5

The complete paper

This question is about Satoshi Nakamoto's original Bitcoin paper, which can be found at:

In an attempt to somewhat date this paper, every source I could find claims that the availability of the paper was first announced by Satoshi on Cryptography Mailing List, on 1 November 2008. At that time, the paper was available via the cited bitcoin.org/bitcoin.pdf address, and the domain bitcoin.org had been registered on 18 August 2008.

Some metadata of the paper

When you open the PDF (I used the archived one) with Adobe Acrobat (I used version 15.010.20060), among the revealed metadata in the Document Properties-panel are the following:

  • Created: 25/03/09 04:33:15
  • Application: Writer
  • PDF Producer: OpenOffice.org 2.4
  • Page Size: 8,50 x 11,00 in
  • Tagged PDF: No
  • Fast Web View: No
  • Fonts Used in this Document (all "Embedded Subset", all "Type: TrueType", all "Encoding: Built-in"):
    • ArialMT
    • CenturySchoolbook-Bold
    • CourierNewPSMT
    • OpenSymbol
    • TimesNewRomanPS-BoldMT
    • TimesNewRomanPS-ItalicMT
    • Times NewRomanPSMT

I don't know whether this is also inherent metadata, or whether this has got to do with my own settings (especially, because it is under a heading entitled "Reading Options"), but the following is also shown in the Document Properties-panel:

  • Language: en-GB

Encoding info

Additional metadata &/or encoding info can be found when opening in a text editor (I used Textmate, version 2.0-beta.8.5), using e.g. the option Encoding: Western — ISO Latin 1 (which I chose).

For example, the document begins with:

  • %PDF-1.4

and (skipped 1 line, which I couldn't decode properly using Western — ISO Latin 1) then:

  • 2 0 obj
  • <</Length 3 0 R/Filter/FlateDecode>>
  • stream

and (skipped about 20 lines of text which I couldn't decode properly) then:

  • endstream
  • endobj

(blank line)

  • 3 0 obj
  • 4448
  • endobj

(blank line)

  • 5 0 obj
  • <</Length 6 0 R/Filter/FlateDecode>>
  • stream

Such streams and endstreams; and objs and endobjs continue throughout the document.

For brevity: for further information regarding encoding, please open the document in a text editor yourself.

Note: Doing so, also reveals that /Lang(en-GB) is indeed present in the original document (and is thus not merely dependent on the settings of the PDF reading application of the reader, as I had suggested above).

Creation of the paper (LaTeX or not?)

Since the identity of Satoshi Nakamoto is a question mark, I have become interested to understand how that paper has been created.

Some people, as did I, intuitively thought that the paper was compiled in LaTeX.

For example:

The study also questioned why the most-cited textual feature of the Bitcoin paper is the fact that it contains double spaces after full stops. The Bitcoin paper was drafted using Latex, an open source document preparation system. Without the base .tex for the Bitcoin paper, which is not available, researchers are unable to tell if the author double spaces between sentences.

The layout (I.e. margins, font, etc) are the default stylesheet for LaTeX. Pretty standard for any academic working in math, physics, or compsci.

Throughout the paper, double quotes are mismatched (I.e. author typed " and " instead of `` and ''). This suggests that the author was not good at using latex.

However, other persons criticize the hypothesis of the creation of the paper in LaTeX.

For example:

"It is not Latex, it is an imitation of Latex.

written in OpenOffice Writer 2.4

8.5 x 11 inches

True Type fonts

converted to PDF after originally written in Writer"

"[...] it was not typeset with TeX; you can tell this by the poor typographic detail (for example, contrast how the letters 'fi' are typeset, and the spacing in the equations). Not to mention the pdf metadata."

More specifically I would like to pose the following 4 questions (of which the last one is perhaps the most normal):

  1. Is it not possible that the paper was typeset/compiled in LaTeX anyway (and the compiled PDF was then imported into an OpenOffice Writer document, just to export it as a PDF again)? And if so, is there any way to find out if that was the case?

  2. Is it not possible that the paper was typeset/compiled in LaTeX anyway (and the metadata of the PDF was; either automatically during compilation, or after compilation; just changed, making it look like anything was done in OpenOffice Writer)? And if so, is there any way to find out if that was the case?

  3. How likely is it that the images in that paper were typeset in LaTeX (7 in total, all presented below). And which LaTeX packages could be used for that purpose?

  4. How likely is it that the formulas in that paper were typeset in LaTeX (4 chosen, all presented below). And which LaTeX packages could be used for that purpose?

It's possible to think about more questions regarding the PDF metadata (such as: is it possible to discover which language-version of OpenOffice was used), but they would be not LaTeX-questions and therefor off-topic, so no need to answer them.


All of the following pictures, I have obtained by first opening each necessary page of the original paper, imported in Inkscape 0.91, while the check box "Replace PDF fonts by closest-named installed fonts" was unchecked. Subsequently, by clicking on images or formulas and then exporting them as PNG's.

Exported (images)

  • Image 1:

Image 1

  • Image 2:

Image 2

  • Image 3:

Image 3

  • Image 4:

Image 4

  • Image 5:

Image 5

  • Image 6:

Image 6

  • Image 7:

Image 7

Exported (formulas)

Note: The Chinese characters were not originally present (cf. the original Bitcoin paper, or Screenshots (formulas) below), but replaced mathematical characters through the Inkscape import-progress.

  • Formula 1:

Formula 1

  • Formula 2:

Formula 2

  • Formula 3:

Formula 3

  • Formula 4:

Formula 4


Screenshots (formulas)

The following are screenshots of the formulas, as facsimile present in the original paper:

  • Formula 1:

Formula 1, screenshot

  • Formula 2:

Formula 2, screenshot

  • Formula 3:

Formula 3, screenshot

  • Formula 4:

Formula 4, screenshot


Note about Satoshi Nakamoto's knowledge of image creation:

It may be important (if Satoshi Nakamoto indeed wrote; and compiled or exported; that original Bitcoin paper) to note that Satoshi claimed not to know "how to do SVG", or in other words: he had no experience in creating SVG images.

Such can be understood by a reply of Satoshi of 27 February 2010, at 4:28AM forum-time, in a Bitcointa.lk thread.

The thread might also be useful to learn about other knowledge of image editing/creation that Satoshi had.

closed as primarily opinion-based by yo', Svend Tveskæg, Romain Picot, egreg, David Carlisle Apr 28 '16 at 6:45

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 3
    Both the formulas and the diagrams look like some standard office thingy, probably version before 2010 (makes sense after all). The sum and the cases bracket are Wordish, and the arrows in the diagrams are OpenOfficeish. – yo' Apr 27 '16 at 23:01
  • 7
    Metadata are just metadata. Like the content of a document, they are under the author's control. At least, they certainly are if, say, pdfLaTeX is used to compile a file. You can tell pdfTeX to use whatever metadata you like, really, and it will use them. Metadata can also be changed after PDF creation very easily. This is not to say that the paper was compiled that way. I doubt it. But the metadata are only useful if the author intended them to be so, just as the contents are only useful if the author intended them to be so. – cfr Apr 28 '16 at 0:28
  • 3
    Of course, you can always use a package which makes LaTeX output look like Word output. If this was made in LaTeX, something like that's been done. It doesn't look like LaTeX output at all. – cfr Apr 28 '16 at 0:36
  • 5
    I agree with @cfr that the document doesn't look like it was produced by LaTeX. A tip-off is the complete absence of discretionary hyphens and, coupled with that, the readily apparent uneven amounts of interword spacing. The presence of Arial also hints at a non-LaTeX origin. – Mico Apr 28 '16 at 3:20
  • You can tell it wasn't created with TeX, by looking at the poor typesetting of mathematics (see particularly the fractions “q/p”, on pages 6 and 7). (Actually, it is no doubt possible to achieve that exact typesetting with TeX, but nobody with enough TeX experience to know how to do that would also have the poor taste to insert inappropriate whitespace in precisely that way.) (Related (deleted) question.) – ShreevatsaR Mar 15 '18 at 22:18