Hacker News new | past | comments | ask | show | jobs | submit login
The Case for Writing Papers in Economics Using FaKe LaTeX [pdf] (illinois.edu)
36 points by PascLeRasc on Mar 15, 2018 | hide | past | web | favorite | 59 comments

I don't understand why LaTeX is considered hard. I wish people would try TeX out first as well. You literally type stuff and a paragraph comes out. Enter a blank line to separate paragraphs. It can't be much easier. LaTeX just adds some macros to make it easy to make a title and section headings etc. Once you become proficient at entering equations in TeX it's incredibly fast. And you won't even need to render them to read them after a while.

> One can achieve something like 95% of the visual appearance of native LaTeX documents with a one-time investment of an hour or less work

Yes, and one can use LaTeX and achieve 100% with a similar one-time investment. I really don't understand this. To produce the document presented here LaTeX has absolutely no extra overhead compared with any other tool.

I suspect the misconception comes from the complicated looking tools like TeXnicCenter etc. If people saw me writing LaTeX in emacs (with zero buttons or other complications) they wouldn't think it was complicated. It's just text.

Still, if people insist on using Word for whatever reason I'm glad if they make the documents look nice like this one, even if the page geometry is way off (use Komascript).

> I wish people would try TeX out first as well. You literally type stuff and a paragraph comes out. Enter a blank line to separate paragraphs. It can't be much easier.

This may be because raw TeX is not very google-able, whereas large amounts of guides exist for LaTeX in particular. The best resource for it seems to be the TeXbook, which seems to still be in print and not freely available.

Side note: This is one of the reasons why I'm personally somewhat partial to troff, be it heirloom-doctools or groff. Many great resources, introductionary or otherwise, are freely available on the Internet at no cost, such as UNIX Text Processing and the introductions to the ms macros, eqn and tbl in the UNIX V7 manual volume 2A.

troff/groff seems attractive to me, but can it achieve the same level of quality is TeX? When it comes to typesetting I'm not prepared to sacrifice quality for convenience and TeX seems to offer a reasonable compromise here.

Straight up: No, it can't.

It will blow MS Word and LibreOffice out of the water. It can get close to TeX, but very noticeably eqn(1) is inferior to math mode. Table of contents generation and putting it at the top of the document is still not happening.

This is covered on page 4, which cites a study showing a loss of productivity on papers that don't have a lot of equations.

”I wish people would try TeX out first as well.”

Do you by any chance have any TeX tutorials to recommend?

My favourite TeX tutorial, which will turn you from beginner to expert and make you fall in love with TeX (maybe), is the book A Beginner's Book of TeX by Seroul and Levy. It comes highly recommended by Hans Hagen in the ConTeXt manual. (It's not free—published by Springer—but you can find it in the usual places you find books: libraries, whatever.)

The TeXbook by Knuth. See if you can find it in a library.

Lots of comments pointing out how obvious it is that the paper was not typeset with real LaTeX, but I think that's a little beside the point. Scott isn't arguing that it looks 100% as good as LaTeX. The article cites research that suggests that writing in LaTeX hurts productivity, even for expert users:

We show that LaTeX users were slower than Word users, wrote less text in the same amount of time, and produced more typesetting, orthographical, grammatical, and formatting errors. On most measures, expert LaTeX users performed even worse than novice Word users

I think that paper came up on HN before and IIRC the tasks they were measuring were things like manually positioning some figure in an arbitrary position. This is not a use case LaTeX optimizes for or one that comes up often in my experience.

Edit: See https://news.ycombinator.com/item?id=8797002 for an earlier discussion of the referenced paper.

Not to mention that this is geared toward people who write papers with a minimum of equations. Unless a paper is math-heavy, it probably doesn't need the typesetting that TeX offers.

LaTeX isn't that hard, Microsoft equation editor is slow, and don't even get me started on how much more time consuming it is to put in references, tables of content, etc.

The first equation in the paper is

$$\Delta P_t = \mu + \sum_{j=1}^J \lambda^j S^j_t + \sum_{p=1}^P \delta_p \Delta P_{t-p} + \varepsilon_t$$

Yes, it's a bit of a mouthful but clicking all of those symbols in equation editor is much slower once you know the syntax.

Not only is the MS eqn editor slow, it also doesn't work consistently across versions of Word. I have a work colleague whose equations never show properly on my machine when he sends over Word docs.

By contrast, LaTeX works the same everywhere -- to the extent that SymPy can derive an equation, output it as LaTeX, and it goes right into a paper or presentation without issues.

Word can automatically manage your references and TOCs. There's even a Mendeley plugin to get your references from Mendeley into Word. It's actually as simple as a couple of shortcuts (Alt - S - Y1 on my machine, to "Add citation" from Mendeley).

Also, like another poster has said, Word now recognises some LateX-like shorthand in equations so you don't have to click everywhere. Frex, adding a subscript is as simple as adding an underscore, the subscript and pressing space.

I don't know how to quickly add superscripts, or both sub- and super-scripts though :)

Not advocating for Word vs LaTeX, mind, but Word's actually pretty handy for writing a quick paper, like for a Master's thesis or putting together an early paper draft to get an idea of the outline etc. things that don't have stringent formatting requirements.

Pseudo-LaTeX syntax works in Word's new equation editor (introduced in Office 2007). I can't test, but I suspect your equation works as-is in Word.

I can test for the class. It's close, the curly brackets aren't used for grouping in MathType:


Fixing that for the summation terms and t-p subscript produces:

    \Delta P_t = \mu + \sum_(j=1)^J \lambda^j S^j_t + \sum_(p=1)^P \delta_p \Delta P_(t-p) + \varepsilon_t

where Word doesn't realize that the terms after the sum should be the content of the sum. Word Help and the equation editor are thoroughly unhelpful in deducing the appropriate syntax of \sum.

What I eventually divined was that the syntax is state-based: If you type it out manually, it will put your cursor into the right position, and once you're done with _t you need to hit the right arrow key before typing the space and plus symbol. Copy-pasting the whole thing is different and does not work.

I mean, it's great that it supports an entry mode that's faster than digging through menus and clicking on symbols with your mouse. But it's still fundamentally different and, dare I say, inferior when the appearance of the equation isn't defined by the sequence of characters you enter (or copy-paste) but by the sequence of arrow keys and escape keys you hit while you type it in. This is exactly the problem with Office vs. LaTeX.

If you type a space after P_t it will figure out it's a subscript. That is, after you copy-paste your text you can hit space a few times to wake it up to the fact that it needs to render it.

In my experience, pseudo-LaTeX syntax pseudo-works.

Page 4 specifically refers to papers that don't have a lot of equations.

In my experience the abstract statement "The downside is that it is a very time intensive and complicated method of writing papers. " is just incorrect.

There is a bit of a learning curve, but once you are past that it is much less time intensive that writing papers in a word processor. This effect is magnified with longer documents.

That may be somewhat domain specific. And to be open about my biases, I also think that Word's "track changes" is a pretty poor collaboration tool.


I might be going to downvote hell for this, but as a computer scientist writing papers for a living, every hour spent writing latex belies your comment.

By the way, Context is a much saner alternative, although it still shares many of Latex' core flaws.

Unfortunately, that doesn't matter because most conference mandate use of their Latex or (ugly) Word template, and using Word is cause for ostracization.

You are also correct to say that version control is a big advantage of Latex. I wonder why Microsoft hasn't come up with a better competing solution.

How many papers have you written? Have you written a master's or PhD thesis? How exactly are you spending your hours writing LaTeX? It sounds like you are doing it wrong.

You're two links away from the answer, person that obviously wants to have a good faith conversation.

It took me a while to figure out what you meant. I've never looked at people's profiles on here before. I couldn't imagine writing my thesis in anything but LaTeX. Or rather, anything that didn't support both high-quality typesetting and programming. I had little programs written in my .tex files which would generate some of the tables and figures. TikZ is wonderful. I aimed for nothing less than TAoCP level of quality.

The output of Tex is superior to anything else. I doesn't change the fact that the user experience is awful.

But if you want Tex quality with less head-bashing, Context is much better than Latex. It's what I use for my thesis, but not for my papers -- templates are all Latex, and I didn't know about it at the time.

Also I disagree with your basic point: hours of my time are worth much more than some fancy ligatures and fluff that makes the text look very marginally better.

Writing papers is always real work, but I’ve never wasted more time than when collaborating with a word processing document...

They cite a study showing that a novice in word is faster than an experienced user of LaTeX.

Why not using LyX? It writes LaTeX for you with minimal effort and knowledge.

This article screams non-LaTeX in the title already. The word LaTeX should be rendered as http://site.uit.no/futurelab/files/2013/10/latex-logo.png


I've found that writing in LyX is faster than writing in Microsoft Word and friends. It forces you to focus on content (hence the WYSIWYM) and greatly accelerates equation entry with keyboard shortcuts and live preview.

I think the point I would make in favour of actually using LateX is that whilst evidence may indicate Word is quicker, the time I spend in Latex is (as a programmer) better spent.

I mean that in the sense that because the interface is basically a compiler compiling my code I can logically reason about what it's doing, and have reproducible results. I can tell it where I can tell it how to take care of images and be confident it will follow my guidance. This is especially important in larger documents. In Word, editted page 3 is as likely to change the layout of page 72 as an edit on page 72 is, and an edit on page 3 will have an equal chance of effecting the layout of page 4-N.

In Latex with very few exceptions I do not manually layout any part of the document, I design the rules for how the document is laid out. Obviously that is more time consuming - as solving the general problem is always more painful than a point fix.

This really pays off when your dissertation is due in 2 days and you're trying to cut 150 pages down to 100. You can focus on content, not why Bill Gates personally believes that any image is a graven image and is therefore forbidden.

"I conclude that appearance/aesthetics is the dominant reason for widespread adoption of LaTeX". I fully disagree. IMHO, the dominant reason is that Microsoft word has been a nighmare for scientific documents. The different generations of equation editors were buggy and incompatible. You can not edit an old document using a recent version of word. You can not exchange documents with people using a different version of word. For most people, when you can not use word, the only choice for scientific documents is LaTeX.

Even if recent versions of word are less buggy, people will migrate slowly.

Link seems broken atm :(

Regarding the efficiency though, I would think (even disregarding new users of LaTeX) it would really depend on the length and content of the paper.

For longer papers with many formulas I'd pick straight LaTeX any time. Beside the formulas and look, this way once something looks good, it probably won't break in many unforeseen ugly ways when stuff is changed or added elsewhere. This in my experience is not the case with word processors, where you really need to review the whole document after any edit. Plus, working with text in a text editor is so much more efficient and enjoyable.

For things like slides and notes though, particularly when there are a lot of screenshots/graphics/tables/etc. in proportion to text, it's somewhere about 50/50 between Word and org-mode, which I then compile into pdf through either LaTeX or html. But I have accumulated a lot of templates and customization for the latter option to work well; without them, I would just be using Word.

Another factor is whether, and how often, it is going to be necessary to update the document with changed figures/tables/etc. This depends on how they are generated: if they come from an Excel spreadsheet, things work much smoother embedding them in Word; if they are generated by some code running, it's much easier to have the LaTeX document pick them up.

If the person already knows markdown, Pandoc is a great option [0]. I find it really useful for creating pdfs. I haven't used it to create official documents that often, but I have used it to create Beamer slides and it was a very smooth experience. If you're an advanced user and you want to do something LaTeX-specific you can just write LaTeX inline and Pandoc will accept it.

[0] http://pandoc.org/

The referenced article concerning the speed of Latex vs Word can be read freely online [0].

I remember reading that article shortly after it was published. They conducted an experiment in which participants were instructed to typeset a piece of work from scratch in thirty minutes in either Latex or Word. Since Latex is most valuable and efficient when writing a large document such as a book, containing multiple chapters, citations, cross-references and potentially images, the study is almost meaningless. The only conclusion I would draw from it is "if you want to type something small up real quick, maybe consider something other than Latex".

[0] http://journals.plos.org/plosone/article?id=10.1371/journal....

If this PDF is supposed to convince me that this 'fake' LaTeX is somehow able to produce results which come close to the real thing it fails rather miserably. Yes, it is set in some variation of Computer Modern [1] and the basic layout sorta-kinda looks like that produced by commonly used LaTeX templates. That said, the PDF distinctly looks like something produced by MS Word due to its rather characteristic (problems with) spacing and kerning. It also does not look the results produced by the AEA template [2] but that might be by choice.

Writing the same document using something like LyX is probably easier than using "FaKe LaTeX" with the added advantage of using LaTeX templates to provide a uniform look. The only real advantage of this "FaKe LaTeX" is the possibility to directly use a spread sheet program for tabular data.

[1] https://en.wikipedia.org/wiki/Computer_Modern

[2] https://www.aeaweb.org/journals/policies/templates

Tangentially related: I recently learned that Satoshi's original paper on bitcoin was not actually written with LaTeX. It seemed to have tricked a lot of people (including myself).


This is rendering very badly for me in Firefox. The font is very hard to read - some parts of it are so thin, they appear invisible. (https://screenshots.firefox.com/LHvHiygRGnaPkhsy/www.farmdoc...)

It looks mostly fine in a desktop viewer though.

Of course, the point is moot if the journal you're submitting to only takes LaTeX.

Sadly, most journals that I might submit to require Word. I published a paper in transportation a while back, and happily submitted my manuscript for review as an immaculately formatted PDF.

It was reviewed and accepted, which meant that they sent me the final, accepted paper submission guidelines that stipulated... Word.

Using LaTeX is like writing a program to automate something you do a lot:[1] lots of effort up front for payback down the line. Once you work out how to use it and iron out the kinks in your templates, it becomes much easier, permanently. Whereas with Word, while you'll get started much quicker, every sufficiently complex document will break in new and strange ways that require extreme dexterity with a mouse and a good deal of random fiddling to fix.

[1] https://xkcd.com/1319/ (ignore the bottom graph)

But is this paper written in faKe LaTeX?

I assume it is, because the author writes "As a non-LaTeX user", but I'm having trouble verifying this by eye otherwise. Which I suppose is the point.

Oh yes, it is. My teeth hurt while reading it. There is much, much more to TeX typesetting than just the Computer Modern fonts (which many find unattractive, by the way).

One of the benefits of LaTeX is supposed to be easy switching between fonts.

I don't think I've ever seen something in LateX that didn't use Computer Modern!

Maybe you just didn't know it was LaTeX if it wasn't Computer Modern? I typeset my thesis using Minion and Optima.

Yes, there are two big giveaways in the article. The first is how the word LaTeX is typeset. This has a command that produces a much fancier typesetting in most LaTeX documents. The second big giveaway is a bit deeper into the paper

> It turns out there is an alternative that satisfies my economist mindset—FaKe LaTeX documents created in Microsoft Word. That is exactly how I produced this note.

There's also the awful justification throughout.

Well he did cite a study showing that people were more productive using Word than LaTeX. That's a very good justification assuming it's factually accurate and the productivity gap is meaningful.

Most of his complaints about how hard it is to use LaTeX seem absurd, but that doesn't mean he isn't experiencing them or that they aren't common among less technical users.

Sorry, I meant justification as in the typesetting sense:


This is something TeX is _very_ good at. Word, less so.

Ah, I think he mentioned that as one of the things Word can't mimic.

Would LaTex hyphenate “Mi- crosoft”. It struck me as particularly ugly.

There was a time in my life when I would have noticed that, but I haven't actually written LaTeX in years.

I can tell instantly from the poor spacing of the justified text. Word makes obviously poor choices that TeX would not. Word could use the same algorithm used by TeX. I have read that the codebase for Word includes a significant amount of legacy code attempting to maintain backward compatibility, so they are probably stuck with the layout code currently used.

On that note, I wasn't aware you could get Computer Modern as a font in Word (having tried once!).

If we care about substance over form, papers should be written in some type of markdown -- leave publishers, or readers even, to impose a format.

The substance often is the equations, not the words. I don't really want to write equations in markdown, and I don't want some publisher trying to impose a format on my equation. I want to format it, so that my readers read my equation, not what some publisher thought it said.

Then presumably, you are either comfortable with LaTeX or else are capable of asking your publisher to respect your wishes. For writers who are not familiar with publishing tools, I would humbly submit that a markdown language that formats deterministically for publication would be superior to wrangling with MS-Word.

Well, I can get LaTeX (or more precisely, LyX) to do what I want, so that on the PDF, the equation is exactly what I want it to be. I don't know if I'm "comfortable" with it (though Lyx is far easier to use than LaTeX). But I think I have no choice. [Edit: I have no choice because the publisher doesn't understand my paper, and doesn't understand the equations in it. The publisher may make a change without even realizing it, and the change may really matter. It's my paper, and it's my responsibility that the equations are right.]

Second best would be "asking your publisher to respect your wishes" in the form of looking over galley proofs, to make sure that the equations came out right.

I can't tell. Is this genuine, or is it some kind of Poe's Law (https://en.wikipedia.org/wiki/Poe%27s_law) thing?

People say pretty silly things about TeX and LaTeX, but this is so obviously silly that I thought of it as being like an April Fool joke. On the other hand, the sprinkling in of admiration for DEK, for instance, doesn't mesh with that.

The OP is definitely serious. Some people just cannot stand TeX.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact