The LaTeX fetish (Or: Don’t write in LaTeX! It’s just for typesetting)

It’s that time of year when students are signing up for study skills classes. One of the skills that science students are likely to be encouraged to develop is the use of LaTeX. Other people may come to LaTeX for other reasons: people who want to typeset their own books; people who’ve heard that LaTeX may have something to do with Digital Humanities; etc. I’ve written this essay as a sort of pre-introduction to LaTeX. It won’t teach you how to use it (I’m not qualified!), but it will try to give non-users a clear understanding of what LaTeX is really for, which may help them to make their minds up about whether the effort of learning it (not to mention simply getting it to work) is really going to be worthwhile. Why such a long essay? Because many of those who evangelise for the use of LaTeX fetishise it to the extent of spreading misinformation about its true benefits and I want to clear some of that up.

1. What is LaTeX?

According to its own website, LaTeX (pronounced ‘lay-tech’) is ‘a high-quality typesetting system’ and ‘the de facto standard for the communication and publication of scientific documents.’ (LaTeX Project n.d., para. 1) I’m not going to argue with that.

Commercial typesetting of books, magazines, etc is typically done using WYSIWYG applications for desktop publishing, such as InDesign, Scribus, or the now-discontinued PageMaker. LaTeX works differently: you set it to work on a file containing text interspersed with code (i.e. markup), and it spits out a Postscript file that another program can convert into a PDF (some variants will generate PDFs directly). If you’ve spent any time reading (a) papers from computer science conferences, (b) open access preprints of scientific articles on arXiv.org, or (c) documentation for R packages, you will be familiar with the look of those PDFs: the titles (but not the headings) are centred, the first line of each paragraph is indented, the lines of type are justified, the margins are usually generous unless a double-column layout is used, the word-spacing is elegant, and everything is (typically) printed in this weird, old-fashioned-looking typeface called Computer Modern.

Technically, LaTeX is built on top of TeX: ‘a special-purpose programming language that is the centerpiece of a typesetting system that produces publication quality mathematics (and surrounding text)’ (TeX Users Group, n.d., para. 8). TeX was created in the late 1970s by the legendary computer scientist, Donald Knuth, who was disappointed by his publisher’s standards of typesetting. Its early adopters were mathematicians, who appreciated its provision of mathematical symbols unavailable on the typewriter as well as the beauty that its typesetting algorithms gave to mathematical formulae. In the early 1980s, another distinguished computer scientist, Leslie Lamport, extended TeX to produce LaTeX (it is the first two letters of his surname that provide the ‘La’ in ‘LaTeX’). Lamport did this by creating TeX macros: that is, programs that write TeX code for you, behind the scenes. Very few people try to write documents directly in TeX. It’s too hard. Writing in LaTeX is easier. But that doesn’t mean that it’s a good idea.

2. A fetish for LaTeX

Many scientists and mathematicians write articles in LaTeX form: as they type the words and punctuation marks of the text that the reader will encounter, they intersperse them with bits of code that instruct the computer on how to typeset the text. The computer interprets these bits of code as instructions such as ‘style this bit of text as a chapter title’ or ‘insert an ellipsis here’. The result is (ideally) a nice-looking PDF that can be submitted to a journal for review; if the journal accepts the article, then the author or authors can upload more-or-less the same PDF to an Open Access repository and send the LaTeX source code to the journal itself, which will typeset it again. LaTeX users often produce other documents using LaTeX too: their PhD theses, conference handouts, and CVs, for example. Perhaps this is just because, having mastered a particular technology for some particular purpose, one may just as well employ it for every other purpose that it can be made to serve. But there may be more positive reasons too. LaTeX-typeset documents are, as I’ve hinted above, for the most part fairly easy to recognise. A LaTeX-formatted CV is the CV of a LaTeX user, and a LaTeX user is to be taken seriously within LaTeX-using academic disciplines. Etc.

Although LaTeX places fewer obstacles in the writer’s way than TeX does, the fact that people write prose in either of them is anomalous. LaTeX is a typesetting system and a markup language. Typesetting systems are not customarily used for writing in, and while markup languages such as XML and HTML often are, this is generally recognised as a bad idea. It has been quite reasonably asserted that ‘making humans edit XML is sadistic’ (Django Project n.d., para. 10), for example, and while it was at one time suggested that the online journal Digital Humanities Quarterly would require all submissions to be in XML (and a unique variety of XML was created specifically for the purpose), it now additionally accepts submissions in the file formats used by popular word processing packages (see DHQ 2016, para. 10). The requirement to use the wikitext markup language when authoring or editing Wikipedia articles has been recognised by the Wikimedia Foundation as a barrier to participation, though its efforts for reform were stymied by the ever-diminishing community of committed Wikipedia volunteers, amongst whom ‘it’s not a fringe opinion that making editing easier is a waste of time’ (Simonite 2013, para. 22). I write this blog using the slightly simplified version of HTML required by WordPress’s ‘plain text’ editor – though every time an essay gets beyond a certain length, I start to wish that I didn’t. Markup’s great for machines to read and write, but for humans, not so much – and this is well understood by the creators of word processors such as Microsoft Word and LibreOffice Writer, both of which store text in XML form, but neither of which ever makes the user deal with the actual XML.

Despite this, much writing is done in LaTeX. What I call the ‘LaTeX fetish’ is the conviction that there is something about LaTeX that makes it good for writing in. As we shall see, arguments in favour of writing in LaTeX are unpersuasive on a rational level: LaTeX is in fact quite bad for writing in (although it could be worse, i.e. it could be TeX). This doesn’t mean that people shouldn’t use LaTeX at all, but it does mean that people probably ought to stop recommending it as a writing tool.

3. The case for writing in LaTeX

LaTeX is better for writing in than TeX, but that isn’t saying much. Some enthusiasts will say a lot more, however, implying – or stating outright – that writing in LaTeX is somehow better not only than writing in TeX but than writing with the use of a word processor. The blog for the ShareLaTeX online editing software for example gives the following advice to PhD students: ‘Your thesis could be the longest and most complicated document you’ll ever write, which is why it’s such a good idea to use LaTeX instead of a common word processor.’ (ShareLaTeX 2013, para. 1) A fuller argument is presented by the LaTeX Project itself:

LaTeX is not a word processor! Instead, LaTeX encourages authors not to worry too much about the appearance of their documents but to concentrate on getting the right content. For example, consider this document:

Cartesian closed categories and the price of eggs
Jane Doe
September 1994

Hello world!

To produce this in most typesetting or word-processing systems, the author would have to decide what layout to use, so would select (say) 18pt Times Roman for the title, 12pt Times Italic for the name, and so on. This has two results: authors wasting their time with designs; and a lot of badly designed documents!

LaTeX is based on the idea that it is better to leave document design to document designers, and to let authors get on with writing documents. So, in LaTeX you would input this document as:

\documentclass{article}
\title{Cartesian closed categories and the price of eggs}
\author{Jane Doe}
\date{September 1994}
\begin{document}
\maketitle
Hello world!
\end{document}

(LaTeX Project n.d., paras 2-6)

The above is a fairly standard explanation of why people should use LaTeX instead of a word processor, and I’ve seen versions or partial statements of the same argument many times. As of the time of writing, the Wikipedia page on LaTeX, for example, says this: ‘LaTeX follows the design philosophy of separating presentation from content, so that authors can focus on the content of what they are writing without attending simultaneously to its visual appearance.’ (Wikipedia 2016, para. 7) In a presentation entitled ‘Writing papers the right way’, staff at the MIT Research Science Institute follow the question ‘Why LaTeX?’ with the unequivocal answer, ‘Presentation shouldn’t get in the way of content’, and then explain this point with a series of comparisons between word processors and LaTex – comparisons each of which sees LaTeX emerges as the winner:

  • With a word processor, you spend valuable time agonizing over what font size to make the section headings.
    With LaTeX, you just tell it to start a new section.
  • With a word processor, changing the formatting means you have to change each instance individually.
    With LaTeX, you just redefine the relevant commands.
  • With a word processor, you have to carefully match any provided templates.
    With LaTeX, you can be sure you’ve fit the template, and switch templates easily.

(RSI Staff 2015, slide 5)

Sound terrible, those word processors, don’t they? And that LaTeX, it sounds great, doesn’t it? Now, I’ll be honest with you. I took these sorts of arguments at face value for a while, and I more-or-less repeated them as part of a published discussion of academic literacy practices:

While it was difficult to change the appearance of letters typed on old-style mechanical typewriters, popular computer operating systems are distributed with a range of digital fonts, giving millions of people access both to modern typefaces such as Helvetica, Gill Sans and Calibri, and to centuries-old typefaces such as those of the Garamond family. Such opportunities can be viewed as a distraction, which is why many scientists reject word processors in favour of LaTeX: a document markup language that encourages the user to forget about what the text is going to look like and concentrate instead on its conceptual structure.

(Allington and Hewings 2012, p. 53)

So this is – officially – why LaTeX is good for writing in. Word processors make you ‘worry too much about the appearance of [your] documents’, which is ‘a distraction’, but writing in LaTeX enables you to ‘focus on the content of what [you] are writing without attending simultaneously to its visual appearance’, usefully ‘forget[ting] about what the text is going to look like and concentrat[ing] instead on its conceptual structure’. Everybody says so. Even me. (My co-author’s innocent, btw: I wrote that paragraph.)

4. The case critiqued

So convinced was I that this was why scientists wrote in LaTeX that I even had a go at writing a LaTeX paper of my own (it never got finished; there was a lesson there). What I eventually realised was that while the argument is (as noted above) widely repeated, it is also wrong. Let’s have a look at that example again. Seriously, anyone who believes that making people type this…

\documentclass{article}
\title{Cartesian closed categories and the price of eggs}
\author{Jane Doe}
\date{September 1994}
\begin{document}
\maketitle
Hello world!
\end{document}

…instead of this…

Cartesian closed categories and the price of eggs
Jane Doe
September 1994

Hello world!

…amounts to ‘let[ting] authors get on with writing documents’ has (at best) a slightly unconventional understanding of the words ‘let’, or ‘writing’, or possibly ‘get on with’. Do any LaTeX users really believe that this is why they use LaTeX? Perhaps a parallel can be drawn with the ‘discursive mantras’ that Matt Hills argues to be prevalent among fans of cult television shows such as Doctor Who. These mantras – standard arguments trotted out again and again by fans – are, in Hills’s analysis, ‘defensive mechanisms designed to render the fan’s affective relationship meaningful in a rational sense, i.e. to… legitimate the fans’ love of “their” programme’ (Hills 2002, p. 67). Try reading ‘fan’ as ‘committed user’ and ‘[television] programme’ as ‘[computer] program’. But it’s more than that. The argument is used in persuading new generations of academic writers – students especially – to take up LaTeX.

So let’s take a still closer look at the full LaTeX Project argument quoted above:

  1. To produce this in most typesetting or word-processing systems, the author would have to decide what layout to use, so would select (say) 18pt Times Roman for the title, 12pt Times Italic for the name, and so on.
  2. In most typesetting systems – yes, arguably. That’s the point of them. But most people who do their writing on computers don’t do it using typesetting systems. They do it using word-processors such as Word, Writer, or Pages. And these have default settings that are (for the most part) allowed to stand. That is why there are so many documents in 12pt Times New Roman (formerly the default font on Word, the market leader among word processors) and 12pt Calibri (currently the default font on the same word processor). If you want to go messing with the font and the margins on your word processor, you can. But you are not made to. And there’s certainly no need for a conscious decision about it before starting to write. It is thus completely false to suggest that one ‘would have to’ select a typeface, font size, etc in order to produce the example document above in a word processor. One would simply have to type ‘Cartesian closed categories and the price of eggs’, etc – which is a lot more intuitive than typing \title{Cartesian closed categories and the price of eggs} (not to mention \documentclass{article}), as in LaTeX one must. Of course, typing the words ‘Cartesian closed categories and the price of eggs’ won’t produce a nicely formatted title, as \title{Cartesian closed categories and the price of eggs} will (once LaTeX has worked its magic) – but there’s a way around this which is just as effective as the LaTeX way and much less obtrusive (we’ll come to it shortly)

  3. This has two results: authors wasting their time with designs…
  4. Authors have many ways of prevaricating, including messing around with designs. There’s no reason to suppose that authors who use word processors would prevaricate less if they didn’t use word processors. They might prevaricate by perfecting their LaTeX setups. I know a number of academic authors who seem to spend considerable amounts of time doing that. I shan’t say that this is worse, but is it really better? Until there’s empirical evidence that LaTeX authors prevaricate less than other authors, the above is a baseless assertion, expected to be taken on faith

  5. …and a lot of badly designed documents!
  6. Well, maybe. But there are other ways around that problem. For example, the last three universities I have taught at all have formal specifications for the formatting of student work – formal specifications that closely resemble the default settings on popular word processors. When it comes to stopping people from creating documents in purple 28pt Comic Sans, teaching them all to use LaTeX is a lot less efficient than stating that you will refuse to read anything that doesn’t match the style guide. (Teaching them to use word processors properly might also help.)

  7. LaTeX is based on the idea that it is better to leave document design to document designers, and to let authors get on with writing documents.
  8. Think about this for a moment. If ‘How can we let authors get on with writing documents?’ is the question, can LaTeX really be the answer? LaTeX does less to prevent authors from getting on with writing documents than TeX does. But if neither of the two existed, and you had to come up with something, right now, in 2016 – would it really be a markup language?

The MIT Research Science Institute argument isn’t much better. If you look at the three comparisons carefully, what’s actually being contrasted is not LaTeX and word processors, but the effective use of LaTeX and the naive misuse of word processors: all three things that the Research Science Institute staff tell their students can be done with the wonderful LaTeX can in fact be done perfectly well with a modern word processor. I’ll get to the details in the following section, but for now it is enough to observe that people who don’t know how to use a particular tool very well are being told to throw that tool away and learn to use an entirely new one on the grounds that it will enable them to do things that they could have done at least as well with the old one – which is (when you think about it) a little peculiar if the aim is really to help people with their writing, and not (heaven forbid!) simply to evangelise for a community’s preferred way of doing things. The really important thing to teach students is the importance of writing in a structured way and using the features of whichever tool they are using in order to facilitate that, but instead we have LaTeX evangelism and the false implication that word processors don’t facilitate structured writing at all. Someone who indicates section headings in a word processor by emboldening them or enlarging the font size is not using that word processor correctly and will be unable to take advantage of its full range of features, e.g. Microsoft Word’s Outline view, LibreOffice Writer’s Navigator, or the automatically generated tables of contents that both will create at the touch of a button. Comparing good use of LaTeX with poor use of word processors is unfair; the most that can really be said is that you are more likely to be introduced to LaTeX in a class taught by someone who really knows how to use it, and more likely to be introduced to a word processor by playing around with it or under the informal instruction of someone who doesn’t understand it very well, and that, for this reason, the number of people who use LaTeX but don’t use its document-structuring features is probably close to zero while the number of people who use word processors and don’t is enormous.

In sum, the case for writing in LaTeX is more than a little weak. TeX solved a genuine problem for scientists and mathematicians, but it made writing the prose that surrounded their mathematical formulae rather hard; LaTeX partially mitigated that problem at a time when few other reliable computer typesetting systems existed and none were designed specifically for academic use; TeX had behind it perhaps the biggest name in computer science since von Neumann; TeX and LaTeX were already-existing, already-working examples of free software at exactly the time that the free software movement started to kick off and evangelise; things developed from there. None of these was a bad reason for certain academic communities to adopt TeX and LaTeX in the 1980s. But none of them has anything to do with any supposed advantages of LaTeX as a writing medium – and the fact that most people use word processors badly is neither here nor there.

5. The case against writing in LaTeX

To cut to the chase, LaTeX documents are very hard to read until typeset, which is inefficient for both writing and editing. This is a point that programmers ought to understand: if the readability of code is important, then so is the readability of text. LaTeX documents can yield beautifully readable PDFs once typeset, but the experience of editing them might politely be described as sub-optimal.

LaTeX is, as already noted, a markup language. Markup consists of text spattered with code. The code gets in the way of the content. Reading marked up text requires interpreting or filtering out the markup in order to reconstruct the actual text in your head. This is not an advantage when writing or editing prose. I’ll show you what I mean.

Here is a screenshot of some text marked up as LaTeX. The text is taken with slight adaptations from an article I published this June. In the screenshot, I’m editing it using the version of Emacs that comes with OS X, which is my usual text editor when I’m working on a Mac:

Editing LaTeX in Emacs 22 (with deliberate error)
Editing LaTeX in Emacs 22 (with deliberate error)

You’ll note (click for a full size image) that the title is inside a pair of curly brackets identified as containing the title (\title{The IF community}), while the section heading is inside a pair of curly brackets identified as containing the section heading (\section{A past that it could not transcend}). That is the nature of markup. What the markup is doing is good – unambiguously identifying the section heading as a section heading will help us later (e.g. when the designer wants to apply some particular style to all the section headings) – but doing it through markup (as opposed to doing it in some other way) is disruptive of the text for human readers (including editors and the original author): looking at the screenshot, we see the text mixed up with a lot of symbols that are not part of the text, and it is up to us to figure out what words and punctuation marks the eventual reader will see. That is a distraction from figuring out whether those words and punctuation marks are the right ones for the writer’s purpose. No matter how accustomed you are to the markup language in question, it’s an unnecessary cognitive burden. This is a particular issue with regard to the BibLaTeX markup for automatically generating citations and bibliography, because that markup doesn’t at all resemble the text that it will generate in the finished product.

By the way, one of the words in the screenshot is not right. It’s a typing error that I deliberately inserted. I know where it is because I put it there, but looking for it is hurting my eyes.

The following is a screenshot of the same text being edited with LibreOffice Writer. The same typing error is present, but this time you may find it easier to spot (I should add that the spellchecker didn’t; again, click for the full-sized image):

Word-processing in LibreOffice Writer (with deliberate error)
Word-processing in LibreOffice Writer (with deliberate error)

Spoiler: it’s ‘devotes’ (for ‘devotees’). Maybe you spotted it here, maybe you didn’t, maybe you spotted it in the markup – but I think you’ll agree that looking for it in a document sans markup was an easier mental task. I should point out that the citations and bibliography were automatically generated (with Zotero), but in such a way as to appear on screen in their final (easily readable) form: where the LaTeX version of the text has \parencite[see][]{bennett_2002} and \printbibliography, the LibreOffice version has ‘(see Bennett 2002)’ and the actual bibliography.

I’d also like to point out that the title and section heading are no less unambiguously identified as such in this document than they are in the LaTeX document above. At the time when I took the screenshot, the cursor was on the same line as the title – though you can’t see this in the screenshot itself. If you look in the top left of the screen (just underneath the Zotero buttons), you’ll see the word ‘Title’ in a dropdown menu box. By using that dropdown menu, I told LibreOffice that the line on which my cursor was placed at the time contained the title, and in the same way I told it that ‘A past that it could not transcend’ was a top level heading (‘Heading 1’, equivalent to the HTML H1 tag). Just like LaTeX, LibreOffice infers that text following a heading but not designated as anything else is body text belonging to that heading (although one can also explicitly designate text as body text using the dropdown menu just mentioned). Move the cursor from one line to another and the label in the dropdown menu box changes, telling you what the text on that line has been designated as. If you want to change the style of headings or of body text throughout the document, you edit the styles and then the word processor will apply the changes automatically to the actual text. You can also save edited styles as a template, then load them into any document you want, restyling that document to match the template. And of course you can share the template with other users so that everybody’s documents are styled consistently.

So, pace the LaTeX Project’s claims, I didn’t ‘have to decide what layout to use’, nor to choose a typeface, a font size, or a font style. LibreOffice does that for you once it knows what to style as a title, what to style as a heading, etc. And pace MIT Research Science Institute staff claims, I didn’t have to ‘spend valuable time agonizing over what font size to make the section headings’ – I just designated certain bits of text as section headings, and LibreOffice handled the font size. And whatever the Research Science Institute staff may want students to believe, ‘changing the formatting [doesn’t mean] you have to change each instance individually’ – just as you can ‘redefine the relevant commands’ in LaTeX, you can redefine the relevant styles in LibreOffice (though for this screenshot, I’ve stuck with the defaults, as I usually do). Moreover, you don’t ‘have to carefully match any provided templates’, because you can switch templates just as easily in LibreOffice as in LaTeX. And you can do all these things in Word as well. Actually this might be a good time to look back at the above-quoted comparisons between word processors and LaTeX, because we have now seen that they are all false: essentially LaTeX propaganda.

But back to the editing process. Okay, there’s a less eye-watering way to proof-read a LaTeX document than scouring the markup itself. This is to typeset the marked up file, open and read the resulting PDF, look for anything that needs to be changed, edit the marked up file accordingly, typeset it again, rinse and repeat.

If you’re a programmer, that probably makes sense to you by analogy with the code-compile-debug cycle. But that isn’t how most people want to write. Most of us would rather have a piece of paper or a screen that is covered in the words and punctuation marks that are actually going to appear in the finished piece than a screen covered in markup like \parencite[see][]{bennett_2002}. Never mind boilerplate code like \documentclass{article} or \begin{document} at the start of your document – running into something along the lines of \parencite[706]{lena_peterson_2008} in the middle of a paragraph and having to mentally parse it into ‘(Lena and Peterson 2008, p. 706)’ interrupts your train of thought and makes it harder to do what you really need to be doing: reading your punctuated words back to yourself to make sure that they ‘sound’ like what you really wanted to say and don’t have any mistakes in them.

Being able to edit the document that you’re looking at (as opposed to what is in effect the source code to the document you’re looking at) should be taken for granted. In the interests of producing fluent prose and avoiding errors, it’s obviously better to edit a document that has ‘…’ where an ellipsis is intended than a document that has \ldots in the same places.

6. So what is LaTeX good for?

Science researchers are the biggest users of LaTeX. There are also some researchers who use LaTeX (or some variant thereof) despite working in the humanities. These people are not using LaTeX simply because it’s the thing to do: they’re going the opposite way from the herd. As a result, they are perhaps more likely to have reflected deeply on the real advantages of LaTeX – as opposed to imagined advantages, such as those pushed by the LaTeX evangelists above.

As far as I can tell, they choose LaTeX for the opposite reason to the stereotypical one about focusing on content and forgetting about design. For example, one argues that ‘The computer should allow an ordinary writer to produce a polished typeset page, but Word makes this extremely difficult to achieve.’ (Goldstone n.d., para. 7) This reverses the above-quoted arguments for writing in LaTeX: that is, such authors use LaTeX (or variants thereof) because they do not believe ‘that it is better to leave document design to document designers’: in fact, they are using it precisely because they want to have a go at being designers (which is in turn because they ‘worry… about the appearance of their documents’).

This is what LaTeX is good for: not helping people to compose text, but helping them to make it look nice. If that is important to you, go ahead and give it a look.

7. LaTeX: a typesetting tool, not a writing tool

As an author, I want to ‘get on with writing documents’, but sometimes I have reason to play at being a designer, and on those occasions I want to think about design.

LaTeX provides one set of options for those occasions. Desktop publishing packages such as InDesign provide others. Actually, they provide a somewhat wider set. Perhaps unnecessarily wide when it comes to typical academic document types such as conference and lecture handouts. Forget all that nonsense about LaTeX being somehow better for writing in than a word processor: it isn’t. But it is better than a word processor for typesetting. And its use requires fewer aesthetic choices and less design expertise than a desktop publishing package, so it’s likely to save you some time in that regard (and give you fewer opportunities to mess up) provided that the kind of document that you want to typeset is the kind of document that people generally use LaTeX for typesetting. The advantages of LaTeX for academics inhere in its having been set up to produce reasonable-looking documents of the kinds that academics most frequently like to self-publish. Outside of its comfort zone, it’s not a lot of use: TeX can be used for almost anything, but the macros that extend TeX to make LaTeX have for the most part been created with fairly scholarly uses in mind. I didn’t use it to typeset the public report from the Valuing Electronic Music project, for example, because that was a public report and I didn’t want it to look like a conference paper.

Here are the PDFs generated by the above LaTeX file (on the left) and by LibreOffice’s ‘Export as PDF’ facility (on the right). I didn’t alter any default settings for either of them (except by enabling smart quotes in LibreOffice). The LaTeX version includes a date as well as section and page numbers automatically because this is default behaviour for LaTeX articles; you can easily do that on most word processors too but it’s not the default on this one so I chose not to have it done here.

'The IF community' (typeset with LaTeX)
‘The IF community’ (typeset with LaTeX)
'The IF community' (exported from LibreOffice Writer)
‘The IF community’ (exported from LibreOffice Writer)

I think you’ll probably agree that the LaTeX version looks better than the word processor export version. Whether it looks sufficiently better to justify the additional effort is a judgement call that you’ll have to make for yourself.

8. If we don’t write in LaTeX, how can we make use of the typesetting goodness of TeX?

The obvious approach would be to write your document in some other format, then – when it’s finished – go through it, or have somebody else go through it for you, marking it up. That’s how I created the LaTeX version of the sample paragraph above, for example: the article I published was not written in LaTeX, but I thought I’d have a go at LaTeXifying it for the purposes of this essay. But markup languages are often better used by computers than by people, and these days you can save time by getting your computer to do your LaTeX markup for you. So just because your work is going to be typeset with LaTeX doesn’t mean that you will have to author it in LaTeX.

Here is a range of options for getting your text marked up automatically:

  1. Write using AbiWord, Scrivener, Emacs Org-mode, or Texts. AbiWord (available for Windows and Linux), Scrivener (available for Windows and Mac), and Org-mode (available for just about anything if you can put up with Emacs) can export your work as a LaTeX file; Org-mode and Texts (available for Windows and Mac) can additionally typeset it with LaTeX if you have a working LaTeX installation. You can also use AbiWord to open a file that you wrote using another word processor, then export it as LaTeX. I haven’t used AbiWord for a long time and I haven’t yet tested Texts, but I can vouch for Org-mode. Note that the version of Org-mode that comes with the version of Emacs that comes with Mac OS X won’t do this: you’ll need a newer version.
  2. Write using Word or LibreOffice, then use Pandoc to convert the resulting files into LaTeX files [1]
  3. Write in Markdown using any text editor at all and use Pandoc to convert your Markdown files into LaTeX files (yes, Markdown is a markup language, but it’s one that was designed to be simple and human-readable; also note that while Texts is a word processor, it saves everything as Markdown). Also: Pandoc’s citeproc extension provides Markdown with a more intuitive citation markup syntax than BibLaTeX, and because knitr works with Markdown, this option also enables you to embed automatically generated tables etc if you use R. You might want to check out Kieran Healy’s detailed guide to this general approach.

Not all of these options are equal, all rely on your structuring your text properly, and it still helps to have some knowledge of LaTeX (especially if you want to do something like embedding mathematical formulae). But the fact is that they exist because writing in LaTeX isn’t what it’s cracked up to be. LaTeX was invented so that nobody would have to write prose in TeX, which is too hard for ordinary mortals. The above were created so that nobody would have to write prose in LaTeX – which is not too hard for ordinary mortals, but still a fairly bad idea.

9. Typesetting your own book

You may be thinking of using LaTeX to typeset a book that you are not self-publishing. Academic publishers expect this sort of thing from scientists and may also put up with it from humanities researchers. It’s what Knuth invented TeX for in the first place. But here are three things to consider before trying to persuade your publisher to let you typeset your book yourself:

  1. There’s already somebody whose job it is to do that
  2. If you’re so keen on doing other people’s jobs for them, then maybe you’d also like to pick up the physical copies from the printers, transport them to the publisher’s warehouse (hey, they might let you drive the forklift yourself!), do the marketing using your own telephone, and deliver copies to customers and retailers in your own car. Wait – why stop there? Why not insist on operating the printing and binding machinery yourself? In fact, why not buy a mechanical hand press and print and bind the copies at home, like Virginia and Leonard Woolf?
  3. Even if you only typeset your book and don’t bother with the rest of the above, Ricardo’s theory of comparative advantage still says you’re wasting your time.

Not put off yet? Okay, maybe typesetting your own book is going to give you personal satisfaction, or maybe the typography of your book needs to have some particular look that nobody else in the world but you can provide. Fine; it’s not like it’s going to do any harm. (If it were ever to become standard practice for book publishing outside science monographs, then that would be another matter. All book designers would become unemployed, and most books would end up either looking awful or looking pretty much the same as one another. Not good.)

10. A brief warning about the endless inconvenience of anything that has anything to do with LaTeX

As we have seen, there are few good reasons for writing in LaTeX but some good reasons for typesetting in LaTeX. I have not yet touched upon the technical problems involved. If you’re not currently a LaTeX user but are still thinking of using LaTeX for typographic purposes, you really need to know what you’re letting yourself in for. Before wrapping up this essay, I shall make a few observations for the benefit of people in that position.

Free and open source software has a strong tendency towards being difficult to install and get up and running. TeX and LaTeX are no exception. Also, if you want to do anything really wild and crazy – like using a typeface other than Computer Modern – then plain vanilla TeX and LaTeX won’t do. (Please: you do want to use typefaces other than Computer Modern. Computer Modern is not some sort of universal, all-purpose typeface. It’s just the digital version of the typeface that happened to have been used for the first edition of the book whose second edition Knuth created TeX in order to typeset, and it really isn’t suited to some of the uses to which I’ve seen it put, especially in slide presentations.) And anything other than vanilla TeX and LaTeX is really, really difficult to install. So difficult that most people seem to give up on installing individual packages and instead install the whole of something called TeX Live. Tex Live wraps up almost everything that a TeX or LaTeX user could ever possibly want into a single, handy download. A single, handy unbelievably large download that takes up over two gigabytes on disk: to be exact, 2.4 gigabytes for the Mac version, MacTeX. For comparison, LibreOffice takes up about one and a half gigabytes of disk space on a Windows or Linux machine, and less than one gigabyte on a Mac. And LibreOffice is a word processor, a spreadsheet, a slideshow presentation program, a drawing package, and a database, all in one. LaTeX is just a typesetting program. This is nuts – especially if you’re using old or cheap hardware. And even if you decide to weigh your hard drive down with TeX Live, you still have a lot of work to do in getting things to work properly, and almost nothing ever seems to be clearly explained. Sometimes it feels as if getting LaTeX to work has become a sort of hazing ritual through which the pledge must suffer alone.

Here is an entirely typical wail of despair from somebody trying to get LaTeX to work correctly with BibLaTeX and Biber (two programs included in TeX Live that supplement LaTeX as an alternative to BibTeX, also included in TeX Live, or to BibLaTeX plus BibTeX – confused already? Just you wait!):

‘I get “There were undefined references” errors and [I] can’t fix it after two days of trying. I have tried switching editors from Sublime Text 3 to TeXStudio on a Mac, then trying both on a PC. I am willing to try anything at this point. …

I have read about doing a compilation trick but I’m not sure how to do this in either SublimeText or TexStudio. … I have run into many problems and taken many detours that led to other problems. I’m at a loss. Can someone please give me a few hints or keywords I can search for to fix these problems, or a complete solution? I can’t even get a minimum working example up. I will install anything.

(user2205916 2014, paras 1 & 9)

Every time somebody new tries to get started with LaTeX, that person is set up for hours or even days of this kind of thing – plus a lifetime of fiddling with TeX and LaTeX’s quirks – even if (like this user) that person employs relatively ‘user friendly’ graphical applications such as TeXstudio rather than trying to cope with Emacs or vi and the command line. This particular user has been driven to such distraction that he or she has not only changed LaTeX editor programs but changed computers in the vain hope of getting anything to work. Now the response. The accepted answer carefully outlines a solution and then explains the underlying problem as follows:

it is possible to set up TeXstudio in alternative ways to achieve the same effect. The key is that you have to ensure that the[re] is a sequence[:]

  • LaTeX
  • Biber
  • LaTeX

which can be done ‘by hand’ (as I have) or can be automated in various ways. Note that the same general idea applies whatever editor is used: this is a feature of LaTeX and not of the editor.

(Wright 2014, paras 3-5)

Well, how silly of the would-be LaTeX user not to have realised that the only way to get LaTeX running properly with Biber was to run it twice, once before and once after! Now note the grateful response to this (unusually clear) answer: ‘The detail of your answer will be very helpful to other neophytes like myself. Other answers assume a level of computer/LaTeX literacy not all have.’

It’s a typical story. Set foot on the path of LaTeX, and sooner or later you’ll be tearing your hair out. You have been warned.

Footnotes

1. Converting an .odt file with a Zotero-generated bibliography to LaTeX with Pandoc is slightly more difficult than it sounds. In my experience, the best way to achieve it is to open your .odt file in LibreOffice and save it as a .docx file (which turns all the Zotero references and the bibliography into ordinary text), then use Pandoc initially to convert the .docx file to a Markdown file and then to convert the resultant Markdown file to a LaTeX file.

References

Allington, D and Hewings, A (2012). ‘Writing in English‘. In: Allington, D. and Mayor, B. (eds.) Communicating in English: talk, text, technology. London: Routledge. pp. 47-76.
DHQ (2016) ‘DHQ submission guidelines’. Accessed on 11 Sep 2016 from http://digitalhumanities.org:8081/dhq/submissions/index.html
Django Project (n.d.) ‘The Django template language | Django documentation | Django’. Accessed on 11 Sep 2016 from https://docs.djangoproject.com/ja/1.9/ref/templates/language/
Goldstone, A. (n.d.) ‘Producing digital documents’. Accessed on 3 Sep 2016 from https://andrewgoldstone.com/tex/
Hills, M (2002). Fan cultures. London / New York: Routledge.
LaTeX Project (n.d.) ‘An introduction to LaTeX’. Accessed on 3 Sep 2016 from https://www.latex-project.org/about/
RSI Staff (2015) ‘Introduction to LaTeX: writing papers the right way’. Slide presentation, Massachusetts Institute of Technology Research Science Institute. Accessed 12 Sep 2016 from http://web.mit.edu/rsi/www/pdfs/new-latex.pdf
ShareLaTeX (2013) ‘How to write a thesis in LaTeX pt 1 – basic structure’. Accessed on 12 Sep 2016 from https://www.sharelatex.com/blog/2013/08/02/thesis-series-pt1.html
Simonite, Tom (2013) ‘The decline of Wikipedia’. MIT Technology Review, 22 Oct. Accessed on 11 Sep 2016 from https://www.technologyreview.com/s/520446/the-decline-of-wikipedia/
TeX Users Group (n.d.) ‘History of TeX – TeX Users Group’. Accessed on 9 Sep 2016 from https://www.tug.org/whatis.html
user2205916 (2014) ‘Biblatex, Biber, and LaTeX: citations undefined’. Stack Exchange, 12 Jan. Accessed on 10 Sep 2016 from http://tex.stackexchange.com/questions/153647/biblatex-biber-and-latex-citations-undefined
Wikipedia (2016) ‘LaTeX’. Accessed on 10 Sep 2016 from https://en.wikipedia.org/w/index.php?title=LaTeX&oldid=736123000
Wright, Joseph (2014). Answer to ‘Biblatex, Biber, and LaTeX: citations undefined’. Stack Exchange, 12 Jan. Accessed on 10 Sep 2016 from http://tex.stackexchange.com/questions/153647/biblatex-biber-and-latex-citations-undefined

84 thoughts on “The LaTeX fetish (Or: Don’t write in LaTeX! It’s just for typesetting)”

  1. Dear Daniel
    I enjoyed reading your article even though I must admit that I do have RALF : Recently Acquired LaTeX Fetish. At least we both agree that the end result does look better.

    On the issue of LaTeX eliminating a lot of time wasted formating the pages I would say that there is a little bit of truth on both sides of the argument. I know that I have spent time setting up my preamble to obtain the results that I want. Sometimes I have to go to a website to look up the code that I need for a function (stuff that can be just a click of an icon in any word processor). For me it is worth it, I enjoy doing it. For a student who may have a deadline to meet for handing in assignments this can be a problem. As much as I praise LaTeX, I do caution young people that three days or less before an assignment is due is not the time to start learning LaTeX. This is something to do before the school year starts. The good news is that the learning curve has been flattened out thanks to the great LaTeX how-to videos on YouTube.

    I find that the new LaTeX text editors both on-line and the software we download can be just as easy to use as any word processor. Yes LaTeX does consume a lot more room on a hard disk but it is not an issue. I have installed LaTeX Live on a Raspberry Pi with an eight gig SD card and still had lots of room to work with files. Today sixteen to thirty two gig SD cards are very inexpensive. For those of you reading this who are looking for a good note taking software that will export to LaTeX or HTML or Markdown check out the Zim Desktop Wiki.

    Computer technology has displaced many jobs, I use to work for Kodak. I left Kodak to take a two year course to become a library technician. Within ten years computer technology did the same thing to library jobs that digital photography did to Kodak. I am still employed in library work part-time but my wife who is a highly skilled pre-press graphic artist, has had to leave the industry.

    It is not all bad. At one time, book publishers had all of the power. But not anymore as anyone can produce a good looking book using LaTeX. As a law librarian, I see this as a good thing. (The environmental benefits of digital photography are huge, it had to happen)

    My first Linux distro was Red Hat 5.0. I looked at LaTeX then and I liked it but was not to happy about the learning curve and my relative low need to use it. Three years ago I took on the task of creating a poster to display at a law library conference and it rekindled my appreciation for LaTeX and now I wish that I had kept on with it. Today I use LaTeX to publish a newsletter and other documents for the library. Yes, it is worth the effort to learn it, and use it, but we must remember that we do not have to learn all of it, just what we need to know for the task at hand.

    Best wishes
    John Kerr
    Guelph, Ontario

    1. Hi John

      Thanks for your comment! What you’re using LaTeX for is exactly what it’s best for: publishing documents. What I regard as a problem is that it’s evangelised as a tool for writing them. I like LaTeX a lot; I just think it’s important to be clear to students both about its real advantages and disadvantages and about the alternatives that exist. It produces beautiful PDFs, but that isn’t important for everybody, and what are usually sold as the advantages of writing in LaTeX are in reality the advantages of structured writing more generally.

      In my opinion, it’s the latter that we really need to be teaching, e.g. introducing the importance of having a hierarchy of headings, and explaining that this is how we do it in Word and this is how we do it in LaTeX, etc. Also, anyone considering working with LaTeX should know that using LaTeX doesn’t have to mean writing in LaTeX: one can e.g. write in Markdown and then convert to LaTeX with Pandoc (and this gives the option of simultaneously converting to HTML and publishing to the web). Or one can use Scrivener, Texts, Abiword, or Org-mode – all of which will export to LaTeX at the touch of a button.

      On the size of TeX Live: that didn’t bother me until I decided to install a relatively minimal Linux setup in order to keep as much disk space as possible free for data (some of my datasets are quite large). I carefully reviewed the size of each package before installing it, and realised that TeX Live would be the largest of all by a considerable margin. That struck me as wrong: I only needed XeTeX and XeLaTeX, which are quite small. So I tried to install just those. While I did eventually manage it, it took a long time and was very frustrating – and what was most frustrating was that when I looked for help online, the advice I found basically amounted to ‘install TeX Live; everybody else does’.

      On computer technology displacing jobs: there’s a lot more of that coming, but you’re right, it’s not all bad. The point I was making above is that if one is publishing a book with a conventional publisher (and not self-publishing it), then by typesetting it oneself in one’s spare time, one is directly taking work away from someone who would have been paid to do it. And that doesn’t really matter if only a handful of us do it, but it would matter if everybody did.

      I’m sorry about what happened to your wife’s job – and about the skills that are disappearing in that industry (skills that were and still are in many cases an important part of their possessors’ identities).

      Best wishes

      Daniel

  2. Hi Daniel,
    I don’t get it. Why are people who do not want to use Latex so aggressive about it? Using words like “fetish” etc. I do not see where this is any less being an “evangelist”?

    Some of the things you claim in your article are not true. I do not know if that is because you did not know about it or if you choose not to mention it?
    For example, the settings you use in your emacs for writing code… damn they hurt my eyes!
    You do not have to type in lime green with black background… you can choose whichever style and syntax-highlighting you want! Additionally, there are tons of different editors you can use. Simple ones like gedit or notepad, all-purpose ones like Atom or Notepad++, or specialised editors like Lyx.

    You can write external style files for Latex which you can share and use in different files! (And work together with others using Git, for example.)

    It is a very individual thing if you find something intuitive or not.
    For me it feels much more intuitive to write \title{This is a headline} then to push a button, a hotkey or let a program “magically” turn something into a headline because it is sure I really wanted to have one now. This has nothing to do with being “a fan” or having a “fetish”. It is another way of thinking. My brain works that way, it is not my fault!

    I am a student and I came to Latex because Microsoft Word and Excel made me cringe. The Libre Office alternatives were even worse.
    We had to learn how to use it for writing our reports in chemistry, biophysics etc.. Used it for almost 2 years on a regular basis and I can say for myself that I can produce anything I need with the Office products. But at what price? Learning it was a pain, working with it feels like a fight. I fight with a program to make it understand what I want. Crashes, freezes, broken files and let me not start on the subject “should science be published with proprietary software?”.

    So I looked for an alternative and I found a book about Latex in our universities library. Yes, it has a steep learning curve. Yes, I do (still – after almost two years again) need longer for some things than I would need with Office.
    But I will happily keep on spending half an hour longer with every task (although I got the impression that Latex is faster when working on big files). Because I do not feel exhausted or stressed after working with Latex. It does not make me angry and I always know exactly why and what is happening.

    And now I use it for diagrams and tables and plots and poetry. I use it for stuff no one ever sees.
    It showed me the world of open source. I recently started learning R and Markdown additionally. And yes, I would recommend EVERYONE to give it a try! You should not try to scare people off from using it. Let them experience it on their own – some will like it and some won’t. There is nothing rational about it, just personal taste.

    You make it look like almost everyone who chose to use Latex could do better if they would just learn how to properly use a word processor. I say people are different and for some Latex fits their thought process better. This is not a competition…

    (I apologize for my bad English.)

    1. Hi Nina

      Thanks very much for your comment (and don’t worry about your English, which is great). I am a little confused, though: you state that ‘[s]ome of the things [I] claim in [my] article are not true’, but the example you provide is ‘the settings [I] use in Emacs for writing code’. Those settings are my personal preference, so they’re not true or false. Perhaps you meant that I was suggesting that LaTeX is a poor writing and editing medium because it looks weird (and exaggerating the weirdness with my Emacs setup)? That’s not the point I was trying to make.

      The problem with using a markup language such as LaTeX or XML as a writing and editing medium is that any text that has been marked up will be less readable than the same text without markup, because it has been marked up. LaTeX is a poor writing and editing medium because it is a poor reading medium, and it is a poor reading medium because it consists not of plain text but of plain text littered with bits of code such as \ldots and \parencite. The less code there is, the closer the text will be to genuine plain text, and the more readable it will be. That’s why we have minimal markup languages like Markdown (which – as I point out in my article – can be converted into LaTeX). But even plain text is less readable than typeset text, which is why LaTeX – and all other typesetting systems, including WYSIWYG typesetting systems – came into existence in the first place. This is not just a matter of opinion, by the way. If LaTeX was good for reading, we wouldn’t typeset it at all, we’d just publish the raw LaTeX files.

      While a word processor such as Word or LibreOffice won’t typeset text as well as a typesetting package will, that isn’t its purpose. Its purpose is to give text some of the readability of a typeset text even while you’re writing it, for example by indicating emphasised text by displaying it in italic type (i.e. like this) rather than by wrapping it in markup (i.e. like \emph{this}). This is better for writing and editing, because it’s better for reading.

      Now to your question about why I put my case so strongly (‘fetish’, etc). This is because it’s not just a matter of individual preference: there is social pressure involved. Every year, students are encouraged to use LaTeX with arguments backed up by statements that are literally untrue. I go over the most common of these arguments in section 3 above, and show that they are untrue in section 4. The result of all that duplicitous persuasion is a tremendous amount of wasted time. You say that you personally feel that it is more intuitive to type \title{} than to press a button. I can’t argue with that, because your feelings are yours alone. But you also say that even after learning LaTeX for two years, it can take you half an hour longer with every task. That’s not a great recommendation.

      Now, if people want to do things in a time-consuming way because they find it more enjoyable, then that’s their own business: we all have our quirks. But if they’re going to try to persuade other people, e.g. students, to do it on false grounds, e.g. that ‘[w]ith a word processor, you spend valuable time agonizing over what font size to make the section headings’, then I think somebody ought to take a stand. There really are better things that students could be doing than learning to use a typesetting package. Unless they’re graphic design students, that is… but graphic design students are unlikely to have much use for LaTeX because it’s nothing like as powerful a design tool as the industry standard, InDesign.

      Lastly, I think it’s great that you’re learning R and Markdown (in which case, I guess you know about knitr and Pandoc, which together remove any necessity to write in LaTeX even if you’re going to be using it for typesetting). But there are many other possible introductions to programming and open source than LaTeX – and some of them are much more directly useful (e.g. R itself).

      So of course you have a right to use LaTeX for writing in, especially if you enjoy it. As I said, we all have our quirks! For example, I write code in Emacs, which you (and most people) clearly don’t like. We’re all different, and many things are subjective. But I want people to think twice before accepting – or passing on – the idea that there are objective advantages to writing in LaTeX, because – on the evidence – there aren’t.

      Best wishes

      Daniel

    2. Written by someone using green text on a green background.

      1. You may not like pale grey on green, but I do.

        Incidentally, this is a nice illustration of the distinction between semantics and presentation — a distinction that LaTeX unfortunately collapses. The semantics of my text are declared in HTML, and have no particular colour. The colour is declared in a separate CSS file, which you can effectively switch on and off, and which I can easily change for another. And in fact, I wrote the text not in HTML but in the ‘HTML lite’ that WordPress generates HTML from. Moreover, while I didn’t do so in this case, I often write in Markdown, which is a purely semantic declarative markup language that I can compile to HTML or to WordPress’s ‘HTML lite’… or indeed to LaTeX and thence to PDF.

        The trouble with starting in LaTeX (i.e. composing a text using LaTeX markup from the beginning and not, for example, generating LaTeX from some other source) is that LaTeX is a mishmash of semantics and presentation on top of an imperative presentational markup language called TeX — and that while it’s less verbose than TeX, it’s much more verbose than several other alternatives that now exist.

        I believe that people are resistant to acknowledging this because of the immense personal, social, and institutional investment in LaTeX within certain communities. That is what I call the ‘LaTeX fetish’.

  3. As you point out, there are two audiences for LaTeX. The more enthusiastic audience seems to be comprised of physical scientists who are drawn to its ability to assist their composition and editing of formulas and equations. Documents also include ordinary language, and herein lies the rub. Too many documents have what I call “the look of LaTeX” – an ugly typeface, awkward spacing, and heading fonts that are out of proportion with the text. Moreover, they feature what, I am sorry to say, is plainly poor writing. Because LaTeX is not a good editing platform, I wonder whether it hides or actually encourages bad grammar, poor sentence construction, and errors of expression and reasoning.

    I suspect that most users of, call it, “editing” software are either using what is at hand or what they are expected to use. At my workplace, copies of MS Office are on every agency-provided computer; employees can purchase a copy of Office Professional Plus for their personal use for all of $10.00. That’s a great deal for us, but an even better deal for a behemoth that wants to keep the office software market in its corner. How can Corel, the current owner of WordPerfect, compete? If there is a competition, it is between products that take entirely different approaches to the task. One is structured into objects and containers, like a relational database. The other is a type stream that you use somewhat like you regulate the water that comes out of a faucet. I doubt that many users of these products understood the difference when they made the commitment or when the product was selected for them.

    The appeal of LaTeX is that it seems on the surface to be an ideal combination of both: It is a type-stream product that applies hierarchical structure in the background. The problem, as you point out, is that by default it does a poor job of both. I endorse your idea of composing and editing with a word processing program, then importing the text into a dedicated typsetting or desktop publishing program that you have learned to use skillfully. The difficulty here is that the LaTeX shops that are set-up mainly around academic departments and disciplines seem to value conformity above all; the camaraderie they exhibit is based on a mythology that all they need to do is put their fingers on the keyboard and the rest will take care of itself.

    1. Excellent points, especially this:

      The difficulty here is that the LaTeX shops that are set-up mainly around academic departments and disciplines seem to value conformity above all

      What is behind that phenomenon, I suspect, is a fear that the technical expertise embodied by those ‘LaTeX shops’ would be devalued if it were admitted that there is no particular advantage to writing in LaTeX. From this, the need for conformity arises: to reproduce the demand for LaTeX expertise, there must be new LaTeX learners. And so learning to use LaTeX comes to be treated as an integral part of becoming a scientist.

      Hence all the false claims about the inferiority of word processors, I think. The idea of writing in LaTeX for purely social reasons doesn’t seem very attractive.

  4. Try setting formulae in Word or most any other Wordprocessor. Latex can be used for literary works but afaik that was never its prime design intention. The article is interesting but it comes over simply as a wordy case of someone who has found it difficult to get to grips with Latex.and chosen to write about why they cannot use it!
    The *really difficult* thing about TeX/Latex is setting diagrams. The GUI-type packages help, tikz, for example, but even they need a considerable investment in time and require a certain standard of mathematics.
    It takes effort to understand and use Latex to a decent standard and it is not the easiest of tasks. But what it does is amazing and well worth the investment – if you need what it has to offer.

    1. The article is interesting but it comes over simply as a wordy case of someone who has found it difficult to get to grips with Latex.and chosen to write about why they cannot use it!

      And your comment comes over simply as a hasty case of someone who couldn’t be bothered to read my article but has chosen to reply anyway. (Sorry to be snarky, but you weren’t exactly polite.) I can and do use LaTeX. The article is about why people should stop advertising LaTeX to students with false claims.

      Try setting formulae in Word or most any other Wordprocessor. Latex can be used for literary works but afaik that was never its prime design intention.

      Stop right there – that’s exactly what I said! The second half of the title is ‘Don’t write in LaTeX! It’s just for typesetting.’

      Of course LaTeX is good for setting formulae. It’s just not very good for writing or editing text – which is an issue not just for ‘literary works’ but also for conference papers, journal articles, student assignments, etc. Paul put it well in his comment above: ‘The more enthusiastic audience [for LaTeX] seems to be comprised of physical scientists who are drawn to its ability to assist their composition and editing of formulas and equations. Documents also include ordinary language, and herein lies the rub.’

      If people want to write prose in LaTeX, then of course that’s their own business. But using false arguments to persuade students that LaTeX is better for writing prose in than a word processor is wrong – both for the reason that using false arguments to persuade people of things is always wrong and for the additional reason that most students find writing really hard and need to spend time learning to write better, rather than wasting time learning to typeset their writing (whether with LaTeX or with anything else). Do people really use false arguments to persuade students that LaTeX is better for writing in than a word processor? Indeed they do; I quote examples of such false arguments in section 3, and show that they are false in section 4. I’ll re-quote one of them in a moment.

      The *really difficult* thing about TeX/Latex is setting diagrams.

      Yes. It’s also pointless unless you self-publish, because most publishers don’t want anything to do with the results. And I don’t just mean publishers of humanities and social science journals, but publishers of leading journals in the hard sciences too. This, for example, is the policy at PLOS ONE:

      PLOS does not accept vector EPS figures generated in LaTeX. Submit TIFF or EPS figures created in standard software.

      http://journals.plos.org/plosone/s/figures#loc-latex

      And while it’s possible to convert your LaTeX diagrams into a format that PLOS will accept (and the above-quoted page gives instructions for doing so), it’s much easier to cut out the (as you put it) ‘*really difficult*’ LaTeX stage and create them in ‘standard software’ to begin with. But many people seem to love doing things the really hard way for the sake of it (i.e. even when there’s no advantage beyond personal satisfaction and the kudos that comes from having demonstrated the ability to do something really hard), and I suspect that this has played a role in the spread of LaTeX beyond the sphere for which TeX was originally intended.

      It takes effort to understand and use Latex to a decent standard and it is not the easiest of tasks. But what it does is amazing and well worth the investment – if you need what it has to offer.

      Sure. I just want its evangelists to be honest about what it really does have to offer. You, for example, were honest about that when you said ‘Try setting formulae in Word or most any other Wordprocessor.’ An example of dishonesty is telling students that ‘[w]ith a word processor, changing the formatting means you have to change each instance individually’, because that is simply not true.

      And that’s a real example.

  5. As an opinion piece it’s a worthwhile point of view and a good read—thank you. As a decisive argument as to why one should not write content using LaTeX I don’t find it particularly persuasive.

    How we write is very personal, whether on a keyboard or with a pen. LaTeX, in the end, is simply code for the purposes of creating a final, human readable, output. ‘Readability’ is an issue for the output; as afar as ‘readability of code’ is concerned LaTeX is perfectly functional, especially on a well set up editor.

    WISIWYG ‘word processors’ are, in their own ways, just as clunky, and in certain cases very inappropriate and clumsy tools for producing good quality output.

    As someone has already stated in their own terms, this ends up being a case of ‘horses for courses’. LaTeX/WISIWYG are not panaceas for writing good content or good output, they are tools for a very complex task. Depending on personality and need one may suit someone very well most of the time, some of the time, or none of the time, but at least there is a choice. We should be free to explore those choices, learn from others, and come to our own conclusions.

    Personally, I find LaTeX very satisfying, both to use and to learn. Do I use it for everything? No, but when it comes to providing consistent high quality output, especially in longer documents I hugely prefer it to the WISIWYG alternatives, but that is just me.

    One other thing: seeing the .tex extension on a file fills me with joy. It will be almost certain that I will not only be able to open the file, but that I will be able to compile it to produce the output the author intended. Other systems are still fighting like cats in a sack over file formats and standards, and even applications in the same family cannot be guaranteed to make sense of files produced by their ancestors.

    1. As an opinion piece it’s a worthwhile point of view and a good read—thank you. As a decisive argument as to why one should not write content using LaTeX I don’t find it particularly persuasive.

      Thank you very much. I love polite people.

      How we write is very personal, whether on a keyboard or with a pen.

      Sure it’s personal. What disappoints me is that LaTeX enthusiasts are willing to feed students misinformation in order to persuade them to learn LaTeX (see examples in my article). That’s not about personal choice, it’s about deception. Maybe the deception is not conscious. Maybe the people running these introduction to LaTeX workshops believe what they’re saying, having heard it years before from the people who introduced them to LaTeX, and so on and so on, and maybe none of these people ever bothered to check whether what they were saying was actually true. Kinda disappointing behaviour given that a lot of them are scientists, I must say.

      LaTeX, in the end, is simply code for the purposes of creating a final, human readable, output.

      Yup. As I have said, it’s a typesetting package.

      ‘Readability’ is an issue for the output

      But also for the input, because you have to read what you’re writing (or editing) while you’re writing (or editing) it. Text shot through with code is less readable than text. If it wasn’t, people wouldn’t have invested so much energy in creating ways to generate LaTeX from less intrusively marked-up formats such as Markdown and Emacs Org-mode. Heck — if it wasn’t, we’d publish the actual .tex files, rather than the .pdf or .ps files that they compile into.

      as far as ‘readability of code’ is concerned LaTeX is perfectly functional, especially on a well set up editor.

      Yes — as far as readability of code is concerned. The trouble is the readability of the text when it’s all shot through with code. ‘\parencite[see][]{bennett_2002}’ is less readable than ‘(see Bennett 2002)’, even if you understand the code.

      WISIWYG ‘word processors’ are, in their own ways, just as clunky, and in certain cases very inappropriate and clumsy tools for producing good quality output.

      Oh yeah — word processors certainly are ‘in certain cases very inappropriate and clumsy tools for producing good quality output.’ But the thing is, they’re not intended for producing good quality output — they’re intended for producing copy that can then be typeset in a separate package if necessary (if what’s being written is a letter, for example, then it’s probably not necessary).

      For what it’s worth, I think that word processors could be much better at this task. It would help if they didn’t make it so easy to apply arbitrary formating to arbitrary bits of text, for example, as putting a brake on all that would encourage people to use styles properly.

      Depending on personality and need one may suit someone very well most of the time, some of the time, or none of the time, but at least there is a choice. We should be free to explore those choices, learn from others, and come to our own conclusions.

      Yes, totally. The trouble is that a lot of people — not you, but people I quote in my article — evangelise LaTeX by disseminating misinformation about word processors. That can’t possibly help people to come to their own conclusions.

  6. In the same spirit as your article, I didn’t bother finishing it before offering my opinion — why invest all that effort just to know what I’m talking about? But I did read the first couple of pages, in which you list a bunch of objectionable (to you) attributes of any LaTeX-generated document, ALL of which are incorrect.

    I’m fine with using WYSIWYG word processors to write text-only papers and letters and notes to their pals, but I have suffered mightily trying to generate readable documents containing mathematics and/or captioned figures using publisher-mandated word processors. The amount of effort required to use those beasts for anything technical far exceeds the net effort I have invested in LaTeX and other adaptations of TeX, and the results are still unacceptably bad.

    I don’t expect anyone who doesn’t know what a command line is to fully grok LaTeX, but if you want to go on a jihad to wipe out LaTeX by convincing neophytes not to use it, perhaps you should eschew the use of “alternative facts” and learn how it actually works & what it can actually do.

    1. In the same spirit as your article, I didn’t bother finishing it before offering my opinion — why invest all that effort just to know what I’m talking about?

      Had you bothered finishing my article, you would have realised that I do know what I’m talking about.

      I don’t expect anyone who doesn’t know what a command line is to fully grok LaTeX, but if you want to go on a jihad to wipe out LaTeX by convincing neophytes not to use it, perhaps you should eschew the use of “alternative facts” and learn how it actually works & what it can actually do.

      I did learn how it works and what it can do — and had you read the article, you would have realised that. And I have no intention of wiping out LaTeX, just of helping neophytes to understand what it’s actually for. This might well result in some of them leaving it alone if the thing that it’s actually for turns out to be not that important to them. Or it might result in some of them exploring some of the alternatives that I suggest, e.g. generating LaTeX from text structured in other systems (e.g. with Markdown).

      I’m fine with using WYSIWYG word processors to write text-only papers and letters and notes to their pals, but I have suffered mightily trying to generate readable documents containing mathematics and/or captioned figures using publisher-mandated word processors.

      You see, this is the problem with commenting on articles that you haven’t actually read. I know that LaTeX is good for typesetting mathematics. In fact, I say so in my article. You know — the one you didn’t read.

      My article is written against the false claim that LaTeX is somehow better for writing text than word processors are.

      You don’t make that claim. However, many introductions to LaTeX do, and I quote examples of such misinformation in my article.

      But hey, have a nice day.

    2. Daniel’s argument, summarized, is that people should use the right tool for the job. When it comes to writing, there’s a place for word processing software and a place for typesetting software. He is encouraging you to learn how to use both, in complementary fashion. To understand Daniel’s idea, you must be able and willing to read what he says. You also must be willing to consider, even provisionally, that Daniel is saying something beyond the usual diatribe of product A is better than product B. Why don’t you surprise me and demonstrate that you are up to the task?

  7. Daniel Allington:

    Thanks for writing, and creating this post. I originally stopped by to skim it… but I ended up heartily enjoyed reading almost everything on this webpage.

    Now… what I didn’t expect to find was this gold mine of a Comments section.

    The commenters, and your responses, kept the fire alive.

    Thank you, everybody who posted here.

    I do have a question

    * Who are these people saying you must write regular text in latex?

    because honestly, I didn’t even know latex existed until I was pushed to the point of insanity trying to typeset f(x) = pi / 2 into a computer, and I kept google searching “how the fuck do I write equations”.

    You almost discouraged me from using latex entirely.

    That’s probably why everybody in comments is bashing on you, your post really does come off that way. (so does the title.)

    I personally find that any notepad program works best for me when just writing text… because then I can ACTUALLY focus on just the writing.)

    I’m glad you mentioned markdown and pandoc =)

    On a side note, you don’t have to use wikitext on wikipedia anyway (it’s :

    See: https://en.wikipedia.org/wiki/Wikipedia:VisualEditor

    I await your responses of rage,
    at those comments in which you engage,
    in this war of pro-grams that you wage,
    y u so mad, that you wrote a whole page?

    ¯\_(ツ)_/¯

    Sincerely,
    Popcrate

    P.S. You might find it interesting that this got caught in the google results: and provided a nice “related articles” link to the reference (because its search-friendly format.)

    1. Hi Popcrate

      Thanks for writing, and creating this post. I originally stopped by to skim it… but I ended up heartily enjoyed reading almost everything on this webpage.

      Thanks for your comment! I’m glad you found the essay (and comments) enjoyable. I’ve argued for years that Open Access isn’t enough: communicating inside and outside the academy requires different voices. So I make a point of writing my blogposts in a less serious tone than my research papers.

      Now… what I didn’t expect to find was this gold mine of a Comments section.

      Yes, I seem to have touched a nerve!

      The commenters, and your responses, kept the fire alive.

      That’s certainly true. I often wonder whether I should go back to my old policy of closing posts to comments after a week or so in order to cut down on spam, but then something like this will happen and remind me of the value of feedback.

      Even the antagonistic comments are valuable because they provide an opportunity for clarification. That said, there are limits. Last year, one of my posts got a comment from a white nationalist — there’s no point engaging with people like that.

      I do have a question

      * Who are these people saying you must write regular text in latex?

      There are plenty of those. See Section 3 of my essay, above. (Although they don’t say ‘you must’ — they just say that it’s better if you do.)

      because honestly, I didn’t even know latex existed until I was pushed to the point of insanity trying to typeset f(x) = pi / 2 into a computer, and I kept google searching “how the fuck do I write equations”.

      Yes, that’s what LaTeX is good for: typesetting equations. However, you don’t need LaTeX-the-computer-program for that if you’re publishing to the web. LaTeX-the-computer-program produces .ps or .pdf files for printing out, but if you’re planning to keep things in the digital realm, you might be better off with MathJax or jsMath. These take equations written out in LaTeX-the-markup-language and render them in the browser. Or you can write your equations out in MathML rather than LaTeX-the-markup-language, and the browser will render them by itself. Whichever of these routes you take, the results will look just as good. Also, you don’t need to write the surrounding text in LaTeX-the-markup-language, which is a win.

      You almost discouraged me from using latex entirely.

      High five! Just kidding. I don’t want people to stop using LaTeX; I want them to think twice before adopting it or recommending others to adopt it. It has advantages but not everybody really needs those advantages, and it also has disadvantages that are not always made clear to newbies.

      That’s probably why everybody in comments is bashing on you, your post really does come off that way. (so does the title.)

      The title doesn’t say ‘Don’t use LaTeX at all’, it says ‘Don’t write in LaTeX! It’s just for typesetting’. The reason people read ‘Don’t write in LaTeX’ as ‘Don’t use LaTeX’ is that they’ve been conditioned to think of LaTeX as a text-authoring system, which it isn’t.

      It’s kinda like if I had spent my entire life turning screws with a chisel and then you said ‘Don’t turn screws with a chisel! It’s just for cutting’ and then I got angry and accused you of being on a jihad against chisels.

      I personally find that any notepad program works best for me when just writing text… because then I can ACTUALLY focus on just the writing.

      Yes, although it depends on what you mean by a ‘notepad program’. Being able to navigate a complicated document by headings rather than by lines really helps, and so does being able to fold away the text below the headings to reveal the overall structure. You can do this kind of thing in various ways with word processors and decent text editors. I do a lot of writing in Emacs and I recently started writing in RStudio too. I guess you could think of those as notepad programs with bells and whistles. Some of the bells and the whistles are more helpful than others.

      Although text editors or ‘notepad programs’ have many things going for them, I must say that, once a text gets beyond a certain length, I weary of reading it in a fixed-width font. And that shouldn’t be surprising, by the way: it’s well known that texts set in fixed-width fonts are less easy to read (unless you’re an OCR program). To get around that problem, you either need to work in a word processor or to write in a text editor but periodically typeset your writing and then read back over it in another window. The latter option turns out to be much more of a faff than the former.

      (Note that I said ‘texts set in fixed-width fonts’. Code is obviously far more readable when it’s set in a fixed-width font.)

      I’m glad you mentioned markdown and pandoc =)

      Glad to have been of help. Btw, one thing that I forgot to mention above is that you can write in Markdown and use embedded LaTeX for the equations, then generate HTML and .pdf versions of your work from the same original. This can be automated with a Makefile.

      I await your responses of rage,
      at those comments in which you engage,
      in this war of pro-grams that you wage,
      y u so mad, that you wrote a whole page?

      As we used to say, lol.

  8. It’s worth mentioning, if you don’t know it, a lovely piece on Josh Parsons’s website, called ‘The LaTeX Cargo Cult’, in which he argues that the routine use of LaTeX by (non-logician) philosophers amounts to little more than an attempt to ape the conventions of the hard sciences and their apparently more solid disciplinary credentials: http://www.joshparsons.net/latex/

    1. Thanks very much for the link to that essay – I hadn’t seen it, but it’s so very much to the point, I wish I’d quoted it above. ‘Cargo cult’ is a perfect metaphor for what I’ve called the ‘LaTeX fetish’ – and of course it’s not just philosophers who treat LaTeX in this not-quite-rational way:

      Despite the fact that LaTeX is not easy to learn, and has few reasonable applications for philosophers apart from typesetting mathematical logic, it is surprisingly popular, especially among graduate students.

      Perhaps it’s not only philosophers who suffer from what Parsons calls ‘philosophical cringe’, and long to associate themselves with more STEMish disciplines. Or perhaps there are a range of reasons, and those introductions to LaTeX that I was ranting about above also play a role: the ones that gloss over all the problems of LaTeX and falsely claim that it helps you structure your writing. I note that they’re mostly aimed at students. And part of it may be just that fiddling with the visual appearance of a document – which LaTeX encourages you to do – serves as a displacement activity for the anxiety of actually writing.

  9. Hello, mathematician here, but one of the few renegade ones who agrees with you. Because I have a learning disability, looking at plain-text markup is especially painful for me, so I was forced to resist the LaTeX indoctrination that every serious math student is now subjected to. Even junior- and senior-level math homework is expected to be typeset at many schools.

    A few observations about equations: Knuth developed TeX more than twenty years before the Unicode standard, so the horrible concatenations of backslashed commands needed to produce mathematical symbols are now obsolescent because they can be replaced by Unicode literals. Furthermore, using real-time renderers like MathJax eliminates the tedious need to compile every time you want to see what you wrote. Since most mathematicians are still partying like it’s 1989 in this respect, they are ignorant of this as well as the fact that every word processing program in existence can structure documents just as well as LaTeX. There’s really no longer any reason to use LaTeX for anything that isn’t going to be published, but the “professional standard” is so deeply entrenched that anyone who bucks the trend isn’t taken seriously.

    1. Hello – and thank you for your witty and well-informed response. It seems to me that any attempt to push back against the LaTeX indoctrination that you describe tends to be interpreted as ‘I am too stupid or lazy to learn to code.’ That is not only unreasonable, but – as your example shows – also discriminatory.

      I hope I can make time to pull together all the new things I’ve learnt from comments such as yours. Since publishing the above piece to this blog, I’ve realised that there is even less reason to use LaTeX than I previously thought. As you show, there is in fact no objective need even for mathematicians to write in LaTeX if the writing in question is not going to be published. But as the ‘cargo cult’ essay that Josh shared in his comment shows, there’s no longer any need to use it even when writing for publication – all it really does is to cause inconvenience for the publisher.

      Perhaps if enough of us say these things in public, the situation will start to change.

      1. Compiling an authoritative resource on reasons not to use LaTeX, together with more usable alternatives for various fields, is an excellent and long-overdue idea.

        You are absolutely correct about the accessibility issue, which is one of many reasons mathematicians should switch to HTML-output formats instead of using pdflatex. Some forward-thinking indivdiuals in trendier fields have already cottoned on to this (e.g. the Stacks Project), but it’s nowhere near a field-wide movement. This is somewhat ironic given that (due to LaTeX self-publishing!) mathematicians make extensive use of electronic distribution of their work, especially through the arXiv preprint server.

        I should back off a bit on my position above: There actually is a feature of LaTeX which currently has no satisfactory direct equivalent in word processing (or Markdown + MathJax, etc.), namely, the “theorem environment.” But since theorem environments are just specially formatted subsections, there is no reason why someone (like the American Mathematical Society) couldn’t write a LibreOffice plugin or Markdown extension syntax to produce theorem environments. However, nobody is going to do this anytime soon because….well, you know the answer.

    2. “they are ignorant of this as well as the fact that every word processing program in existence can structure documents just as well as LaTeX.”
      Although not a mathematician I find that hard to believe. In principle it may be possible but one needs to create a Template that does all this things that LaTeX does very well per default. I once read that Knuth spend years in adjusting the spacing around math-formulars … to think that one can just easily achieve something like this does not seem realistic. But maybe I misunderstood your point…

      1. Thanks very much for your comment, Thomas. I believe that Elizabeth is right here: every word processing program in existence probably can structure documents just as well as LaTeX. The point you raise about spacing around formulae relates to typesetting (which is a matter of presentation) rather than to text structuring (which is a matter of content). This is what I’ve been trying to get at, although it seems to be a more subtle point than I initially realised because in many cases it’s not getting across: authoring and typesetting are different tasks, and are best supported by different tools.

        The thing is, there’s no need to make a word processor do all the things that LaTeX does well by default. LaTeX is a document-typesetting tool. Word processors are document-authoring tools.

        If a writer (especially a student) wants to ‘focus on the content of what he or she is writing without attending simultaneously to its visual appearance’ (which one of the bits of LaTeX evangelism I quoted above suggests is a good thing), then my recommendation would be for him or her to quit worrying about spacing until the document is finished, and then switch from an authoring to a typesetting tool if he or she really needs to create a PDF.

        If he or she really needs to create a PDF. I’m not convinced that PDF is all that great as a publishing format, unless you’re actually printing the thing out. But that’s another story.

        1. Again I tend to agree with the more pragmatic point and your critic of »one of the bits of LaTeX evangelism«. Also I am writing from the perspective of someone who has to create a pdf in the end – for online publications things may look different.
          But I don’t understand your distinction between typsetting and text structuring. Maybe I don’t understand what you mean with text structuring when you say that it is a matter of content.

          I think by the way that one of the fields where LaTeX should be used as a typsetting tool is in the publishing of Novels. With the tool writer2latex I created really beautiful typset books out of odt. Files »on the fly« so to speak. For these kind of texts LaTeX is superior to Indesign in my opinion because one can produce the pdf result incredibly fast and can change the page settings than very conveniently.

          1. Text structuring is a matter of identifying (say) which words are the heading of a chapter, and which words are the text of that chapter, or which words in the text of that chapter are in one paragraph and which are in another. Typesetting is a matter of (say) centring and emboldening all the chapter headings, and indenting the first line of each paragraph. Text structuring is best done with authoring tools such as LibreOffice Writer, Microsoft Word, Scrivener, etc. Typesetting is best down with typesetting tools such as InDesign, LaTeX, Scribus, etc.

            If I’ve understood your workflow correctly, what you’re doing is authoring your work in LibreOffice Writer, and then using writer2latex to convert the resulting ODT files into LaTeX for typesetting. That is exactly the sort of thing that I advocated doing in the essay above – although the conversion tool I recommended was Pandoc rather than writer2latex. It’s possible that writer2latex does the job better than Pandoc, so I’ll check it out: thanks for drawing my attention to it. As I explain above, converting from ODT to LaTeX via Pandoc is not as straightforward as one might wish!

  10. I really enjoyed reading this long post and parts of the discussion, too. I tend to agree with the general point that there is something like a cult or fetish when it comes to LaTeX – with all the pretensions an snobbish contempt for people that are not using it.
    I think there also can be not much discussion about the pragmatic point: when publishers want an rtf.- or docx.- file one shouldn’t write in LaTeX.

    But you have the tendency to be a bit unfair on two things:
    a) how hard it is to read LaTeX code: I think it is pretty easy to prove your point when you use the terminal version of emacs with a monospaced font – the black background is very helpful for that, too. But when you use a gui-editor like texstudio for example and adjust it a bit the text does not look so bad – but of course it is not like word or libreoffice. In the end it comes down to how important WYSIWYG is. And people are just different when it comes to that.
    b) how hard it is to learn it: I am a philosopher (for mathematicians it may be different)– and I just cannot agree with you here. I think that you can learn the commands that are sufficient for most of the stuff people need in this field in 5 minutes – I meant stuff like: text-formatting (italics, bold), sections, quotation, footnotes, title, tableofcontents, lists with items – that is basically it.
    But of course: people don’t use it like this – they get into the details and become apostels of the fetish…

    1. It’s not just about WYSIWYG vs. markup. If you’re someone whose writing process is pretty much the same on a computer as on a Selectric, then doing everything with a text editor makes perfect sense. But many people actually *like* the “bloat” of a word processor or (worse!) an integrated writing environment.

      I have trouble writing anything at all unless I can import/export with a mindmap, color-code my text, add hyperlinks to other applications, drag & drop images, etc. These are all integral parts of the “process content” of my writing (as opposed to “product content,” both terms that I just made up), so being forced to use a text editor seriously inteferes with my ability to work.

      As for typesetting mathematics with LaTeX, the party line is that you get used to it. But the honest people will own up to writing anything nontrivial on paper (remember that?), transcribing it into LaTeX, and then compiling–usually more than once–to be sure the LaTeX-gibberish matches the handwritten mathematics. And don’t even get me started on LaTeXed diagrams.

  11. Thank you for writing this well reasoned and thought provoking essay! I guess I can subscribe to the main takeway of your essay (although I am not entirely on board in all your conclusions). And I know I am late to the party and would otherwise have refrained from commenting, but there are a two points which I missed in this discussion so far and find important enough to weight in. These points would be collaborative writing and, relatedly, copy and pasting (because copy-and-pasting that research paper you wrote a year ago is, in a way, also a collaborative writing process). And there will be a quote from the movie “Quigley Down Under” somewhere below, so be alert: spoilers ahead!

    I am writing this from the perspective of a computational physicist who spend the last 20 years or so of his life rewiring his brain in a particular way so that working with source code – no matter what language or purpose – feels natural and everything else is … not difficult but definitely not natural. Or, put in another way: If you’re getting really good at hammering, things that tend to look a bit like nails cross your way at their own risk. I know I am odd that way and I hope that I can restrain myself from forcing this particular point of view onto everybody else. I really hope.

    In my current job, I tend to often participate in writing proposals for research projects whose most endearing features are that they are typically in the end 200-odd-pages long and which are strictly regulated with respect to structure and content. They are usually created in a big hurry and under great pressure with little regard to style and aesthetics, they typically represent a cut-and-paste best-off of the last research papers and proposals written concerning this or vaguely related subjects. They also tend to be written, edited and then spliced together concurrently by a group of persons in different time zones, working with different versions of word processing software on different operating systems with minimal coordination. And, for whatever reason, they are typically a complete mess. Especially references, text marks, bibliography, captions to tables and figures, everything which is not plain text.

    The proposals are usually required to be written in MS Word (or rather: they should be submitted in the arôme du jour of OOXML and yet, in striking similarity of what you describe in your essay, everybody thinks this, positively, means they should be actually written and edited in whatever version of MS Word is at disposal; But I digress). And even though recent versions of MS Word offer tools for collaborative writing, nobody in these teams of scientists (mixed teams with STEM background) under pressure typically uses them. Partly because they never found it important enough to learn them (because it “works” also if you do not know or use them), partially because they are not always suited for the task and not (imho) exactly intuitive either. It boils down to a copy-and-paste frenzy and one or two guys doing an all-nighter in the end to fix all the dangling references and stuff.

    I tend to comment in the later stages of these activities that if we would apply the same approach to writing software than we apply to write certain documents, nothing would work. I am usually rather snarky at this point, with very little patience. I am also usually mostly talking to myself at this stage. Go figure

    There is a rule-of-thumb that I like to annoy people in my environment with, if they listen: If you design a software (or build a tool or a measurement device; or setup a process or a procedure in your organization, etc.) it should be easier and more convenient to do it right than to do it wrong.
    Note that it should not necessarily be “the easiest” to do it right. Local minimum is usually fine.

    If people copy-and-paste LaTeX (or, for that matter: Markdown, (X)HTML, etc.) they have a good chance of taking the content together with the relevant context. It’s not perfect because there is a good chance to miss something or even to override some previously introduced label or similar by some autopiloted copy-and-pasting but since the resulting combined document is interpreted in the same way as a fresh document, written from scratch, would be interpreted, there is a fair chance of either ok-ish results or at least horribly and non-intuitive error messages.

    The guy who pasted his text with hard-coded bibliography markings into the larger document the other day had not spend a single thought on how his reference [32] would magically find its way into the bibliography at the end of the proposal. He consequently also applied the same strategy and rationale by labeling his figures and tables. In a way, MS word (or whatever it was he was using) made it easier for him to write something that looked correct to him than something which would be correct in the great scheme of things, e.g. the final document.

    Now, don’t get me wrong: People whose approach to creating documents is along those lines will have a hard time writing something more complex than a shopping list regardless of the involved tools. And problems with management, financing, staffing, etc. which all contribute to the hollapalooza described above can not be cured even with the perfect software package for writing proposals.

    But, and this is my second point: The mere fact that with text based markup/input files we can – trivially – split a complex document into individual files, write, edit and combine them in any way or form we like during the process of document creation by mere “including” them (you know, like when you write software) and get exactly the same document as if I would write a single monolithic file; the fact that we can have a proper version control (with granularity of diff’s and changes at the file and then at the character level and not at the group-of-XML-tags-in-OOXML-file-seems-to-have-changed,-oops level), including the comments; the fact that we can meaningfully branch and merge, the fact that we can formalize push/pull workflows on top of that, … : It would help a lot, not in writing but in editing and compiling documents as complex and hairy as these dreaded proposals.

    (I know that SVN for example can version control MS Office files and delegate the diff view to the proper program. But it’s a far cry from what would be possible if you could split the damn thing into smaller, managable chunks and let a kind of compiler do all the stitching and splicing. Sigh).

    My own apporach to WYSIWYG editors in comparison to approaches like LaTeX (or markdown) can probably be best described by this:

    There is this movie “Quigley Down Under”. It’s a Western-type movie, set in Australia where the protagonist, an oddball American called Quigley who excels in shooting things at insane distances with uncanny precision using his trusty sharps rifle. When asked about his choice of rifle (as compared to the more ubiquitous and apparently more practical revolvers), he always replies that “he has no use for them”.

    In the end, the mandatory showdown takes place in the form of a duel with the main villian and his two henchmen. And, in a cruel twist, they only him a revolver to defend himself. Which he draws so quickley as to shot all three of them before they can even blink.

    (It’s a family movie, obviously)

    The reason I am bringing this up are the words he utters to the dying villain in the dust:
    “I said I never had much use for one. Never said I didn’t know how to use it.”

    I try to be fluent enough in both realms to be frustrated by what my colleauges do. But if I had my way and would write on my own, then I know which of the two approaches I prefer :-)

    Again, thank you for your time and your efforts.

  12. This is a good pragmatic discussion of the difficulties of using LaTeX. I have acquired a “just good enough” proficiency with it to get by tweaking existing templates for documents where I find the final presentation to be most important (like resumes). I could devote more time to learning LaTeX, but using it at all for my purposes is somewhat of an overkill, so that wouldn’t be pragmatic.

    I like the suggestion of using Markdown and then a conversion engine to convert that Markdown to LaTeX. Right now, I keep notes in Evernote using Markdown and the “Markdown Here” extension to convert that markdown to formatted Evernote pages when I want to read the notes in a nice format. This is a very pragmatic alternative to say learning Emacs and using Emacs org-mode for note-taking, even though I’d really like to devote the time to learning it.

    I find writing in Markdown to be the least unpleasant experience…Evernote even with Markdown still leaves something to be desired in terms of tables and tables not allowing for formula input (this is from what I can tell one very nice feature of Emacs org-mode).

  13. As a mathematician who has (been forced to) spent a lot of time learning LaTeX (and now has a very minimalistic approach to it) I would be more than happy to have an easier alternative!

    I would never use LaTeX to write a non-mathematical text, however so far I haven’t seen any reasonable alternative for texts with a more than minimal amount of mathematical formulas…

    does it exist out there ? I’d be happy to know!

    1. Thanks for your comment! There are a few simpler options that I am aware of. One is to write your text in LibreOffice but insert LaTeX equations using the TexMaths extension:

      https://extensions.libreoffice.org/extensions/texmaths-1

      Another option is to write in Markdown (which is almost plain text and certainly better than LaTeX for writing prose in) and incorporate LaTeX equations where required. RStudio, Scrivener, and Markx all support this. RStudio’s the only one of the three that I’ve used and it has problems of its own, but I’ve been hearing really good things about Scrivener.

    2. I feel that for professional-level mathematical writing, TeXMaths doesn’t quite do it. It’s fine for inserting display equations here and there, but it’s too awkward if you’re making a lot of use of inline mathematics.

      My split-the-difference solution is to use LaTeX via MultiMarkdown (I like Scrivener, but there are other options) and use Unicode literals to write the mathematics so that it’s human-readable without compiling. But this means loading the unicode-math package and compiling with XeLaTeX or LuaLaTeX instead of pdfLaTeX. Although there are numerous advantages to using a more modern compiler, the disadvantages are (1) it will probably annoy your collaborators, and (2) the arXiv still doesn’t take XeLaTeX source, although they seem to be coming around about that.

      1. More good advice! If I can make time, I hope I’ll be able to go through all the comments on this page compiling a list of the most helpful points to republish.

  14. The best analogy that I’ve come up from reading this article is as follows: using LaTeX is like editing a Rich Text Format (*.rtf) file using Notepad.exe, while a Word Processor is like editing the same Rich Text Format (*.rtf) file using Wordpad.exe. Obviously, Notepad.exe is the default plain text editor for Windows, and Wordpad.exe is the default and minimal word processor for Windows.

    PS: This reminds me, in the late 90’s, MS Works had a hard time opening Rich Text Format directly. Was it import, or something?

    1. Funnily enough, I remember creating a Rich Text Format file in Notepad, some years ago, and adding all the markup by hand. I can’t remember why I did it. It wasn’t too bad an experience, but when I opened the file with a program that was actually designed for working with RTFs and then saved it again, it became virtually unreadable in Notepad because of all the extra markup that the program had inserted. In WordPad or Word, of course, it looked exactly the same, because all that markup was hidden. Which is how things ought to be, really.

    2. There are two persistent misunderstandings in discussions of Word vs. LaTex. Daniel has disabused us of both, but some of the comments continue to exhibit them. One is that all editors are essentially alike. The other is that Word is intended for novices and simple documents. Here is a long-winded explanation of why they are wrong and misleading.

      There are two kinds of software for composing and editing text documents. One is a type-stream, which conceives of a document like water coming out of a faucet — one character at a time. Characters produce words; words become sentences; sentences comprise paragraphs; paragraphs are grouped into sections, and sections are organized into documents.

      Text editors are type-stream software. Formatting is applied principally at the character level by entering codes directly or indirectly. Secondarily, codes are applied at higher organizational levels. Not all type-stream software is alike. Plain text editors such as Windows’ notepad.exe lack the capability to navigate characters, words, sentences, and paragraphs with efficiency. This is one way that type-stream word processing programs such as WordPerfect and NotaBene are superior. This difference gets conflated with the discussion about whether the LaTex coding language is superior, for instance, to WordPerfect’s or NotaBene’s. You can see the embedded commands in the latter from their reveal codes views.

      In contrast, knowledgeable users of Microsoft Word compose and edit text in a structured fashion. A document is divided into sections (by default, Word documents comprise 1 section); sections are divided into paragraphs; paragraphs into sentences; sentences into words, and words into characters. Every level of the hierarchy has distinct attributes. The attributes together comprise what is called a style sheet or template. You can modify document attributes directly at any level, but at the paragraph level and above capable Word users generally (though not always) elect to modify the template.

      The advantage of this method is that once you have settled on a template, for instance, for writing a business letter, a memo, or a paper, you treat the document like a type-stream and apply the appropriate formatting when it suits you. To assist your work, MS Word and other word processing programs allow you to compose and edit in draft view, which allows you to use a draft font (if you wish) and text that wraps to the window. NotaBene gives you more viewing options than Word. However, Word’s outline view is one of its most valuable features, especially for large documents.

      In choosing between Word and Latex, you are selecting first and foremost structured or type-stream software. In choosing between Word and other structured applications such as LibreOffice, you are selecting between ways of structuring and editing at the paragraph level and up. In choosing between LaTex and other type-stream software, you are selecting between ways of coding and navigating a document. In all cases, you are deciding how much money to spend. All too often, you are also deciding which team colors to wear.

  15. I do agree that writing in bare naked LaTeX doesn’t make much sense. That being said, you don’t need to! You can just use LyX, which a graphical editor for LaTeX (but one that stills keep content and presentation separated). Of course it lets you still use the markup when you feel it’s best (writing math in LaTeX is faster than using any graphical editor IMHO) and also lets you use your OS’ fonts very easily. Seriously, LyX is probably the single best piece of software I have ever used, and it’s completely free and open source.

    1. Sounds good. As I’ve said a few times on this page, I’m open to various solutions. I probably ought to have included LyX in the list of alternatives to what you amusingly call ‘writing in bare naked LaTeX’. There are a lot of those now, and some of them are really good. But that’s kind of the point of my article.

  16. Thank you very much for your well thought-out article. I am very glad that there is a choice of tools, so that people can use whatever they’re most comfortable with and I fully agree that people should not be forced to use tools that they dislike. As you note, thanks to the amazing pandoc, we can relatively easily convert from one format into another.

    That said, I personally strongly prefer LaTeX to conventional WYSIWYG editors. LaTeX’s markup is not verbose enough to be painful (like HTML is, which I would not write manually in any large amounts), but is expressive enough to represent the semantics of the document (as you write, it is possible to have a semantic document in Word/LibreOffice, but it requires greater self-discipline). I enjoy the fact that I can write anything, just using a keyboard, without having to use the mouse to click on buttons etc. I appreciate using tools that I have some possibility of understanding. Finally, with AUCTeX and preview-LaTeX in emacs, I can have a pseudo-WYSIWYG experience, when I need it. (With writeroom-mode, which results in the text being given margins on both sides and having a nice, readable width, I actually prefer to read the LaTeX version of my documents in emacs, than their compiled version in a pdf-viewer. I’ve even mulled writing an arXiv LaTeX source importer, so that I could read other people’s articles in this way. I might be weird…)

  17. idk bro I spent like 4 months learning latex and been using it ever since, wasn’t hard really, you’re prolly just mad ’cause ur a noob lol

    1. I’ve been using it for about six years. But if you had read my article you’d have seen that how hard it is to learn is far from being my main concern. The big problem from my point of view is that it is sold as a way of helping writers to focus on content and forget about presentation, when in reality it’s anything but.

  18. I think your conclusion is way to broad. Since you don’t have a background where lots of formulas are required, this is kind of understandable. However, there are reason why LaTeX is the preferred format for papers in e.g. maths and physics.
    I tend to agree that for simple texts with few formulas, one might use a simple word processor or markdown with interspersed latex formulas (or something similar). However, at least in my experience in physics, you will sooner or later face some nontrivial problems with these tools that you can’t solve easily because they are ‘just’ markup languages with formula/LaTeX tacked on top.
    For example (I am currently writing a physics PhD thesis with LaTeX), I define semantic commands like \magneticFluxThroughSurface[\surfaceElement] for recurring symbols. To me, this approach has two main advantages. First, redefining the symbol in the whole document is very easy and amounts to changing one line in the header. Second, using this textual representation of the symbols makes proofreading a formula/proof easier since I can quickly grasp the meaning without having to decipher a host of greek letters. Of course, having a live preview of the rendered formula next to the latex source would be/is even better.
    LaTeX also really shines is scientific text with lots of figures and graphs. While sometimes painful, it is possible to generate many figures directly from within LaTeX. As you can use the macros of the main text, all symbols in all the figures match the style of the remaining document. While you can achieve this by other means, you will often find yourself redoing figures after some data or symbols changed or when you want to highlight some aspect of the plot.
    I would also argue that a simple text-only representation of documents makes it easy to use version control (so you have all past iterations of your document available) and collaborate with multiple coauthors. In my limited experience, writing a word document with multiple people caused a lot of extra work for us compared to say latex + some version control system.

    To conclude: yes, LaTeX is old and has its quirks; yes, LaTeX sometimes causes a lot of extra work; and yes, for simple texts you better use a simpler program. However, there are legitimate use cases where latex beats any other free system that I know of like collaboration with many coauthors, easier version control, and an overall unified look for documents with lots of formulas and figures.

    1. A very useful comment, and I wouldn’t argue with any of what you say. As I’ve said in my replies to a lot of other comments, my real beef is with introductions to LaTeX that sell it on pretended advantages instead of on its real ones.

      The only thing I’d quibble with is your implied definition of a ‘simple text’ as one without many equations in it. A document can also be complex in the sense that it has a complex structure. LaTeX markup isn’t an especially good way of representing a document’s structure. It’s really a repurposed presentational markup language.

  19. Over the last 20 years I have written a significant amount of technical documentation in both ms-word and LaTeX. And I must say that I vastly prefer the latter for several reasons that I will cover later.

    First of all let me agree by saying that one indeed does not *write* in LaTeX.

    Writing LaTeX is usually done in a *text editor*. Personally I’ve switched to vim several years ago after having used emacs for more than a decade; emacs was getting too slow for my taste. But which text editor suits you is a highly personal choice. A good text editor with the right macro packages can work wonders. Like every tool however, it requires an investment in time to learn to use it. (The same goes for word processors, by the way.) For example in my LaTeX editing I make heavy use of snippets (using vim’s ultisnips plugin). These are letters or words that can expand into text with the touch of the tab key. For example I have a snippet for section headers. Not only does this snippet wrap the section header that I type in the right LaTeX command, it also automatically creates a label for that section based on the section title. I have also defined snippets for creating figures and tables. For myself, I find that the color scheme that I use also has a significant influence. Again personally I like the Solarized Dark scheme. Features like search and replace with *regular expressions* are to me also very valuable tools that a text editor offers.

    But if an editor is not your thing, there are several GUI’s for Tex, like Texmaker and LyX. As I have not used these myself, I cannot say much about them except that you might want to check them out.

    Second, let me state for the record that for anyone who is happy with the feature set of ReStructuredText or Markdown I would recommend to use those over LaTeX. Especially since those can relatively easily be converted into other formats. And if you’re happy with a word processor, you should keep using that too.

    With regard to fonts in LaTeX, these days I would suggest to use “xelatex” with the fontspec package instead of plain “latex” or “pdflatex”. With that you can relatively easily use basically any font you have on your system.

    Using text-based formats like ReStructuredText, Markdown or LaTeX has several advantages over word processors. Since I’m not sure how this comment box does formatting, I will list them as separate paragraphs and mark the relevant points between **.

    *Universal.* Text (in this day and age that means unicode) is a pretty universal interface. Every word processor or typesetting program can import plain text. And it you don’t like the TeX markup commands, you can use “detex” to remove them.

    *Revision control.* While ms-word can track changes in an editing session and it can merge two different versions of a document, it is a poor substitute for proper revision control in my experience. Generally, I keep all the materials for a LaTeX document in a directory tree, which is kept under revision control with git. This means that if I accidentally delete a file, I can just check it out again and I won’t have lost all my work. Additionally, I can easily see when I e.g. added or changed certain text or formatting.

    Now there is nothing stopping you from using git to check in word processing documents. But part of the power of revision control is lost when using the opaque binary formats as found in word processors; in general you cannot see differences between two revisions.

    *Document size.* I routinely produce documents of 200-400 A4 pages in LaTeX, where it is not uncommon to have hundreds of images and tables. Over the years I have tried several times to “port” such documents to ms-word only to have it crash and corrupt the file partway through.

    *Automation 1.* My documents often include the results of calculations. These calculations are typically written in a suitable programming language like Python or Lua. (The calculations that I do often are a poor fit for a spreadsheet because of size and precision issues.) The output of these programs (which might be plain text of include LaTeX commands) is included in the LaTeX document using a simple “\input” command, while typically the listings package is used to include the program itself in the appendices. Of course you could copy and paste both the output and code into a word processor. But what if the program changes? You’d have to do it all over again.

    *Automation 2.* I use the “make” program to build my LaTeX documents. And actually the *last* thing that this does is typeset the LaTeX code. It first runs all programs to create the necessary calculations, figures and graphs. Then it uses the “latexmk” program to actually call LaTeX and friends. If you can use something like “latexmk” it will certainly be helpful. It runs LaTeX, bibtex or biber and makeindex in the right order and keeps doing that until all references are solved or a number of runs is exceeded.

    Now word processors on the ms-windows platform like ms-word can embed other documents like spreadsheets. Unfortunately this embedding is done using absolute paths. So anytime you move the documents or send them to somebody it will break all links. Not fun. There is a workaround for that but it involves digging into field codes and macros.

  20. When I start writing documents digitally (many years ago) I start, as many others,
    with WYSIWYG traditional world processor and I’m not happy: I see good-looking documents around my life and I can’t reproduce anything comparable to them. I start “investigating” and found some “professional” typesetting WYSIWYG software: they were a PAIN, a super-long learning curve and a real wast of time for writing any document. After I meet LaTeX and DocBook, the latter seems to be easier but does not produce anything as good as LaTeX, the former it somewhat
    frustrating but with “pre-cooked” template produce superbe result with very little
    effort. Many year after I still use LaTeX. Now I have my own template for letters, CVs,
    thesis, articles, even for printing mails and bottle labels! Learning LaTeX was perhaps as fast as a
    traditional WYSIWYG wp but was never be a problem: start from precooked templates
    let you produce high-quality documents fast, as you feel the need or curiosity you
    learn “growing” without problems when you have time.

    But you are right, LaTeX is about typesetting. So for the rest? My personal choice is
    org-mode (and yes, I tried to export to LaTeX but is longer and harder than write
    plain LaTeX with yasnippet). I use it for nearly anything, spreadsheet included, far
    simpler, integrated and powerful than any WYSIWYG counterpart. Emacs calc/imaxima
    and the power of any org-babel-supported programming/scripting language is something
    you can only dream with WYSIWYG products, integration of all this stuff in a pure text,
    easy to read document is something you can only dream with WYSIWYG suite.

    There were also some development effort to “easy” LaTeX typesettings (for instance ConTeXt) but they not became widespread as “WYSIWYM” editors such LyX or TeXMacs, IMVHO because they are in fact less flexible than plain LaTeX and plain LaTeX is easier enough, at least it’s give you a easy and simple “learning path”. This concept is alien in today’s world but is a must to grow, to acquire knowledge. Average LaTeX, Emacs, Vim, Zsh, … users grow more or less constantly their knowledge far beyond their software, average WYSIWYG users after years have nearly the same amount of knowledge of the first year with their “favorite” software and are generally nor invited nor willing to learn anything else.

    The long story short, I’d like to tell anyone ask me about LaTeX: “you can choose between grow as a Ford model worker or as a scientist. The first path is easier at start but is nearly impossible to escape or grow, the letter is much longer, it’s start is slow but this slowness does not generally pose problems, sometimes you may feel frustrated sometime you may feel happy, and you always grow. So the choice is being a child in an adult environment or be an explorer. Make your choice!”

    I apologize for my poor English.

    (1) I HATE slides since IMVHO if you need them probably means you do not know
    what are you talking about, but unfortunately they are often requested…

    1. Thanks for your comment. And I agree that org-mode is a good choice for writing in. Personally I prefer Markdown but it’s a very similar experience. (The main difference of course is that for org-mode you must use Emacs. Emacs is my favourite text editor, but I don’t want to inflict it on my students!)

      1. Thank you for the article :-)

        IMO Emacs is not about “inflict” but about teach, the same is valid for zsh, for (g)Vim, for minimum scripting abilities, for Python, for Scheme, for LaTeX etc. My “whole” point is that quick learning normally produce short-term useful results, calm/deep/long learning produce mid/long-term useful and far more valuable results.

        For instance I do not use Emacs only for org-mode but for feeds (elfeed), ng (gnus), mails (notmuch), contacts (a bad mix of BBDB/Hyperbole and org-mode), notes (deft), agenda (org-agenda), accountings (ledger-mode), … and this is the most productive and comfortable way of work I found. The same for the desktop (i3) etc. Of course if someone give me all this setups twenty years ago at the beginning of college I probably remain at least “perplex” and probably not so “productive” for many months but it’s a path and IMVHO teaching paths is far more important than teaching “notions”. I arrive at Emacs after years of gVim, I arrive at LaTeX after years of traditional editors; if someone teach me “a good path” years ago perhaps I obtain my actual solution years ago with far more pleasure and know perhaps I learn far more interesting things. So I may have grown far more. That’s is valid also for “modern stuff” of course simple things like proper merging A4 scan of an A3 sheet (Fotoxx in my case), proper remove red eyes from photos (GiMP in my case) etc are “small potatoes” but ease the life and can be learned easy if someone give you the “opportune” tip…

        1. Emacs is my favourite text editor but it would obviously be counterproductive to impose it on my students. If I’m teaching — say — a course on statistical programming with R, then my students will come with a reasonable expectation that they are going to spend their time learning statistical programming with R. They’re not going to want to have to learn to use an idiosyncratic text editor from 1976 first just because teacher says so. (And pointing out that you also can use it for manipulating files, reading emails, and playing MP3s isn’t going to cut much ice because they already have other apps for these things.)

          So, much though I like Emacs, ‘inflict’ is the right word for what I’d be doing if I made my students use it. Something like RStudio is a much better choice for teaching purposes (or at least it would be if it worked properly on a PC).

  21. Hi, I also liked very much reading through the article and clearly identified myself as one of those fetish LaTeXians ;-) and I can clearly state why and where I use LaTex, no one else in the comments has mentioned it yet:

    I use it to write all my letters, using the scrlttr2 class from the Koma-script package. This produces good looking letters, where the address always fits into the windowed enveloped and the folding and punching marks are always 100% correct. Something I never achieved on any word processor so far ;-)

    I must say, I am a somewhat evangelistic Linux user, forced at work into a Mac or Windows, but by using a git repo for the .tex files, I can just anywhere fire up a text editor and produce a professional looking letter without having to think about correct formatting. Of course some templating work and the use of include files was involved, but once I had it, it started to work like a charm.

    On the issue of using word processors I am here with the author and in my experience most people are simply using it wrong, as they tend to play around with the options for formatting without a clue about proper settings. How often have I seen people manually formatting their text with fontsize and Bold/Italics into the desired form of a title or headline instead of using the templates…
    I use a word processor just like I use LaTeX, I write a text and then I mark it as what it should be. This said, yes, I am a fetish LaTeX user, but only where it really fits for its purpose :-)

    Thank you for this great article and some real appreciation for all your answers in the comment section :-)

    Cheers, Stefan

    1. On the issue of using word processors I am here with the author and in my experience most people are simply using it wrong, as they tend to play around with the options for formatting without a clue about proper settings. How often have I seen people manually formatting their text with fontsize and Bold/Italics into the desired form of a title or headline instead of using the templates…
      I use a word processor just like I use LaTeX, I write a text and then I mark it as what it should be.

      Yes, me too! The biggest problem with word processors is that most people use them badly. These days I generally write in Markdown or in a word processor, but if I’m using a word processor I use it as if I was using a markup language (i.e. with the same discipline).

  22. Hello
    Thank you for sharing about Latex and you are mostly right I think about the argument that its main interest is the Typesetting. But you didn’t mention about equations and this is where Latex for me is still interesting. My solution is to write my text in Libreoffice and use the Latex extension Texmath. I think that this tool should be integrated in the LibreOffice distribution :-).

    Thank you for your advises regarding the odt to Latex convertion. I’ll give it a try.

    Another interest for using Latex ? Well maybe versioning that is quite difficult with odt, non ascii text … but I have not invested in exploring this myself.

    1. Thanks for your comment. By the way, I did mention equations! Also a lot of the comments (and my replies to them) have discussed equations. See Elizabeth Henning’s comments in particular.

      You make a good point about version control, but I should point out that version control also works with Markdown and with genuinely plain text. You don’t have to use LaTeX to use it.

  23. Thank you, Daniel, for writing this essay. It was a pleasant read.

    In the above post and in some of the responses you are sort of agreeing that LaTeX is best typesetting system and best for producing printable outputs – PDFs etc.

    The only thing that made me uncomfortable is that, while you are agreeing to LaTeX’s typesetting capabilities, you are comparing it with a word processing system. This is an unfair comparison. Word processing systems like MS Word and LibreOffice should be compared with other word processing systems or other WYSIWYG editors like Lyx.

    Content is valuable. Whether it is creative writing or research theories. I feel that it is always better to keep such work in pure text, inside a versionable repository then keeping it in a proprietary format. I don’t need a specialized software even after a century, to extract and grab the knowledge in it. That cannot be said of a proprietary format. This is where LaTeX stands out and not because it isn’t good for editing. It isn’t a WYSIWYG word processor. It isn’t for writing.

    1. Thanks very much for your comment!

      The only thing that made me uncomfortable is that, while you are agreeing to LaTeX’s typesetting capabilities, you are comparing it with a word processing system. This is an unfair comparison.

      True. But the point I was making is that introductions to LaTeX often begin by comparing it to word processing systems, and they do so in a way that is unfairly biased against the latter. The things they typically say that word processing systems can’t do but LaTeX can are things that word processing systems can do just as well as LaTeX.

      There is at least one thing that LaTeX can do much better than any word processing system, which is making nicely formatted PDFs for printing out. But this is not a very important goal for many writers, especially students.

      Content is valuable. Whether it is creative writing or research theories. I feel that it is always better to keep such work in pure text, inside a versionable repository

      I agree. But that’s more of an argument for using Markdown (or actual plain text) than for using LaTeX.

      I don’t need a specialized software even after a century, to extract and grab the knowledge in it. That cannot be said of a proprietary format.

      I’m not sure about that. Take a LaTeX source file from the 1980s and try to compile it using a recent version of LaTeX. It seems to me that for the sake of future-proofing content, the best thing would be to store it in genuine plain text, or something that’s as close to it as possible, and not in an elaborate markup language whose main purpose is to produce PDFs (which we almost certainly will not be using in a century’s time!).

      It isn’t a WYSIWYG word processor. It isn’t for writing.

      I’m glad we agree!

    2. Daniel may disagree with my interpretation of his essay, but I did not read it as claiming that LaTex is “best for producing printable outputs.” Any well-written software program can produce a document for paper editing. The combination of printed output and software with a WYSIWG view makes for easy editing on screen. That’s an advantage of WYSIWYG that LaTex users seem to overlook. Instead, they focus on composition and insist that WYSIWYG gets in their way. However, they seem not to understand that current word processing software allows writers to switch quickly and easily, without compiling, between draft and layout views. Experienced Word users also laud its outline view, but understanding how to exploit these capabilities requires a level of skill or interest that vast numbers of LaTex users seem to lack. The trade-off is that switching views with a keystroke requires a WYSIWYG feature. To emphasize: WYSIWYG is an optional feature; it is not a type or kind of software. I haven’t conducted a survey, but suspect that experienced word processing users generally compose in a draft view and edit in layout view. This tack is not available to those who insist on keeping their work “in pure text, inside a versionable repository.” The irony is that these self-proclaimed pursuits are relying on high-tech computing equipment and binary software. Real purists use pencil and paper. Finally, to the comment that “content is valuable,” I say that the value of content depends principally on the writer, less on the tool.

      1. Thanks very much for your comment, Paul!

        Just a quick response to your first point:

        Daniel may disagree with my interpretation of his essay, but I did not read it as claiming that LaTex is “best for producing printable outputs.”

        The statement you quote is ambiguous. It accords with my essay in the sense that I argue that LaTeX is better for producing printable outputs than it is for writing in. I didn’t, however, mean for my essay to suggest that it’s better for producing printable outputs than all other software applications.

        My response to your other points is simple agreement.

  24. Back in the ’80s, this dinosaur discovered General Markup Language, which were macros for the Document Composition Facility aka Script.

    Dead easy for a beginner to pick up the simple markup language, and adapted beautifully from impact printers to laser and ink jet.

    Styles were implemented with a Profile file. The same document could be formatted for dumb screen display, impact or page printer. Mathematical formulae could be included with extensions.

    Bookmaster could produce a publishable manual from the file.

    I implemented the Xerox Publishing Standard in GML. We could produce technical manuals for our software in a fraction of the time it took another group to encode them in XICS, great for typesetting, but a pain for authoring.

    WYSIWYG is a giant PITA because you have to interrupt your word flow to mouse over to wrestle with the GUI to pick out what you want.

    The art of writing is rewriting. There’s a big difference between looking at a screen and a printed page. Formatting and looking at the proof reveals all sorts of stuff that needs revision.

  25. For pure prose writing(not mathematics or scientific), there is nothing that cuts down on fiddling than writing in Markdown. When I did write with LaTex(picked it up because it sounded like fun), I certainly enjoyed the output. But, I spent time making sure the page looked good while I was writing it. This is clearly the opposite of separating content and layout. With Markdown , the mark up is so simplistic, it requires little effort to write or read. This is not the case with LaTex, where sometimes I would spend hours trying to find the write magical incantation to get what I wanted. With Markdown, since I have no control of styling while writing, I just continue writing. I try and stay away from Word processors (until the final edit), because I just like to play with the settings.

    1. Yeah, this is one of the most important points for me. LaTeX does not separate content from layout. So if the need to separate content from layout is a good reason for not authoring documents in a word processor (as so many introductions to LaTeX will tell you), it’s also a good reason for not authoring them in LaTeX. I mostly write in Markdown now (and I generally compile to HTML rather than PDF because there’s no real use for PDFs unless they’re going to be printed out). Fine-tuning LaTeX markup or settings is just as much of a time sink as playing with the layout and settings on a WYSYWIG word processor.

  26. I never came across a more elaborated LaTeX critique which came to totally different conclusions I came to.

    My answer is quite lengthy and I needed some formatting as well. Therefore I wrote my answer as a blog article on my own blog: http://karl-voit.at/2017/08/26/latex-fetish/

    Please do read it there.

    1. Thanks. It would have been nice if you’d made some sort of an effort to understand my arguments rather than just assuming that I don’t know what I’m talking about, but this presumably isn’t your first language so I guess I shouldn’t expect too much. You might want to have a look at some of the other comments on this page, though, if you ever check back here, because I’ve been through the points you make in some form or other several times now (as have several of the commenters).

      All best,

      Daniel

  27. Daniel,

    Thanks for this. I wrote my Master’s Thesis in LaTex, and I’ve returned to school for a doctorate, so I was Googling the state of LaTex. I really needed it for setting formulae, but I’m hearing your point that I don’t have to write everything in LaTex to get set formulae at the end.

    I’m also super optimistic about MathJax, which has arisen in the intervening decades. I’m very excited about rejoining the world of academic publication.

    Ion

    1. Thanks for your comment, Ion. And yes, that’s exactly my point: you ‘don’t have to write everything in LaTex to get set formulae at the end’.

      Best wishes for your doctoral studies!

      Daniel

  28. Daniel,
    An interesting essay and point of view, but one that has a very blinkered viewpoint. Both TeX and LaTeX were originally designed for writing scientific documents, i.e. documents containing lots of mathematics, symbols, chemical formulae etc. If your writing does not require any of these, then it is arguably quicker to use something like Word (a piece of software that has improved dramatically since TeX and LaTeX were created). Using the equation editor in Word is a painful experience (less so than it used to be, but still painful) and is a great distraction.

    I agree with some of the other commentators on here, your article does contain some statements of dubious veracity. As just one example, Computer Modern is by no means the only font available in a LaTeX system and it is trivial to change fonts. With Word, there is only ONE mathematics font, so using something other than the default text font is not good, whereas with LaTeX there are multiple mathematics capable fonts.

    Also, your complaints about installing TeX systems do not in any way shape or form agree with my experience, or that of the many students (from those who know the innards of a Unix system to those who know how to switch it on and use Word) I have taught, all of whom have installed TeX systems (ProTeX or MacTeX) on their computers without a single hitch; it’s been seamless every time and that’s been the case for many years.

    1. Thanks for your comment, Adrian.

      I agree with some of the other commentators on here, your article does contain some statements of dubious veracity. As just one example, Computer Modern is by no means the only font available in a LaTeX system and it is trivial to change fonts.

      Like some of the other commentators, you seem to be starting from the assumption that I can’t possibly know what I’m talking about. But give me a little credit here. Trivial to change fonts in LaTeX? Oh yes, provided that you’re changing to a font for which you have a TeX font metric file and a virtual font file and an encoding file and a mapping file. There are a handful of fonts for which people have already created these files, but creating them is the opposite of trivial. They are needed because TeX is – at the end of the day – a very old piece of software that predates numerous standards.

      To work with fonts in a reasonable way, one needs to use not LaTeX but XeLaTeX. Perhaps you already are, and that’s what you mean by ‘LaTeX’. But see my next point.

      Also, your complaints about installing TeX systems do not in any way shape or form agree with my experience, or that of the many students (from those who know the innards of a Unix system to those who know how to switch it on and use Word) I have taught, all of whom have installed TeX systems (ProTeX or MacTeX) on their computers without a single hitch; it’s been seamless every time and that’s been the case for many years.

      As I explain above, the problem with packages such as proTeXt (which I didn’t mention) and MacTeX (which I did) is not that they are hard to install but that they are bloatware. I’ll quote myself:

      And anything other than vanilla TeX and LaTeX is really, really difficult to install. So difficult that most people seem to give up on installing individual packages and instead install the whole of something called TeX Live. Tex Live wraps up almost everything that a TeX or LaTeX user could ever possibly want into a single, handy download. A single, handy unbelievably large download that takes up over two gigabytes on disk: to be exact, 2.4 gigabytes for the Mac version, MacTeX. For comparison, LibreOffice takes up about one and a half gigabytes of disk space on a Windows or Linux machine, and less than one gigabyte on a Mac. And LibreOffice is a word processor, a spreadsheet, a slideshow presentation program, a drawing package, and a database, all in one. LaTeX is just a typesetting program. This is nuts – especially if you’re using old or cheap hardware.

      As I said, I didn’t mention proTeXt, but it’s a download well over 2gb in size. Not sure what that expands to once unzipped, but the problem appears to be the same, and exactly as I describe in my essay. In order to use anything other than vanilla TeX/LaTeX (for example, XeLaTeX) without driving oneself insane, one has to install a vast quantity of software that one neither wants nor needs. The fact that anybody even considers doing so shows how bad the underlying problem really is.

      That said, I think we’re almost in agreement about the main point:

      Both TeX and LaTeX were originally designed for writing scientific documents, i.e. documents containing lots of mathematics, symbols, chemical formulae etc. If your writing does not require any of these, then it is arguably quicker to use something like Word (a piece of software that has improved dramatically since TeX and LaTeX were created).

      As I emphasise in my essay, technology has indeed moved on since TeX and LaTeX were created. There are, for example, packages that will enable you to write your prose in something better suited to the writing of prose, and drop into LaTeX syntax only for the equations etc. Moreover, thanks to MathJax, you can export the results in a more Internet-friendly form than the PS or PDF files that LaTeX and co create, namely HTML.

      I’ll refer you to Elizabeth Henning’s comments above (see in particular the first one).

      1. Here you really don’t seem to know what you are taking about. XeLaTeX is not not LaTeX. XeLaTeX as well as Lualatex are alternatives not to LaTeX but to Pdftex. Also the argument with the old hardware seems very tendentious (to say the least) as hard drive space is really not an issue anymore and the space difference between TeXlive (or MacTeX) and the Office Suite of Microsoft are only marginal. Also note that the functionality of TeXlive is in not necessarily smaller than that of the mentioned Office Suites (you can create beamer slides, draw flowcharts, create tables and so on) – so why not call the later software “bloatware”?

        1. You clearly haven’t bothered to read my article. Where do I say that XeLaTeX is LaTeX? Nowhere. Where do I say that XeLaTeX or LuaLaTeX is an alternative to LaTeX? Nowhere. What I do say is that XeLaTeX is the only reasonable option when it comes to installing new fonts, which is true. LaTeX – actual LaTeX, not XeLaTeX – makes that ridiculously inconvenient because it’s built on top of TeX, which was created for the sole purpose of re-typesetting a book whose font had already been chosen.

          Oh, and ‘the functionality of TeXlive is in not necessarily smaller than that of the mentioned Office Suites’ because you ‘can create beamer slides, draw flowcharts, create tables and so on’ – really? MSOffice is more than just Word and PowerPoint, and LibreOffice is more than just Writer and Impress. They also both contain spreadsheet programs, etc. You ever tried doing your taxes in LaTeX?

  29. I agree with most of this critique!

    I write. I don’t put formulas in documents. Latex makes little sense for me.
    Latex works for formatting documents, but there are better tools for my purpose.
    I’m not wild about word processors either, but I’m starting to appreciate their utility.
    I like to write in vim most of all. Nothing compares to vim for editing and manipulating text. For this reason, I was seduced by Latex, but Latex is a great deal of work for very little benefit in my case.

    One problem I have with vim is curly quotation marks. Vim doesn’t do them well. There are plugins and workarounds, but I’ve chosen just to write with dumb quotes. Latex will fix the curly quotes problem, but only if I remember to type “ and ” instead of “. It’s easier to run a file through groff than Latex if all I want to do is fix quotation marks. Easiest of all is a word processor. If I import a file to Libreoffice with dumb quotes, select all the text, and apply autocorrect, it fixes all the quotes for me. It’s like magic, it’s easy, and I can just write in vim without all the hassle of markup. Curiously, this only seems to work if default is selected as a style. But it works!

    Latex is oversold and unnecessary for many people. Others need what it can do. For my case, groff is a much easier markup language to deal with. Basically, I have to do nothing or very little to a straight text file before running it through groff. If I have to put a wrapper template around a file before calling groff, I can get away with a very minimal file. Almost nothing is needed. As I see it, the need for Latex for most people begins and ends with complex equations and/or very specific formatting requirements. Anything short of that, and there are easier and faster ways of doing things.

    1. Thanks, Robert – good points. And I haven’t tried groff but I’ll give it a look!

      As I see it, the need for Latex for most people begins and ends with complex equations and/or very specific formatting requirements. Anything short of that, and there are easier and faster ways of doing things.

      Basically, yes. And even with the equations: it’s now possible to write those in LaTeX and embed them in a document the rest of which is not written in LaTeX.

  30. Pingback: trollheaven
  31. Hi Daniel, an interesting article and debate in the comments. I wanted to give a bit of an ‘outsider’ perspective as someone who’s worked in academic publishing and wider communication efforts.

    Most of the time, the first thing I have to do when I get a document is to clean up the styling. Often that means printing out a copy for reference, then removing everything and starting from scratch with plain text. This is a Word to InDesign workflow. You can map styles in Word to styles in InDesign, of course, but many of the Word docs I’ve worked with have been very messy and inconsistent in their use of styles, including such wonders as the footnote style being applied to paragraph text and adjusted to look the same.

    As you say, when you’re preparing your own book, presentation or paper, it’s up to the author. But for edited work, standardization and productivity are most important, and it’s usually quick to batch content in-house using a standard system than it is to check if people are applying style guides correctly.

    I’ve always advised editors to demand 12pt, 1.5 line space Word documents because it’s easiest to proofread. That’s a big flaw in writers typesetting their work as they write, it often makes the editor’s job harder because errors are harder to spot.

    Finally, if I can be a bit cheeky, most researchers would be better advised to invest more time in learning how to write clearly than in becoming typesetters. Beautiful typography can not make plodding or wooly text any more interesting.

    1. I totally agree, especially with this: ‘most researchers would be better advised to invest more time in learning how to write clearly than in becoming typesetters’.

  32. I’m not going to try to “correct” your opinion because I agree that composing in LaTeX is not for everyone. All I can tell you is why it is more efficient for me to do so than using a WYSIWYG editor. It isnt JUST about making it look better.

    So here are the reasons I prefer LaTeX, just to give another viewpoint on why some people might find it useful and to help decide whether it is worth the learning curve.

    – I use both chemical and mathematical equations and formulas in my writing. For me, writing chemistry/math expressions and equations are simply much faster (and also have the advantage of lookig better) in LaTeX than in a word processor, particularly for chemical equations and formulas. In a word processor when I come to a formula/equation, I have to spend too much time thinking about how to create what I want. It takes me out of my writing groove.

    – I also do presentations in LaTeX (using the Beamer package). Using LaTeX, it is MUCH easier to have different versions of presentations depending on what I want to do. For example, I use the same file for the handouts I give to students, the presentation I give in class, and the annotated version I post for them to study from. To produce one output or the other I just change the “mode” before I compile the presentation. No other presentation package has anything like this.

    – Text editors are simply far better than word processors for actually editing things: surgically altering text as quickly as possible. Watch someone who is adept at Vim, Emacs or even editors like Sublime Text and you’ll see what I mean.

    – As you mention, the output is indeed much better with LaTeX than with word processors. I also find word processors not so good at things like cross-referencing, optimal figure/table placement, or working with large documents.

    1. Hello Chris! Thanks for your perspective.

      The point about equations has been dealt with above a few times; my answer is that you can write your equations in LaTeX and embed them in a Mardown document, and the end result looks exactly the same but you don’t have to use LaTeX for the words surrounding the equations. The point about Beamer slides and handouts is valid although again I would observe that you can write them in Markdown and convert them to LaTeX. (Personally I’m not overly keen on Beamer slide presentations as they all tend to look the same. With manuscript drafts, sameness is a good thing, but a slide presentation is more of an end product.) Actually, are you sure that no other presentation package enables you to create slides and handouts from a single file? As far as I know, most WYSIWYG presentation packages also do that. But maybe I’m not understanding you fully and there’s some specific feature of Beamer handouts that they don’t support?

      When it comes to editing a document, there are two tasks involved. One is the technical task of making a specific alteration to a text. And as you correctly point out, text editors such as Vim and Emacs are very powerful when it comes to this sort of thing, and that is indeed something that I didn’t mention above. However, you don’t need to write in LaTeX to use Vim or Emacs. You can use them when writing in Markdown or org-mode and then convert the Markdown or org-mode source to HTML or LaTeX (via Pandoc or Emacs itself) – a point I mention in the article and that has come up a few times in the comments above.

      This brings me to the second task involved in editing a text, which is reading it in order to decide what alterations to make. Markdown is easier to read than LaTeX source code because it’s a more minimal markup language. And text in the editing window of a word processor is easier to read than Markdown because there’s no markup at all. The PDF output of LaTeX (or Markdown via LaTeX) is arguably easier to read than text in the editing window of a word processor because it tends to be typeset more attractively. But you can’t apply your edits to that directly – instead, you have to apply your edits to something else, i.e. the LaTeX or Markdown source, and then recompile.

      As for the other things you mention: cross-referencing works fine in Word using bookmarks, but you’re right that it often struggles a bit with long documents (e.g. there’s an 800-page document I’ve been working on with some colleagues and it takes a while to load into memory). And when it comes to what you call ‘optimal figure/table placement’, yes, word processors are awful. But here’s the thing: figure placement is not a matter of writing but a matter of publishing. As I’ve said over and over again in my replies to comments, my argument in the essay above (which is even contained in the title of the essay) is that LaTeX is not for writing (i.e. deciding which words to use in which order) but for typesetting (i.e. making documents look nice after you’ve written them).

      Anyway, it’s great receiving comments like yours because they let me know what I ought to clarify if I was writing this essay again (which one day maybe I will), so thanks again.

      1. Besides the points mentioned by Chris, here is another major difference between LaTeX and word processors like LibreOffice (or Word). There are users who prefer to primarily interact with the computer via the keyboard and there are users who prefer to do so via the mouse. For the first group, LaTeX is a godsend if you have to typeset a equation heavy document or make a presentation. For the latter, something like LibreOffice is probably a better choice.

        You are right, inserting pictures in a LaTeX document is like a game of dice. Even dealing with long tables can be quite tricky. However, despite all its shortcomings I cannot imagine writing a report or a thesis in anything other than LaTeX (or maybe use Pandoc to convert the MarkDown to LaTeX, don’t quite like that, though). I have even taken notes in LaTeX. It is just so convenient. Editing a 800 page document with LaTeX and Vim would be a breeze compared doing the same in a word processor.

        Maybe all you need to turn to the dark side is a good mechanical keyboard. Cheers.

        1. It’s not really about keyboard versus mouse. Power users of MSOffice and LibreOffice also control them with keystrokes rather than mouse clicks. This is exactly the sort of thing that committed LaTeX users seem not to be aware of.

          Just as importantly: what when you want to convert your 800-page document to some other format than PDF? For example, HTML. Yes, there’s a Perl script for turning LaTeX into HTML, but it won’t work properly if you did anything complicated… and if you didn’t do anything complicated, then why didn’t you just use Markdown?

          I find this over and over again when presented with arguments for why LaTeX is actually a great solution for writers. Scratch the surface and none of those arguments holds up. The truth is that people persist in writing in LaTeX for social reasons, because of the force of habit, and through ignorance of the alternatives. If you as a 21st century programmer were to sit down and invent a document preparation system, you’d be very unlikely to invent LaTeX, or anything like it. You’d be much more likely to come up with something purely semantic, like HTML. Why? Because making people mix imperative statements for the construction of a PDF into a text document in order to write anything makes no sense.

          LaTeX is a dinosaur from the world before the web, when programmers hadn’t yet figured out that presentation and semantics ought to be separated from each other.

  33. If you’re looking at LaTeX in isolation, you’ve really missed 99% of it’s power. Remember you don’t write *in* LaTeX. It is part of an ecosystem consisting of multiple programs. On my computer, I write exclusively in emacs. Why? It’s superfast– I can write and edit far more quickly than I ever could in a WYSIWYG editor. Just moving around a screen is incredibly fast. LaTeX is the perfect complement to that.

    The setup really shines when you are managing or editing large volumes, such as folders with hundreds of files, a thesis or a book. Let’s say I’m writing a 500-page book about the character “Rachel.” During the editing process, we decided to change her name to “Lisa.” If all my files are LaTex, I just use a quick bash one-liner to convert all instances of Rachel to Lisa. Poof!

    I am writing my second book, and I can’t tell you how many times I’ve used the command line to grep for some little detail in vast filesystems with thousands of notes and documents– and pulled up the needle in the haystack because of LaTeX. When collaborating, you can use git for version control, and your team can just make and commit changes as a group. It’s fast, flexible and easy; nobody has to email attachments around or wonder who changed what.

    Not to mention headers. I am so happy that I never have to fight with Word header numbering again. They are ALWAYS RIGHT with LaTeX. You never have to struggle and wonder why the numbers suddenly start over at 1, or silly things like that. It is customizable and there are lots of free, community-built packages.

    This is not to say that LaTeX is perfect. Nothing is. And if you are just using LaTeX in some dumb text editor, without any experience with the command line, I could see how it would seem silly. But LaTeX is part of something much bigger. Before you wave it aside, you should know that LaTeX is just one tool in a powerful Swiss army knife designed for file manipulation. If you take the time to learn to use it in this context, it is massively faster and more powerful than a standalone editor.

    1. Although these are good points, not a single one of them is specific to LaTeX. You could keep every single one of those benefits – editing with Emacs, using command line tools, putting your text under version control with git – while working in Markdown, say – or indeed in any other so-called ‘plain text’ format (e.g. Org-mode).

  34. Dear Dr Allington,

    I was delighted to read your blog post “The Latex Fetish”, because I do so agree with it. In the 1990s I worked for a company that used Latex for its documents, and came to intensely dislike manually preparing and editing them. I do not want to repeat your words, but would like to take up some particular points in more detail, as I have direct experience of these.

    My employer argued that MS Word evolves, and becomes incompatible with old documents. True, but our Latex documents used some in-house macros which evolved without proper configuration control, and it became impossible to reprint old documents. Furthermore, Libre/Open Office today can read Word6 documents (early 1990s) and (I believe) Word2 documents (1980s). Even MS Word 2013, which I use, can read Word98 documents.

    I defy any reasonable person to prepare with Latex a set of address labels, 3 x 7 labels on each sheet of A4 (Avery 2160 format), using a membership list from a database or spreadsheet. The labels must take account of the fact that some British addresses are longer than others — they are not standardised as US addresses are.

    I find MS OneNote an ideal product for preparing new documents when I am not sure what their final shape will be. Its structure of sections, and pages within sections, allows snippets to be collected that can be organised afterwards, and does allow the author to “just write”. Then the material can be exported to Word for final layout. Or, as with these comments, pasted into a blog.

    1. All good points here. LaTeX is less backward-compatible than Word, it’s very poorly suited to typesetting anything other than a research paper or an academic slide presentation, and it locks you into certain ways of structuring a document from the very beginning.

  35. A well-written post, and well-deserving the long discussion it sparked. However, seeing the attention it got, I’d like to point out two things:

    >Free and open source software has a strong tendency towards being difficult to install and get up and running.

    Using Sharelatex or Overleaf websites you don’t need to install anything, just like with the Google Docs. You can set up an account and start creating your documents using the full extent of LaTeX capabilities, and collaborate with other people on them.

    >To cut to the chase, LaTeX documents are very hard to read until typeset, which is inefficient for both writing and editing.

    Once again – in Sharelatex and Overleaf you have a preview of your document on the right side of the browser. While it’s not as fast and convenient as seeing your document in real time, it largely alleviates the issue. You can look at the formatted document and edit the source code simultaneously.

    Despite ommiting this websites and their advantages, your critique is fair, and I agree that pressure to use LaTeX can be too high. Personally, I’m a programmist and working with code feels intuitive and natural to me. Therefore, LaTeX truly let’s me focus on the content, as I know that my text will look good and I’ll only need to fix minor things at the end. For a similar effort I get a worse-looking document from a text-processor(you provide an example in your text). However, that’s a personal opinion, and I can understand how someone with lesser affinity and/or love for working with pure text might have very different experience.

    All best,
    Grzegorz

    1. Thanks, Grzegorz. Those are good points. But here’s a further point I’d like you to consider: LaTeX is not ‘pure text’. It’s text mixed up with an imperative programming language used for controlling the appearance of the text once printed. If you like working with pure text, why are you using LaTeX and not – say – Markdown? Which isn’t pure text either – but is much closer to being pure text than LaTeX is.

      (I should perhaps also note that the approach to LaTeX editing that your recommend is by your own admission ‘not as fast and convenient as seeing your document in real time’. How about actually seeing your document in real time? At some point people are going to have to stop putting sticking plasters on LaTeX and start afresh with something else that doesn’t have as many problems to begin with. Something designed with the benefit of the experience of the decades that have elapsed since TeX and LaTeX were invented.)

Comments are closed.