Print a Wikipedia article properly

Why the Print button disappoints

Printing from the browser prints the page. Wikipedia tidies it on the way out and its own Download as PDF does the job, but both give you one layout and no say in it: you cannot move the margins, you cannot change the measure, and the type size is whatever it is.

For something you are going to read on paper — a long article, a set text, something you will mark up — that matters. A line of 85 or 90 characters is genuinely tiring to read, and the fix is not the printer’s settings. It is setting the article for the page it is going on.

Start with the paper

Open the app, search for the article, then open PDF Options. The first thing there is the paper size, and it is already set for where you are: Letter if your clock says United States or Canada, A4 otherwise. On a phone it starts at A5 instead, because a phone fits the page to the screen and A5 gives readable type where A4 does not.

The full list is A4, US Letter, US Legal, and the whole ISO A series from A0 to A10, plus B and C. Portrait or landscape. The sizes are computed from the ISO formula rather than looked up, so A4 comes out at 596×843 pt — a shade off the round numbers, and correct.

Set this before anything else. Every other measurement on the page is relative to it, and it is the one setting a printer cannot fix afterwards.

Margins, and the line length they buy

Margins are in points: 72 pt is one inch. The defaults are 50 pt top and bottom and 64 pt at the sides — about 22.6 mm, which is ordinary book and report practice.

That number is doing real work. On A4 at 14 pt, the narrower 17.6 mm side margins this app shipped with produced lines of 76 to 80 characters, against a comfortable band of 45 to 75. Widening them to 64 pt brings the column to 468 pt and about 65 characters, and the page count barely moves. If you want a wider or narrower measure, this is the control — not the font size.

Justified text that does not gap

Body text is justified by default, and justification without hyphenation is what produces those rivers of white you see in badly set documents — with nowhere to put a long line’s slack, 37 to 47% of lines end up set at 1.6× normal word spacing or looser.

So hyphenation is on, using the same Liang/TeX algorithm and pattern data TeX uses, in 48 languages — most of the Latin, Cyrillic, Greek, Armenian and Georgian editions. Words break at real syllable points, not by counting letters. A language with no pattern set simply gets no hyphens, and nothing else changes.

If you would rather have ragged-right text, Text Alignment → Left turns the whole apparatus off.

Turn the pictures up before you print

This is the setting worth knowing about, because the default is chosen for the screen and not for paper.

Image Compression defaults to Compact, which embeds pictures at the size they are placed — about 96 ppi. That is screen resolution, and it looks visibly soft in print. Balanced, High and Original all fetch images at twice the placed size instead, which is what gives a printed figure its detail.

It is not free: on a long, picture-heavy article, moving off Compact roughly doubles the image data and can add around a quarter to the file size. That is a poor trade for reading on a phone and an obvious one for something going through a printer.

Page numbers, headers, contents

A printed document needs furniture a screen does not, and most of it sets itself once the document is long enough to need it.

• Running header Appears once the document passes about ten pages. You can turn it off, or set it to the article title, the current section, or both.
• Page numbers Placed automatically so they keep clear of the header — or pick any of the six corners, or none at all. The wording can be 1, 1 / 12, Page 1 or Page 1 of 12.
• The first page Left unnumbered, as a title page should be. Number it like the rest if you would rather.
• Contents page Added past about ten pages, where it starts to earn its place. Can be forced on or off.
• Sources & licence Appended by default. It is the CC BY-SA credit that makes a copy you hand on legitimate, so leave it unless you have a reason.

There is also a QR code you can put in the corner of the first page. It links back to the live article, which is exactly the thing a printed copy otherwise loses — anyone holding the paper can scan it and get the current version. It is off by default; turn it on under the first page settings.

Printing less of it

Paper and ink are finite in a way a screen is not, and a Wikipedia article carries a great deal you may not want on it. Before generating you can untick whole sections — the references, “See also”, “External links” — and there is a row of filters for Text, Figures, Images and Tables that drops a whole category at once. Turning images off gives you a text-only copy that prints in a fraction of the pages and none of the ink.

Whatever survives is what gets typeset, and the contents page and page numbers are built from what is actually there. There is more on this — trimming a section to its first few paragraphs, cutting a single table row, and what happens to the citation markers when the references go — in choosing which sections to print.

What it will not do

Worth saying plainly, because these are things somebody about to print would reasonably look for.

• No binding or gutter margin, and no mirrored margins. The four margins are the four margins. For a document you intend to bind, widen both sides rather than one.
• Page numbers do not alternate between recto and verso. You choose one corner and every page uses it.
• No one-click greyscale. You can set the link colours and switch table colouring off entirely, which covers most of what prints as colour in an article, but there is no single mono switch — your printer’s own greyscale setting is the better tool.
• Arabic and Hebrew print correctly but as images, so that text is not selectable or searchable in the PDF. On paper it makes no difference; in a file you intend to search, it does.

There is a fuller account of what does and does not work well, and of exactly what leaves your browser, on the about page. If you want several articles in one printed document, that is a Wikipedia book.