When I started using LaTeX a few decades ago, one of the main arguments for using it was the separation of content and editing. The idea was that the author would concentrate on the content and LaTeX would take care of the editing.

After several decades of use, I think this promise has been at least partially kept: LaTeX does a pretty good job of managing formulas, fonts, distributing text over multiple lines, positioning floats, etc. even without additional packages. Only the microtype package and the sloppy-like commands are used occasionally.

However, I am less sure if this also applies to page layout, i.e. things like line spacing, paragraph spacing, line length, etc. As a rule, there is usually only one serious limitation imposed by the publisher, namely the page size. So, when I finish my books (scientific notebooks), I usually consult graphic designers, none of whom have any experience with LaTeX, and they usually butcher all the default settings completely.

So are design/typographic principles/rules for page layout implemented by standard LaTeX or perhaps by more complex packages like KOMA? I have to say that I was a bit disappointed with KOMA when I read in its manual that "line lengths beyond 80 characters are unacceptable", and the length of this particular line is - 95 characters! So at first glance, I am not sure how many, if any, typographic rules for page layout have been implemented in LaTeX and KOMA. Incidentally, when I look at my favourite textbooks, they all seem to stick to the line length of 13 cm, which interestingly was also the advice of the last graphic designer I consulted.

So my question is: How much can a user rely on the standard LaTeX page layout? To what extent, if at all, is KOMA an improvement on the LaTeX specifications in this respect? Where can one get reliable information or even paid advice on how to design a page layout appropriately, preferably from a person who is both professional in page design and has some LaTeX experience?

PLEASE, do not close this question. It has received a large number of upvotes (with some downvotes) very fast, so I think this is a relevant question that deserves a chance to receive the conclusion.

EDIT: Answers and discussions usually help to understand the question better, and also in this case I would formulate the question a little differently if I were to ask it now:

To increase the efficiency of production, it is always good to distribute the work among different specialists. In the case of publishing, this would mean that the author creates the content, the typographer/designer creates the layout, the printer puts this on paper and finally the publisher distributes the result. My question can be boiled down to the following: LaTeX most likely does the typographer's job, but can it replace a designer? Can ambitious academic notebook projects be published without the involvement of a designer, by tweaking only one or two options? Because as soon as designers are involved, and they almost universally don't know LaTeX, the result deviates far from the default LaTeX settings.

I have also opened a chat room if anyone would like to discuss these issues:


  • 9
    First of all, what is considered good typography is already hard to answer. Although there exist some rules that might be considered golden standards, most of these are only valid for certain languages or writing systems or domains etc. Also, what is considered good typography might well change over time. I would say that the default layout of standard LaTeX classes is already quite good but it is only meant as a very basic layout. What is good typography, I would argue, depends on the specific use case and it hence cannot be achieved by just using the default settings of one single package. Commented Jun 30, 2023 at 9:42
  • 14
    The online manual of KOMA is not optimized, if you want a typographically pleasing version, buy the book. Apart from this: the default papersize in most of the world is A4 or letterpaper and this formats are too large for good results, no real book uses this format. So the default A4/letterpaper layout is always a compromize. Commented Jun 30, 2023 at 9:56
  • 7
    @UlrikeFischer to add to the "no real book uses this format": and the books or magazines which do use biggish page dimensions close to A4 or letterpaper use two columns or extremely wide margins (which then are sometimes used to place figures or additional information).
    – Skillmon
    Commented Jun 30, 2023 at 10:57
  • 2
    You could well print a book with the default settings of the book class and just adjust the paper size. As I said, it really depends upon what kind of book you wish to print. Commented Jun 30, 2023 at 15:27
  • 4
    A good layout is a very subjective issue. The same that for some "experts" on art, some oil painting is wonderful whereas for normal people is a complete crap. Often in this site someone ask how to destroy a good LaTeX default just because is not what normally see using a word processor or just an odd taste ...
    – Fran
    Commented Jun 30, 2023 at 17:46

3 Answers 3


This is a difficult question, because what is considered good type design depends on many different things and also on personal taste. You ask specifically about page layout, but note that the page layout is always related to the text and cannot truly be separated from other metrics such as font size, line spacing etc. Therefore, this is not really an answer, but rather an attempt to push you towards finding a good solution for your project.

If you want to prepare a PDF which is meant to be printed as a book, I think that simply using the default settings of the standard LaTeX classes won't really help you, because they only provide a very basic set up and the default paper sizes of these classes are probably not optimal for a book. So, you would at least need to use a package, such as the geometry package, to adjust the paper size and probably together with it the margins.

You also mention the KOMA classes which, in combination with the typearea package, can help you to set up a traditional paper layout where the outer and lower margins are typically larger than the upper and inner margins. In this aspect, the KOMA classes indeed provide an improvement regarding the settings for the page layout. While adhering to these settings for the page layout is indeed considered a good choice in many cases, it still might not be the right thing if you want to typeset an academic book with lots of diagrams or tables.

You also ask where to find reliable information about page layout and type design and I would again argue that this depends on what you are trying to do. There are surely some rules-of-thumb or even golden standards one should probably stick to in most cases (you could look out for literature written by famous typographers such as Robert Bringhurst or Jan Tschichold), but a lot of design choices still depend on the contents, the medium and the intended readership.

I don't think that there exists a single best solution for everything. But I also don't think that you need to create everything from scratch. What is probably a good idea in order to decide about the "right" page layout is to start with a standard class and, in a first step, just change the paper size as required (if there are other requirements by the publishing house, you should of course also keep them in mind). The result, of course, won't be an optimal page layout, but it can serve as a starting point for further improvements.

In step by step improving the page laout, you should always think about the why: why should the margins of your book should have a certain width? Why should the lines of the text should not exceed a certain length? You will probably find that it more or less always boils down to enhanced readability and usability: Larger lines require larger line spacing, so that the reader can easily find the start of the next line (and this is also the reason why lines should be not too long in general). The page number should probably be on the lower outside, so that the reader who searches for a page can quickly find it using their thumb. The white space around blocks of text or between headers and paragraphs should help the reader to quickly identify what belongs together. The reader should have some place on the page to put there finger in order to hold the book ... If you do your design choices always with the "user" (that is, the reader) in mind, I think you will eventually come up with a good layout.

To answer your questions added in one of the lastest edits: The more ambitous the project gets, the less probable it is to get what you need with only tweaking one or two settings. So I'd tend to answer this with no. But I would also say that there are surely good designers that know LaTeX. So, if you haven't found a good designer who is proficient in LaTeX yet, maybe this means that you just haven't yet found the right designer for you ...

  • I think this is really an overly wordy non-answer, so I am happy about any comments! Commented Jun 30, 2023 at 17:41
  • 1
    I thank you for the answer. I have zero artistic talent and very little knowledge of typography, so I think in my case it's best if I just concentrate on the content and hand over the page layout completely to designers. Since I do not know a single designer (not even an amateur) who uses LaTeX, the designer comes up with something that does not resemble much the LaTeX/KOMA default. So I just use KOMA-related (scrlayer-scrpage, geometry) and some unrelated (marginnote, tcolorbox) packages and implement their demands. Of course, they argue why they propose certain solutions...
    – Pygmalion
    Commented Jun 30, 2023 at 18:03
  • 1
    @Pygmalion I added a paragraph to my answer in order to relate to your latest edit. Commented Jul 1, 2023 at 10:47

Let me share some views and experience. I've been working with all programs mentioned during the past decades.

Put black and white ...

... (La)TeX focuses on the word, while graphic designers focus on visuals and visual concepts.

For (La)TeX it's clear from its history and inner working mechanisms, doing some typesetting to "create beautiful text", as TeX' inventor phrased.

Now I admire what graphic designers can do, like drawing with pens, mouse, drawing pads, meeting proportions, putting contrasts, creating visual relationships between elements, conveying visual messages.

However, you probably would drive them nuts by requesting them to "draw" using Tikz or pgfplot ...

Freedom ...

... is probably what graphic designers favour. They don't have to follow any rules, though they often do by intention.

So for handling of graphics and text I'd be not surprised, if they'd favour programs like Scribus (Open Source), FrameMaker, Microsofts Publisher and the like. Basically you put frames, arrange them (visually), link them for flow-text etc. So basically, you still do kind of graphic design, only with frames. And yes, you have full control (and burden) over any font, baselines, kernels, borders and what have you.

Standardization ...

... is a restriction which gives content wings to fly. Let me explain.

I once worked with electronic engineers, who where used to write their own data sheets first with Ventura Publisher, and later with FrameMaker. Freedom prevailed ... driving the central management of the companies data sheets (quite "some few") nuts. Error checking, corrections, compatibility issues, left alone checking content was a night mare, bound to specialists.

So they decided to combine FrameMaker with a visual XML feature, again separating content from representation. Consequences: you had to build content structures first from a predefined set of elements, and fill in engineers content next. Layout fixed, everything fixed, plus extra rules. Freedom gone.

It were big deals to convert medium sized data sheets (say 50-80 pages) into this new system. Once they were available, the next correction or derived data sheet were much easier to create or adapt.

And from the centrals point of view changing the look&feel was much easier for a huge set of data sheets in one go (much like you'd change the .css for a webpage). And they were able to extract meaning from all those document, through the elements.

Freedom returned in a different way: focus on content, focus on what's relevant to the reader, i.e. those engineers who buy such chips ... and they bought convincing products ...

Latex and me

I heard about Latex in the mid or late 90's, but had neither resources nor incentives to try or even use it. However, I always admired the typographical style I read in scientific or engineering articles. Universities seemed to have some magic tool at hand ...

So only recently I started using Latex, Tikz, pgfplot, and am happy with the standards they provide. I like it.

I tried mimicking the layout of brochures, and the best you can do then is using package flowfram from Nicola Talbot. To some degree it joins the Frame-Making approach and Latex ... it can be tricky and has a learning curve. But then you may be close to a graphics designers visual approach to documents using text as just one design element. // BTW, when mimicking just a few pages of any paper you find in your mailbox you rediscover the graphic designers focus on freedom again. Little to no systematic, from page to page, from rubric to rubric ...

What do publishers do?

In the scientific community that's simple: they provide some Latex template, e.g. for a paper in a journal, and process it into their production system, which probably is or is close to FrameMaker with add-ons.

By tradition authors sent in manuscripts, i.e. handwritten pages, which were converted by a lengthy process into their production system (including printing machines, paper handling etc.). It never was the authors task to even bother about page layout etc.

I'm not aware of publishers who do all with Latex. It probably will have limitations when it comes to the machine level of their production system. Though there are packages to insert all those markings a print-engineer needs, e.g. to adjust the colour drums.

Page, layout, authors and graphic designers

With all this in mind it may be a good idea to work out, or just write down, a useful workflow. Of course, you can do anything, using all kinds of combinations of the above ... it just depends on the effort you are willing to spend.

  • e.g. converting Latex into the Frame-World, e.g. Scribus, is possible, painful and time consuming
  • e.g. letting graphic designers work on your Latex files may not be the best idea, while specifying graphics may work fine

With a step back I'd suggest to view your book projects like an industrial production process and ecosystem:

  • what do you need?
  • what do you supply?
  • how to make it run smoothly at a high rate?
  • etc.

Also I suggest to take readers preferences into account (content communication)

  • 95 characters or more spanned over 13 cm or more
  • may not be a nice idea for a book
  • there is a reason why newspapers invented columns.

Finally, thinking out loud: After this analysis your question about layout may spot more a symptom than the real problem?

  • 1
    Thanks for the poetic reply. When my colleagues who are engineers write a book, they prepare everything in Word, have someone else do the figures, and then leave it to the publisher's designers to put it all together. I do not like this concept, because often the fonts in the text, in the figures and in the equations do not match, and if you outsource the drawing of pictures, you also do not have full control over how you present the ideas. As a physicist, I know LaTeX, and when tikz was invented, I felt so empowered that I decided never to create the material any other way. 1/2
    – Pygmalion
    Commented Jun 30, 2023 at 20:17
  • 1
    On the other hand, I am not good at design, which becomes problematic when a page layout needs to be streamlined. Of course there are graphic designers, but they do not know LaTeX. So they pick a font and design the page layout from the scratch, which I implement with different packages. Is there another way, by relying more on the LaTeX defaults, which are supposedly a good starting point for the page layout? 2/2
    – Pygmalion
    Commented Jun 30, 2023 at 20:17
  • A few suggestions. 1/ Open a chat room, as we tend not to discuss here. 2/ I'm still wondering about the role of graphic designers in your approach: why are they allowed or even encouraged to restart from scratch and change everything? May be good if you'd clarify this in your question. 3/ This question will probably close soon, so: 1/
    – MS-SPO
    Commented Jul 1, 2023 at 8:18
  • 1
    I have opened a chat room, chat.stackexchange.com/rooms/146970/default-latex-page-layout, is there a way to connect it with this question or invite people?
    – Pygmalion
    Commented Jul 1, 2023 at 8:30
  • 1
    The American Mathematical Society (AMS) journal, and most book, production system is based entirely on LaTeX, from receipt of files to creating printing plates on a RIP. The format of AMS journals, and most books, is "traditional"; that is, a current journal issue very closely resembles an issue produced 80 years ago by Monotype. Content is refereed before a file is received for production; after acceptance, if it's not LaTeX, it will be rekeyed. The AMS document classes, and other packages, are posted on CTAN. Commented Jul 3, 2023 at 16:05

A might be surprising fact mentioned several times in The TeXbook and typesetting manuals at the time of moveable types, and modern typography book such as The Elements of Typographic Style (a very good book to read for anyone want to make their own books), is, when it is not likely to typeset a paragraph in a visual appearing way, the first thing to do is ask the author to rewrite the paragraph. Only start to seek suboptimal workarounds if the author was deceased.

If write with deliberate control, even the default document class would work well. In fact, there are many published computer science texts made with the default book class (or variants) from LaTeX, without changing the default Computer Modern font, going fancy with microtype etc, and I would consider it to be a pretty solid option. Example of such books includes Algorithms for Functional Programming by John David Stone, published by Springer, Domains and Lambda-Calculi by Roberto M. Amadio and Pierre-Louis Curien, part of the Cambridge Tracts in Theoretical Computer Science series (some other books from the series also use slightly tweaked default LaTeX class.

If pushing TeX to its limit, even with the fanciest extension and packages it would fail. Page breaking and layout decisions are based more on psychology factors more than can be judge by a simple algorithm, and there is no way leave everything to LaTeX before AI technology been implanted into it.

For KOMA-Script, from a personal perspective I would say it could be a bad idea to rely on an "external package" that is rapidly changing its interface. You would learn it the hard way when trying to modify the LaTeX document years later.

I would recommend you make design decisions based on what you know you can make LaTeX work, rather than trying to implement everything told by your designer while don't understand the potential side effect of the changes. For example, microtype works by having a predefined set of configurations for a limited number of font, so you might have to tweak the parameters to make things right if a different font other than default is chosen, and it would easily be messed up if you don't have the font tweaking knowledge yourself while your designer cannot communicate the change their desire to you.

Not the answer you're looking for? Browse other questions tagged .