the slightest differences in character widths when multiplied by hundreds of characters will affect copy flow. Ask any middle schooler who has typed a paper: choose the right font, and a nine page paper becomes ten--and for the most part, your teachers will still think any serif font is Times New Roman (along with most of the members of the press). Obviously the the differences here are more subtle, but the principle is the same. It's really tricky to match copy flow exactly without the exact type, because the metrics will vary slightly, even from vendor to vendor for a particular face. Some notable exceptions are "parasitic" fonts specifically designed to replicate metrics, such as Arial, allowing for substitution without changing copy flow.

I still had some doubts about the forgeries with all of the old Selectric Composer guys coming out saying "typewriters could do anything computers can as early as 1492", but my suspicions were confirmed when I started to see accurate descriptions of Composer metrics. I commented on this over at Typographica. (By the way, what's up with typophile going down the weekend typography was being discussed in every news venue?)

Tangent: Didn't Adobe do some stuff with "universal fonts" at one point that would replicate font metrics to replace document fonts without changing copy flow/document layout?

I wouldn't call the Adobe Reader tech parasitic either. It's very useful, in fact. I was referring to the practice of marketing your font as a cheaper replacement, able to slither in to documents where it's competition was originally spec'd without 'changing the design'.

> Yeah. I fixed it. Wouldn't mind if typophile ditched the old > discusware.

That's a great joke, Mark. But are you saying that people
shouldn't try to suss out the truth if it might favor someone you
disagree with?

In effect, what I hear is people (not here necessarily, but on the
news) saying, "Well, even if the documents were forged, we still
think they're right."

Sorry Mark. I know better than to make a joke into more
than it's intended to be.

Delve! That is hilarious! Thanks for point it out.