313

I've read that—unless I know a lot about typesetting—I shouldn't change the margins of a LaTeX document. The default margin size of the article class is really big and it feels like a lot of space is waste. Thanks to the geometry package, it is simple to change the margins, but I'm not sure if I should.

Can someone tell me why the default margin size is so big? Also, why shouldn't I change the margin size?

jlconlin
  • 10,172
  • 21
    There is space reserved for marginal notes. And I don't see why you shouldn't change the margins (within reasonable limits). – Gonzalo Medina Sep 11 '12 at 18:21
  • 10
    You should have a look at the KOMA-Script Doku. There is a detailed explanation on how margins should be computed. Also, the size is big because you should have a maximum of 80 characters per line, but better in the range of 60-70. – Juri Robl Sep 11 '12 at 18:25
  • 51
    The margins are so big because the paper is so big. – StrongBad Sep 11 '12 at 18:44
  • 5
    See this related question: http://tex.stackexchange.com/questions/14348/best-ratios-of-number-of-letters-per-line-and-of-number-of-lines-per-page . – Jim Hefferon Sep 11 '12 at 19:10
  • 13
    Please, do not think of whitespace as "wasting space"! Whitespace has a function beyond marginal notes and paper<->typeblock aesthetics; it gives your eyes a place to rest, even if you do not consciously notice that. – morbusg Sep 13 '12 at 18:08
  • 1
    To actually render the margins see the layout package. I found this very insightful. – math Sep 17 '12 at 12:01
  • I just picked up a Tufte book at-random (Visual Explanations) and counted an arbitrary line on an arbitrary page: 71 characters. – isomorphismes Sep 05 '13 at 22:19
  • 41
    It is to avoid a similar situation to Fermat's Last Theorem, whereby we'd be bereft of a wonderful academic proof for decades simply because the margin was too small in some smart guy's book. http://en.wikipedia.org/wiki/Fermat's_Last_Theorem – McGafter Dec 18 '13 at 09:18
  • 2
    @isomorphismes Sample size of 1 is not all that telling. – user29020 May 10 '14 at 00:38
  • 3
    I'm not so sure. A sample size of 1 is ample if the results are good or bad enough. Lots of events only need to happen once to shake the universe. You might think that avoiding the risk of another Fermat was worth a lot of margin space... – cfr Oct 25 '14 at 02:02
  • 1
    Without proper outer/lower margins, how would you hold the paper in your hands? (Besides that, lines shouldn't become too long because that would hamper readability.) –  Jul 30 '15 at 14:43
  • 1
    According to my professor, the wide margins are for them to scribble notes for us. So, if anyone wants to give any feedback, they can use that space to write their comments/feedback. – Neil Mar 04 '21 at 13:49

4 Answers4

403

It's not that the margins are too wide. It's that the paper is too big!

As is surely by now well-known, the real question is the size of the text block. That is a really important factor in legibility. As others have noted, the optimum line length is broadly somewhere between 60 characters and 75 characters.

Given reasonable sizes of font which are comfortable for reading at the distance we want to read at (roughly 9 to 12 point), there are only so many line lengths that make sense. If you take a book off your shelf, especially a book that you would actually read for a prolonged period of time, and compare it to a LaTeX document in one of the standard classes, you'll probably notice that the line length is pretty similar.

The real problem is with paper size. As it happens, we have ended up with paper sizes that were never designed or adapted for printing with 10-12 point proportionally spaced type. They were designed for handwriting (which is usually much bigger) or for typewriters. Typewriters produced 10 or 12 characters per inch: so on (say) 8.5 inch wide paper, with 1 inch margins, you had 6.5 inches of type, giving ... around 65 to 78 characters: in other words something pretty close to ideal. But if you type in a standard proportionally spaced font (worse, in Times -- which is rather condensed because it was designed to be used in narrow columns) at 12 point, you will get about 90 to 100 characters in the line.

So what is to be done? One rational answer would be to use smaller paper: that is what (most) books do; but presumably the standard paper sizes have become very deeply ingrained, and I suppose there are things (like illustrations and graphs) for which they are useful.

A second possible answer -- which is what most wordprocessors did -- was to stick to the standard "document design" (margins of an inch or so) and just use proportionally spaced fonts as if they were typewriter text. This produces very long lines, which are not comfortable to read. But that discomfort can be somewhat alleviated by increasing the space between lines (1.5 or double space), which helps prevent "doubling", and by avoiding type sizes below about 11 or 12 points (depending very much on the design of the font).

A third possible answer -- which is what the designers of the standard document classes did -- is to stick with a reasonable line length. No choice then but to "increase the margins": but that's really just a side-effect of the unreasonably large paper they were obliged to work with: the end result is very much what you get if you copy a page from a well designed book onto large paper, without enlarging it.

A fourth possible answer, one I rather like, is to adopt a design like the tufte class, which has an unconventionally assymetric layout which enables you to have a reasonably narrow textblock but make good use of a wide margin for notes and illustrations. Of all the ways of approaching the fundamental design question (how to combine a reasonably narrow text block with uncomfortably large paper) this seems to me to be the most carefully thought out.

A fifth possible answer is to use a two column format. In terms of the efficient use of space that probably makes the best sense of all. If you look at (most) properly designed books which have a page size as large as letterpaper or A4, or at magazines, you will find that a multiple column format is used.

A sixth possible answer would be to use very large type -- but that somehow feels wrong (unless you are going to be reading at a great distance). It's a time-honoured practice for things like lectern bibles, but for "standard" documents it looks odd.

These are all potentially valid design choices. I happen to think that the most conventional one (stick with 1 inch margins, and add line spacing to prevent doubling) is probably the worst of them, and that it only seems "right" because we are accustomed to it. And it doesn't generally save paper, because unless you use single spacing you lose vertically the extra space that you gain horizontally.

Still, none of these choices is hard-wired into LaTeX. You can rearrange things how you like. But don't think in terms of "wide margins": think in terms of a text block which needs to be legible, and the imposition of that text block on the page.

Paul Stanley
  • 18,151
  • 72
    The lines in this answer are about 100 characters long (on my screen), but I felt comfortable reading it. – rgrig Sep 12 '12 at 07:44
  • 2
    Because it has increased space between the lines: font-size: 14px, line-height: 18px. Also the paragraphs are very short and spaced between on and another, so it's easier to find the start of the next line. – Juri Robl Sep 12 '12 at 08:09
  • 7
    +1 I heartily agree with this answer. This is an important issue for browser, too. I asked a question about that here: http://superuser.com/questions/299898/firefox-addon-for-managing-text-width – Francesco Sep 12 '12 at 13:14
  • 52
    It is no argument if you feel comfortable with it. Design is not made for specific people but for a majority. To find out what the majority feels is done by scientific research studies involving many people. These studies typically show that a good length for a line is around 60 characters. – Alfe Sep 12 '12 at 14:59
  • 15
    It's not an argument. It's pointing out an irony. I like the answer very much. ("I felt comfortable reading it" was supposed to refer to the content.) – rgrig Sep 13 '12 at 08:32
  • 10
    @alfe I often wondered if there are such studies. You seem to know more. Could you provide some references? Ideally to freely available research of course, but any pointer would be highly welcome! – cgnieder Sep 13 '12 at 22:13
  • 7
    Actually, I always took for granted what I was told while studying communication sciences and computer sciences. But as it turns out after some quick Google research (cough), the only easily findable studies concerning this come to the conclusion that around 95 cpl the reading speed has a maximum. Of course, this is a simplification; more aspects like colors (black on white vs. white on black), medium (paper or screen etc.) and other things are of importance. What I found were these: http://is.gd/sTaElj and http://is.gd/SmnZvZ – Alfe Sep 15 '12 at 22:44
  • 7
    Apparently the number of 60 or whatever cpl isn't backed up by hard research and probably is not a scientific fact but part of a social culture. Make a thing unusual, then its perception changes; in matters of design this can distract; distraction can lead to the worse, even if the change itself would otherwise (in a culture-free environment whatever that may be in terms of typography) have a benefit. There's a reason the Dvorak keyboard layout never surpassed the Qwerty layout ... – Alfe Sep 15 '12 at 22:58
  • 2
    What is "doubling"? – Mohan Nov 26 '12 at 11:54
  • 8
    @Mohan: Doubling is the phenomenon that, when going from one line of text to the next you, by accident, starts to read the same line again. This is most common when lines are long and set close together. – Johan_E Jan 31 '13 at 18:14
  • 1
    I really don't see anything so wrong with 14pt type. Especially as audiences age it seems kind to be easy on their eyes. Young readers or those with excellent eyes can just read at a distance, but should old readers be forced to pull out a magnifying glass? Given how much I hate "fine print" the opposite seems like a good thing. – isomorphismes Sep 10 '13 at 06:55
  • I do agree and I find it comfortable myself to typeset my notes in 14pt and then scale down and trim to a booklet. – Antoine Chambert-Loir May 07 '15 at 21:00
  • 4
    The right cpl number, 60, 95, 101, for comfortable reading probably depends heavily on the type of material (fiction, news, technical, mathematics) and the experience of the reader. A mathematician will be used to either skimming for content -- headings, formulas, theorems, figures -- or reading carefully every word of a paragraph. In both cases, having more on the same page (vertical compression) is actually a benefit (fewer pages to turn, easier to refer to previous statements). – Assad Ebrahim Oct 25 '15 at 13:40
  • +1 for the tufte class mention. Very excellent, and what most good textbooks are going with now – physincubus Dec 18 '17 at 03:27
  • For option 1 changing the paper size, what paper size should one use? it looks like most books are 6" by 9", but that doesn't seem to be a standard paper size. the closest ones are probably a5 or c5. – Shuodao Oct 07 '19 at 19:52
  • I am appalled by all these people asserting that "the optimal line length is 60 to 75 characters" without even thinking that this may depend on what you read. I don't read a math article like I read a "magazine". For a math article, I don't need to read at lightning speed but I want as much information as possible on a single page. An example: https://www.math.ias.edu/~lurie/papers/Conceptual.pdf – user0 Jan 29 '21 at 13:48
  • (I just noticed I basically wrote the same thing as Assad Ebrahim in his comment, which deserves way more upvotes!) – user0 Jan 29 '21 at 13:52
65

I don't know why Leslie Lamport set the margins to be so large.

The traditional length of a line in books for ease of reading is about 65 characters long, and when set in two columns the line length in a column is about 45 characters. If you go with those numbers then adjust the margins according to the final page size. In commercial printing the pages are typically trimmed down from their initial size (if you print on A4 paper you do not necessarily deliver on A4 paper).

My A Few Notes on Book Design (try texdoc memdesign) goes into some detail about different kinds of page layouts.

Use whatever layout that best suits your purposes.

Torbjørn T.
  • 206,688
Peter Wilson
  • 28,066
23

Here is the simplest explanation (not necessarily the most accurate):

Wide margins are convenient to hold the sheet with his the thumbs!

Paul Gaborit
  • 70,770
  • 10
  • 176
  • 283
19

In addition to Paul Stanley's excellent answer: A seventh possible approach ist to use fonts which feature a large character width for a given point size. In practice, a) one should use anything but Times New Roman b) Palatino is a good choice.

\documentclass{article}% default font size: 10pt

\newcommand*{\alphabet}{abcdefghijklmnopqrstuvwxyz}

\makeatletter
\newcommand*{\showalphabetwidth}[2]{%
  \fontfamily{#1}\selectfont
  \settowidth{\@tempdima}{\alphabet}%
  \alphabet~-- width for #2 at 1\@ptsize pt: \the\@tempdima
}
\makeatother

\begin{document}

\showalphabetwidth{cmr}{Computer Modern}

\showalphabetwidth{ptm}{Times New Roman}

\showalphabetwidth{ppl}{Palatino}

\end{document}

enter image description here

lockstep
  • 250,273
  • 4
    Out of curiosity, what's with all the "hate" for Times New Roman? – Argo Mar 12 '14 at 18:32
  • 13
    @Argo Times New Roman with its small character width is suitable for small (e.g. newspaper) columns -- it was designed for the London Times. Choosing it as the default font for A4/Letter paper documents (as Microsoft did for Word) was a very bad decision. – lockstep Mar 12 '14 at 22:10
  • Ah, makes sense. I think Calibri is the default for Word not. – Argo Mar 13 '14 at 16:17
  • 2
    @Argo It is now. – lockstep Mar 13 '14 at 17:15
  • Yet anytime I reset a document which is in Latin Modern in something else, I end up with more pages. Which doesn't make sense at all... Try 'The quick brown fox jumped over the lazy dog.' and Palatino comes out wider than CMR/LMR. [The interword spacing matters and not just the width of the letters. Also, kerning pairs will affect the overall width. So it can make sense. A straight \alphabet comparison is misleading.] – cfr Oct 25 '14 at 02:15