I like the idea of the paper that comparing usability and efficiency of document generation software is important for many aspects: especially efficiency, correctnes and enjoyment in writing scientific texts.
However, I do not agree with the conclusion and the proposed changes for manuscript acceptance.

If the conclusions would have been "Reproducing a page of an existing document favors a WYSIWYG program in favor of a markup language with aribtrary choice of non-WYSIWYG editor" I would agree because the data supports this conclusion. But the authors claim a much stronger point.

I would be very interested in what the authors think about the following points

Especially:

"For all other types of documents, our results suggest that LaTeX reduces the user's productivity and results in more orthographical, grammatical, and formatting errors, more typos, and less written text than Microsoft Word over the same duration of time."

I cannot find how to extrapolate from the experiment to real world multi page documents. The first problem is I do not see the concrete defition of formatting, typographic and grammatical errors. Neither is analyzed whether the users had type checking enabled (which is not the default in every latex editor) or not. Also the point that LaTeX users make more typos or grammatical errors is quite strong if you consider that, in principle, you could write the text in Word and then copy and paste the text into LaTeX without having to change anything. So I would question if there are other reasons for this discrepancy in typos and so on.

"it remains an open question to determine the amount of taxpayer money that is spent worldwide for researchers to use LaTeX over a more efficient document preparation system, which would free up their time to advance their respective field."

This is rather misleading because it implies that LaTeX is not the most efficient tool. I would really like some cited sources on this claim, even if the authors were quite careful not to formulate it as a claim.

It would be interesting to see the difference attributable to latex editors: emacs and vim are vastly different then kile or texmaker and both a different to lynx. "Another characteristic of our study is that it is practically impossible to evaluate LaTeX without also evaluating the used editors. In fact, our research measured the efficiency of Word against LaTeX in combination with som" Yet i cannot find where this is underlined by data.

Furthermore, if it is practically impossible to evaluate LaTeX without the used editors, then how they support their conclusions?

A point that is very important in real life manuscript preparation that is however not addressed is collaboration and version control. From personal experience the managing of plain text documents paired with version control software is vastly superior to most other methods. Whereas Word's tracking changes future shines if the number of persons editing the same document is small

I am particularly disappointed by the conclusions:

"We therefore suggest that leading scientific journals should consider accepting submissions in LaTeX only if this is justified by the level of mathematics presented in the paper. In all other cases, we think that scholarly journals should request authors to submit their documents in Word or PDF format."

As in: the pdf produced with Latex? I do not think this makes much sense at all.

"First, we think that the appearance of the text is secondary to the scientific merit of an article and its impact to the field. And, second, preventing researchers from producing documents in LaTeX would save time and money to maximize the benefit of research and development for both the research team and the public."

Okay, You can think that. But making this kind of claim without source is questionable at best. It's also quite easy to produce a counter example of atrociously typeset text that is clearly detrimental to understanding. Take your favorite article, use comic sans in size 6 and make everything yellow. Format it for A0 paper and try to not be annoyed.

Secondly, LaTeX is excellent for disentangling presentation from content. Most publishers provide LaTeX templates which reduces the time needed for formatting on your own dramatically.

The last point is not supported by the data nor does it cite a source. It also does not discuss helpful remedies, e.g. using Markdown and converting the Markdown to Word/Latex which let's you concentrate solely on manuscript production.

Also, it is a bit confusing that the text does not distinguish between latex, the programming language or its implementation lualatex,pdflatex and so on. But that's a minor point.

Furthermore, the advantages of being able to define semantic markup in LaTeX is not mentioned at all, things like \vector{a} or \mass and so on. The very useful libraries of LaTeX are not mentioned e.g. SI units , tikz.

Writing of reproducible documents with knitr etc. is also ignored.

The quality of the type setting is, as far as I can see, ignored.

Merits of open source against closed source in conclusions are also not mentioned.


In conclusion I think the article's idea is good, the experiments interesting but the conclusion and presentation heavily biased. The authors did a very good work with the experiment and I think if the authors address the points above the article could be very useful in e.g. discussing what document generation software should be taught to students.