I am a professional writer. By that, I mean I earn my living by writing. Writing on the subjects of this Web site. I am also a freelancer. I am not employed by anyone. I write stuff and sell it. Somewhat more than a million words over the past seven years, I estimate, of my writing has been published in a wide variety of newspapers and magazines.

I am not a journalist. When people use that as a shorthand for me, I normally disclaim it. There are several reasons for that, including the fact that I have no formal training as such, nor am I a member of any professional journalist associations. I join very little. I value my independence too highly.

But, also, I do not do most of what a journalist does, which is 'report'. What is 'report'? It is tell a story using other people's words. If a reporter starts using his or her own voice, he or she is creating an opinion piece, not a report. I have, from time to time, by request, written reports. But I don't like doing so. I far prefer to write using my own voice. My work is almost exclusively opinion pieces. I think this or that sounds good. I reckon LCOS technology is better than LCD. And why I think so. A reporter, on the other hand, would quote some professor or industry participant to say that LCOS is better than LCD.

Far too often, the reporter him or herself simply doesn't know if it's true or not. They are in the hands of the experts. In my field, I am the expert.

This is highlighted by this article from the Columbia Journalism Review, discussing the 'Rathergate' scandal. For those who sensibly confine their interests to matters of home entertainment, this concerns a report on a US weekday edition of the CBS Network's '60 Minutes' (sometimes known as '60 Minutes II') in September 2004, which was based on four documents, which suggested that President George W Bush had been naughty as a young officer in the Texas Air National Guard back in the early 1970s.

Almost immediately a number of Blogs took the running claiming that the documents were fake, principally Little Green Footballs and Powerline (the latter was named Time magazine's 2004 Blog of the year. These Blogs were rather strong supporters of the President and his activities in Iraq, so their criticisms could easily be written-off by those with a conspiratorial mind set. But two things were obvious to me from the outset.

First, neither took the point of view that the documents must be fake because the President wouldn't have done such a thing. They attacked the documents on purely technical grounds.

Second, to me their criticisms were utterly convincing, not that I needed to actually read them. The moment I first saw a copy of one of the documents on the Web, I knew they were fake. That's because I know the subject matter. That subject matter is not the details of the Texas Air National Guard practice of three decades ago, but fonts and the faking of documents. And without that knowledge, any 'report' on the subject, such as that Columbia Journalism Review piece, is quite worthless.

The reason I know about the faking of documents is because on two separate occasions in my career, I undertook to fake documents.

Before getting into that, let's get back to that CJR piece. The giveaway that its analysis was useless is this:

Three types of evidence were used to debate the documents’ authenticity after Rather and 60 Minutes II used them in the story. The first, typography, took many detours before winding up at inconclusive. The second, military terminology, is more telling but also not final. The third, the recollections of those involved, is most promising, but so far woefully underreported.

The reporter behind this paragraph is clearly a people person, not a technical person. But tell me what you think. If in a court of law a seemingly credible witness says that the accused never touched the murder weapon, but fingerprints of the accused are found on said weapon, which evidence tends to be more persuasive? Clearly the bland technical stuff.

And so it was with the Rathergate documents. The technical points were as follows:

1. Charles Johnson from Little Green Footballs opened up his copy of Microsoft Word and keyed in one of the documents. He used the default settings for Word -- font, margins and so forth. When he compared this to the electronic copy of the document from the CBS 60 Minutes Web site, he found it differed in only two respects. First, the odd pixel here and there due to image degradation from the multiple copying of the 60 Minutes document. Second, the superscript 'th' in '187^th' was noticably lower in his version than in the CBS version. Actually, LGF didn't make the latter point, and I was in the midst of constructing a lengthy email about it when I discovered that someone else had resolved the issue. It turns out that while the 'th' was lower in a screen shot of the Word display, when printed out it lifted in height to match the CBS documents.
2. the font used in the CBS documents was -- again, within the constraints of multiple copies and noise -- Times New Roman. This is a proportional font, which means that character's such as 'i' aren't as wide as character's such as 'm'. Proportional fonts, in the good old typewriter days, were used in books. Non-proportional fonts were used on typewriters. The vast, overwhelming majority of typewriters. Yes, some very expensive, high end typewriters were available which used one or other proportional font (there are hundreds), but none of these would produce a precise match of the CBS documents. Unlike the current version of Microsoft Word.

Back in 1984 I moved from the Australian Government's Department of Health to the Health Insurance Commission (HIC). At that time I was a medifraud investigator, having moved to the DoH from the Australian Federal Police. At the AFP I had most recently been in the National Criminal Investigation Branch and, before that, the Fraud Squad.

On moving to the HIC I fell in love with the culture of the place, which at the time was still that of a leading edge technology user, and hard-arsed competitor in the private health insurance industry. It had only recently (early in 1984, following the election of the Hawke Labor government) taken over the government health program Medicare. What's relevant to this story, though, is that on my desk was a Raytheon computer terminal, connected to an IBM mainframe computer. The HIC had been good at Medibank Private because it was the first to computerise nationally.

Now for reasons that no doubt speak of some kind of dreadful pathology of personality, I am somewhat of a stationery freak. Ever since childhood I've been entranced by envelopes and paper and pens. But most of all, devices that allow me to create stuff that looks like professional material. In my late teens, after I went off motor vehicles as a result of having rolled my first car, I purchased a portable Royal typewriter. I did virtually nothing productive with it. I just liked the thought of being able to produce professional looking documents. Except, of course, they didn't look professional because the typewriter used the non-proportional Elite font at 12 characters per inch.

Later, when I joined the (then) Australian Capital Territory Police Force, recruit training included Wednesday night attendance at the local Technical college for typing lessons. I was already a pretty fast four fingered typist, but this taught touch typing and we had a considerable incentive to learn. Fail to reach 25 words per minute by the end of 16 weeks, and you were out of a job. Subsequently I bought a full sized manual Olympia typewriter. Must of weighed several hundred kilograms! Kidding, but it was very heavy indeed.

The Raytheon terminal at the HIC office offered me independence from secretaries, at least for mundane things like file notes. Just type them up on-screen, then press 'Print Screen' and they'd chug out on local printer. Then I discovered that the big system printer around the corner, and TSO. That's 'Time Sharing Option' on old IBM computers. Got myself a logon, and learned DCF (Document Composition Facility).

DCF was an incredibly powerful type layout program that had computer programming elements in it. I created a macro, for example, that would allow it to draw diagonal lines across a page between specified co-ordinates. In form it looked rather like HTML, except it used Wordstar-like 'dot' commands rather than < and > signs. And it controlled a suite of proportional fonts beautifully.

I became somewhat of an expert in the use of DCF. Several job changes led me a few years later to become responsible for maintaining copies of the legislation relevant to our investigation responsibilities. First the Health Insurance Act 1973 and later the National Health Act 1953. These pieces of legislation had been amended dozens of times, and were still amended several times per year, so I created 'consolidations' of them with all up to date amendments and distributed them throughout the organisation.

Initially these were printed out in an intensely unattractive manner: as plain text, using the default proportional font. But then I undertook my first document faking task. I decided to make them look as though they had been produced by the Government Printing Office (official consolidations were only issued every decade or so, thus the need for my little effort).

So I created the macros necessary to format the document so it looked just like the real thing, except produced on loose leaf A4 rather than the folded book form of the true legislation. This did not involve faking down to the look of each letter. That would have been difficult anyway, since the system printer was only capable of 240dpi, little better than modern fax.

Some years later, I faked another document. By then I was working in the Pharmaceutical Benefits Scheme which the HIC had taken over. A change in the legislation required a change in the claim form. That I created in -- I'm amazed, now that I think back -- WordPerfect 5.1 for DOS. All the boxes. All the precise page layout. And nothing in the lest intuitive. In that I was concerned in closely matching fonts (a simple sans serif type) and extremely concerned about line breaks and so forth so that everything was in exactly the right spot. I scanned the PBS logo, then hand-traced it into a vector graphics format so that it was scale cleanly. I have a feeling that the PBS logo, as it presently appears on forms, is the same one I created back in the early 90s.

I learned something from these exercises in document re-creation. Doing it precisely, let alone overlay-perfect, was very difficult indeed.

Then I learned something else. We switched from WordPerfect 5.1 for DOS to AmiPro 3.0, to WordPro (a dog of an 'upgrade' for AmiPro) to, finally, Microsoft Word. With every change the look of the documents changed. Letters did not come out in the same place. Lines did not break in the same place.

It was around this time that I started writing reviews of computer software. My July 1993 review of WordPerfect for Windows 5.2, which appeared in Australian PC World hammers relentlessly on problems of font display and printing inaccuracies. So does my later update, and to some extent my review of WordPerfect for Windows 6.0.

The relevance of all this to Rathergate?

Simple: reporting on the matter is no good. The reporter is subject to choosing which of several opinions he or she will accept. But the reporter has no idea which opinions are of value, and which aren't. So the reporter does what all of us do from time to time, the reporter accepts the opinions of those he knows or he finds most congenial, or the opinions most in accord with his or her own views. If one considers Bush to be a liar, one will be more receptive to an opinion that says that the documents in question could possibly have been created on an early 1970s typewriter.

But because I know this subject, I have my own opinion. And that is, that these documents could conceivably have been created in the early 1970s. And the easiest way to have created them would not have been to use an IBM Composer or Selectric II typewriter.

No indeed. The easiest way by far would have been to employ a fine artist to imagine an attractive proportional font, and then painstakingly create the documents by hand. Because no then existing machine could have done it.

© 2005 - Stephen Dawson