News for hackers. Stuff that matters.
Re:An Efficiency Comparison of Document Preparation Systems ... by dang on Wednesday January 14, 2015 @11:51AM (#8887384)
Re:An Efficiency Comparison of Document Preparation Systems ... by ajarmst on Wednesday January 14, 2015 @09:43AM (#8886483)
I only read the abstract, but it seems to imply that they believed that the end result from using Word and the end result from using LaTeX were the same. That is, they were two techniques to achieve the same goal. But LaTeX users don't choose it because they think it makes them faster typists -- they use it because they feel the end result is of objectively higher quality as a typeset document. This may make the statement "you can do it faster in Word" moot or inaccurate because "it" is something different.
Re:An Efficiency Comparison of Document Preparation Systems ... by voidpointer on Wednesday January 14, 2015 @10:06AM (#8886633)
However, by their measure, the results that were prepared in LaTeX were inferior to the results created in in Word. Even for papers that were heavy on equations. So while the LaTeX users would think they create a better end-product, this isn't actually true under the quality metric used in the study.

I have created academic work both in LaTex and in Word and while I have not measured it in an objective way I also had the feeling that LaTeX could create better looking results once you've got over the first hump.

On the other hand, the integrated nature of a system like Word were spellchecking, annotations etc. are all part of the main tool you work with makes it very effective. It takes pretty long however to develop a workflow that avoids all the pitfalls.

In a sense, LaTex makes you believe that once you got over the initial hump, you have typesetting super powers but you will miss out on some of the integrated UI stuff that may boost productivity. Word on the other hand makes you believe that you already know everything while you still need to keep fine-tuning your workflow until you don't mess up constantly.

Re:An Efficiency Comparison of Document Preparation Systems ... by ObviousScience on Wednesday January 14, 2015 @10:29AM (#8886786)
> However, by their measure, the results that were prepared in LaTeX were inferior to the results created in in Word.

They didn't measure the quality of the typesetting in any fashion, though, which is what the GP's claim was about; they measured the rate of errors typists made under time pressure.

It's also unclear what LaTeX software they used, since the one I use has integrated spellcheck, etc.

Re:An Efficiency Comparison of Document Preparation Systems ... by voidpointer on Wednesday January 14, 2015 @10:58AM (#8887030)
> It's also unclear what LaTeX software they used, since the one I use has integrated spellcheck, etc.

That's a hugely important point and in a way it shows that they are comparing apples and oranges. Word is a complete document editing system (well, more or less) while LaTeX is a type setting systems and you need additional tools to use it. Those tools can range from a simple plain-text editor (say, nano) to totally integrated environments that are similar to word in their feature set.

Re:An Efficiency Comparison of Document Preparation Systems ... by ygra on Wednesday January 14, 2015 @10:30AM (#8886792)
In my experience creating a high-quality end result requires knowing about typography, whether you use LaTeX or Word. Things like improperly typeset ”quotes”, or typesetting functions like sin, cos, etc. as variables in math mode, are mistakes that arise because people don't know what they're doing. And that's true regardless of the tool.
Re:An Efficiency Comparison of Document Preparation Systems ... by carlob on Wednesday January 14, 2015 @10:01AM (#8886602)
Exactly! A fun anecdote on this: four professors from my alma mater wrote a book on computation physics in LaTeX, the deal with the editor was that since they were taking care of typesetting and graphics they would be paid both as authors and as typesetters.

As it turns out since the authors' fee was per-copy, while the typesetters' fee was a one shot deal, they made much more as typesetters (though this situation might get reversed if the book keeps selling for decades…).

Re:An Efficiency Comparison of Document Preparation Systems ... by fluidcruft on Wednesday January 14, 2015 @10:07AM (#8886638)
Journals typically do the final typesetting though, so it seems like a premature and irrelevant optimization. Although, I do see the point for preprint.

My experience with LaTeX is that it enables and encourages my OCD behavior and it is very easy to be sucked into an enormous game of package dependency hell and whack-a-mole to achieve some desired aesthetic result. Whereas with Word I already expect an inferior result so I'm somehow more easily happy getting 95% of the way there and calling it a day. It's not too hard to learn "what not to do" in Word and how to use counters and references. Certainly if you've got the capacity to learn LaTeX, you can learn how to markup Word properly and use styles etc.

Re:An Efficiency Comparison of Document Preparation Systems ... by ajarmst on Wednesday January 14, 2015 @10:31AM (#8886801)
Another important point is that I doubt that the LaTeX users were permitted to use personal libraries of macros and styles. Typing a manuscript in raw LaTeX is simply something most LaTeX users never do---I have a directory of style files containing macros for tasks I frequently do. For example, if I type "\rights" into one of my documents, it inserts a nicely formatted copyright notice, boxed and aligned with the Creative Commons symbol. It would take me a half-hour to recreate that in raw LaTeX. Such libraries of personal macros (or organizational ones - when I wrote my thesis, I used a style file from my University that made formatting overhead essentially nil) are the default behaviour for regular LaTeX users, which is not true of Word users. So, again, I think they're measuring the wrong thing because the LaTeX users are not doing their task the way a typical user would, but the Word users are.
Re:An Efficiency Comparison of Document Preparation Systems ... by ajarmst on Wednesday January 14, 2015 @09:48AM (#8886523)
"For example, LaTeX users in our study attained better performance in the typesetting of mathematical equations, and it is not surprising that LaTeX users are typically in disciplines where mathematical formulas are frequent (e.g., mathematics, engineering, or computer science)." I see. So, when you get a better measurement for Word, it's because it's a better tool. When you get a better measurement for LaTeX, it's because your user population is different. I think we're done here.
Re:An Efficiency Comparison of Document Preparation Systems ... by josinalvo on Wednesday January 14, 2015 @10:25AM (#8886765)
I dont think that text means what you think it means =P

It is in a paragraph saying "different tools for different jobs"

Re:An Efficiency Comparison of Document Preparation Systems ... by unfamiliar on Wednesday January 14, 2015 @10:46AM (#8886924)
I think you are misinterpreting it. They seem to be saying that LaTeX is used in those fields because it is good at those things, not that latex performed better in that area because it's users did it a lot.
Re:An Efficiency Comparison of Document Preparation Systems ... by cbsmith on Wednesday January 14, 2015 @09:56AM (#8886574)
Also, there is a small matter that papers where mathematical equations are more prevalent are likely to have fewer words in them (equations are very terse representations of ideas).
Re:An Efficiency Comparison of Document Preparation Systems ... by robinhoodexe on Wednesday January 14, 2015 @10:05AM (#8886625)
This study has been debunked previous here on HN:

https://news.ycombinator.com/item?id=8807497

https://news.ycombinator.com/item?id=8797002

Most of the study is, to be honest, bullshit.

Re:An Efficiency Comparison of Document Preparation Systems ... by falcolas on Wednesday January 14, 2015 @10:45AM (#8886915)
Hmm. It seems that most of the arguments against the study were for the limited time allowed, and for the act of "reproducing" a target document.

IMHO, those are honest goals for a test. Your target may be your university's (publisher's, peer's) expected format which will cause a drop in your grade (pay, funding) if you miss. I also feel that we are frequently time constrained when trying to get work done, so I don't feel that the criticism against that is legitimate.

Now then, this is certainly not the definitive study to decide which is better, Word or LaTEx, but it's certainly an important study for those who want to be productive and have a particular format for a document that they need to match.

Re:An Efficiency Comparison of Document Preparation Systems ... by duaneb on Wednesday January 14, 2015 @12:12PM (#8887543)
Much of the beauty of latex is the fact that you don't have to 'aim' at a target document, because it's already easy to get a decent looking paper together with about ten minutes of copy/paste formatting. You don't HAVE to worry about it because it's already producing a good document! Granted, you CAN make arbitrary layouts with latex, but the beauty is in the macro system removing the need to think about the layout.
Re:An Efficiency Comparison of Document Preparation Systems ... by cies on Wednesday January 14, 2015 @09:47AM (#8886521)
I agree that I'd be faster getting a small document done in Word than in Latex.

But when the document gets huge: ToC, key phrase index, quotes half-way in the margin, output to several paper sizes, multi-page tables, both footnotes AND endnotes, source code listings, a large number of references to journals... In that case I'm glad I used Latex for the smaller documents, because now I'm acquainted to it in time for the big thing.

Re:An Efficiency Comparison of Document Preparation Systems ... by Osmium on Wednesday January 14, 2015 @10:16AM (#8886708)
I agree with you except for:

> multi-page tables

You won't be able to persuade me that's easier in LaTeX than Word! I say that as an otherwise very proficient LaTeX user who finds himself much more productive with LaTeX than Word.

As for this study, I imagine it's a question of proficiency–good LaTeX users will be better with LaTeX, and good Word users will be better with Word. I have colleagues who can only barely use LaTeX, so it takes them forever, and they'd be much better off with Word. We do all our paper collaborations using Word, for example (which drives me nuts but there we go).

Re:An Efficiency Comparison of Document Preparation Systems ... by maxerickson on Wednesday January 14, 2015 @09:32AM (#8886421)
Re:An Efficiency Comparison of Document Preparation Systems ... by dgacmu on Wednesday January 14, 2015 @11:26AM (#8887223)
CS professor here - I use LaTeX, and my wife, also a professor in a different field, uses word. From watching both of our workflows over quite a time, a few things jump out very clearly:

1) LaTeX + revision control is better for multi-user concurrent collaboration. Word + track changes can get the job done for one-person-at-a-time as might be seen more often when making revisions. (We haven't compared word+revision control). But the typical Word workflow involves emailing documents to collaborators - barf.

2) LaTeX+bibtex is better for references. As far as I've been able to tell, EndNote is a steaming pile of pain that crashes frequently and takes Word for the Mac out with it. And glitches in horrible ways.

3) Word users are more likely to get their table formatting the way they want it in less time. I hate playing that game in LaTeX. Particularly if you want fancy nested stuff that looks right, is space-efficient, etc.

The study linked in this article notably omitted two of those that are quite important at least in engineering disciplines -- multi-author collaborations are the norm.

But the recommendations are asinine. In the engineering and sciences, at least, the amount of time that goes into the physical preparation of the paper, seated at your computer, writing de novo text, is nothing. It's trivial. You could crank out a full-length CS paper (12 pages, 2 column, 10 point text) in a few hours, were you just limited by your typing speed and the time taken to format it. If that were the case, I'd have more papers than Erdos!

Re:An Efficiency Comparison of Document Preparation Systems ... by w1ntermute on Wednesday January 14, 2015 @11:31AM (#8887247)
> and takes Word for the Mac out with it

This is another important point.There are incompatibilities between the Windows and Mac versions of Office. I've never seen any such cross-platform issues with LaTeX.

Re:An Efficiency Comparison of Document Preparation Systems ... by billsix on Wednesday January 14, 2015 @10:12AM (#8886677)
I've never seen effective version control of Word documents, but plain text formats such as LaTeX work as expected with git/mercurial/subversion. Very important for works with multiple authors.
Re:An Efficiency Comparison of Document Preparation Systems ... by ObviousScience on Wednesday January 14, 2015 @10:23AM (#8886754)
There are two major faults in this study:

1. They compare only "errors", but don't have any measure of overall document quality. It's certainly the case that it's easier to make mistakes in LaTeX, but I find that people who use LaTeX tend to end up with more professional documents.

2. They compare only on very short documents using very little time. In much the same way anything I'm going to type in 20-30 seconds is better served by Notepad than Word (mostly because Word isn't always even usable in that time window), they're failing to compare LaTeX to Word in the situations where I'd argue that the distinction matters - namely, large, complex document writing. Things with >40 hours of work involved, >100 numbered items, >50 internal references, etc.

Re:An Efficiency Comparison of Document Preparation Systems ... by daly on Wednesday January 14, 2015 @10:47AM (#8886930)
Three major points were missed.

First, TeX is archival. I can still tex documents from my work in the 80s and get the same results. My MSWord files won't load. Thus, MSWord documents die.

Second, TeX allows Literate Programs. I can write a trivial program to extract my source code from TeX documents that explain the code. MSWord documents are opaque. Thus, MSWord limits what you can do with the sources.

Third, TeX is a language. I can (and do) write style files containing macros for my markup styles. In MSWord, what you see is all you are ever going to get. Thus, MSWord is fine if you can't program.

None of these features were mentioned. All of them are important, at least to me. I care that my work survives (archival), that my code can be explained (literate programming), and that I can optimize my workflow (style files).

If these additional constraints were added to the study then I think you'd find that some tasks are impossible in MSWord.

That said, this is a "bikeshed" paper. Emacs vs VI. Noise. Worse, it suffers from "hasty generalization". Sigh.

Re:An Efficiency Comparison of Document Preparation Systems ... by wtbob on Wednesday January 14, 2015 @09:42AM (#8886476)
Already discussed, and issues with the article itself are identified in http://serialmentor.com/blog/2014/12/27/post-publication-rev...
Re:An Efficiency Comparison of Document Preparation Systems ... by fluidcruft on Wednesday January 14, 2015 @10:14AM (#8886699)
FYI, the criticisms there are actually trivially handled by using the MathType and Endnote packages.

The biggest challenge is I work in a field where Word dominates and collaborators don't do LaTeX and journals won't accept it. So, there is very little payoff for slogging over the LaTeX hill, unfortunately. The LaTeX to Word translators all blow and Track Changes rules the world for collaborative writing.

There is no workflow that is not a pain in the ass to collaboratively write LaTeX with someone that can't use LaTeX or that sucks at LaTeX.

Re:An Efficiency Comparison of Document Preparation Systems ... by jjgreen on Wednesday January 14, 2015 @10:34AM (#8886830)
This is so true, I've tried Word/LaTeX collaboration twice, each time I spend more time cleaning up the conversion mess than writing the sodding paper; learn from my fail.
Re:An Efficiency Comparison of Document Preparation Systems ... by jwl on Wednesday January 14, 2015 @10:11AM (#8886668)
Did they take into account something like version control? Because, I have a hard time seeing using Word on a collaborative paper - or even just as the single writer - being very effective in terms of version control.
Re:An Efficiency Comparison of Document Preparation Systems ... by Fomite on Wednesday January 14, 2015 @11:16AM (#8887157)
This is something of a double-edged sword. Writing in LaTeX requires all the collaborators to be familiar with it - nearly everyone is familiar enough with Word to make that work, but as multidisciplinary papers become more common, the same is not necessarily true for LaTeX.

I encountered this when working with a group of both applied math people and clinicians - the applied math people all wanted to use LaTeX, but the gains just weren't worth the learning curve for the clinical co-authors.

Re:An Efficiency Comparison of Document Preparation Systems ... by duaneb on Wednesday January 14, 2015 @11:51AM (#8887391)
Surely google docs is better for collaboration than word is, though, unless you're doing serious math typesetting.
Re:An Efficiency Comparison of Document Preparation Systems ... by Fomite on Wednesday January 14, 2015 @12:21PM (#8887615)
Google Docs is decent, though I do usually have a fair amount of math typesetting, but one of the key issues with Google Docs is you have to trust everyone doing the writing to not change the tone of the piece, etc. It's less conducive to the "Single Approving Authority" model that I've found useful in some recent papers I've written.
Re:An Efficiency Comparison of Document Preparation Systems ... by phreeza on Wednesday January 14, 2015 @10:18AM (#8886731)
The more senior people in my research group actually prefer Word (and OpenOffice) for the reason that it does have an inline, visual change tracking feature. Not as powerful as git of course, but it does actually get the job done and is a lot easier to use.
Re:An Efficiency Comparison of Document Preparation Systems ... by Thrymr on Wednesday January 14, 2015 @12:11PM (#8887535)
> I have a hard time seeing using Word [...] being very effective in terms of version control.

If you're writing a paper with other software developers who already use git, this is very true. If you are collaborating with people who have never used a version control system, it is very much the opposite.

Re:An Efficiency Comparison of Document Preparation Systems ... by kh_hk on Wednesday January 14, 2015 @10:15AM (#8886703)
I don't think time spent and performance are good metrics on the quality of the result, not meaning the output look but the quality of the work itself. Anyone that has spent some time on SQA knows that.

For the study they used a predefined text, so there's no real way to measure if there's an inherent quality on documents created using LaTeX, or if LaTeX does something to your brain to conceptualize your ideas while you are punching them on the keyboard. Or not.

    > Given these numbers 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. [...] 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.
This just sounds hasty. LaTeX may not be the tool for everybody and should not be enforced. Neither MS Word.
Re:An Efficiency Comparison of Document Preparation Systems ... by debacle on Wednesday January 14, 2015 @10:02AM (#8886607)
This has been my experience. Watched a junior dev spend 3 months of 60 hour weeks trying to create a "documentation" platform in LaTeX for what amounted to very basic algorithms.

LaTeX is a good tool, but people who use it tend to abuse it. If you need LaTeX for your solution, there's really nothing better for the job, but you rarely need LaTeX for your solution.

Re:An Efficiency Comparison of Document Preparation Systems ... by jeremysmyth on Wednesday January 14, 2015 @09:23AM (#8886365)
Well this is disappointing.

I wonder if they extended the study to include Markdown would that tilt the balance somewhat. The overhead associated with LaTeX might come from the fact that there's a compile cycle to get the formatting right, although they mention Kile as an editor so that's not clear.

Markdown also has fewer features---therefore a smaller learning curve than LaTeX)---and as a frequent user of Word (and lowriter), I know I interact directly with fewer features of the wysiwyg editors, given that many features are implicit or default, whereas I must manually invoke all features that I use in a LaTeX document (those I haven't placed in a .sty file anyway).

Re:An Efficiency Comparison of Document Preparation Systems ... by klancaster on Wednesday January 14, 2015 @09:37AM (#8886454)
I use markdown (in Scrivener) for my academic writing, and include raw LaTeX as needed for equations. I used Word in my masters' thesis many years ago and will never go that route again - too many issues with Word being "smart" with bullets, etc.
Re:An Efficiency Comparison of Document Preparation Systems ... by mdisc on Wednesday January 14, 2015 @09:43AM (#8886485)
I think continuously compiling tools like WriteLatex (now https://www.overleaf.com/) help close the gap a little bit, but I admit to still spending stupid amounts of time getting things like captions, spacing, and other random formatting to work.
Re:An Efficiency Comparison of Document Preparation Systems ... by Ar-Curunir on Wednesday January 14, 2015 @09:53AM (#8886557)
Editor of choice makes a big difference.

I can write LaTeX in Vim much faster than I can edit text anywhere else. Plus I have automatic refresh on saving and rudimentary spell check in Vim. Why would I use Word?

EDIT: Plus I have autocomplete set up that makes life even easier.

Re:An Efficiency Comparison of Document Preparation Systems ... by Ar-Curunir on Wednesday January 14, 2015 @09:33AM (#8886429)
Re:An Efficiency Comparison of Document Preparation Systems ... by phreeza on Wednesday January 14, 2015 @09:44AM (#8886490)
I think LyX kind of hits the sweet spot between WYSIWYG editing, which helps immensely with tables and free text, and the power of LaTeX for equations.

One thing I never understood is why it has its own file format and doesn't operate on plain latex files. That makes interoperability with latex and in particular colaboration with non-latex users harder.

Re:An Efficiency Comparison of Document Preparation Systems ... by jdpage on Wednesday January 14, 2015 @09:58AM (#8886583)
It's because LyX is a structured document editor, and its format is strictly simpler than the full power of TeX. (La)TeX isn't actually that structured, and with the ability to do weird stuff with macros, it becomes impossible to convert losslessly from TeX to any more structured format (e.g. docx; something like ps or pdf is less structured).

In other words, it can't operate on plain LaTeX files because someone not using LyX might edit the files in such a way that LyX can no longer handle them. By using their own format, while they don't prevent someone from editing the file without LyX, they can at least guarantee that a valid LyX file will always be loadable by LyX.

Re:An Efficiency Comparison of Document Preparation Systems ... by phreeza on Wednesday January 14, 2015 @10:12AM (#8886683)
The weird thing is, LyX can in fact import LaTeX files, and dump any unrecognized command into a red box. If it were just possible to use the imported latex file, and keeping latex as the default format while avoid reformatting the entire file upon re-exporting it to latex, that would make the interoperation possible.
Re:An Efficiency Comparison of Document Preparation Systems ... by Menge on Wednesday January 14, 2015 @09:48AM (#8886524)
Interesting. My first experience in undergrad with LaTeX, I submitted the paper over 3 weeks late with a grade per week penalty and got a B+. I never asked the teacher about my impossible grade, but I would assume I had one of the best papers in the class; usually I'm not good at structuring papers.
Re:An Efficiency Comparison of Document Preparation Systems ... by ajarmst on Wednesday January 14, 2015 @05:58PM (#8889724)
Re:An Efficiency Comparison of Document Preparation Systems ... by ajarmst on Wednesday January 14, 2015 @06:02PM (#8889760)
Re:An Efficiency Comparison of Document Preparation Systems ... by iseewhatdidt on Wednesday January 14, 2015 @11:00AM (#8887041)
I think that there is more fundamental cause for this correlation. See reference https://mjambon.github.io/vim-vs-emacs/
Re:An Efficiency Comparison of Document Preparation Systems ... by josinalvo on Wednesday January 14, 2015 @10:36AM (#8886844)
surprising amount of conclusion from a small amount of fact.

I am not even sure time writing is worth optimizing in the first place.

BUT the evidence in increased errors and time lost might mean that the cognitive load of writing latex is higher, and that might be really relevant

Re:An Efficiency Comparison of Document Preparation Systems ... by guardian5x on Wednesday January 14, 2015 @09:38AM (#8886457)
I wonder which version of Word the study used. I couldn't find anything about that.
Re:An Efficiency Comparison of Document Preparation Systems ... by cefstat on Wednesday January 14, 2015 @10:06AM (#8886632)
The following appears verbatim in the "Supporting Material":

> The study was conducted when most of the participants in the Word groups used Word Version 12. Only 3 participants used older versions of Microsoft Word. Newer Versions of Word include a new Equation editor that makes the software even more powerful and easier to use.

The authors of the article are, obviously, completely unbiased.

Re:An Efficiency Comparison of Document Preparation Systems ... by dubya on Wednesday January 14, 2015 @10:41AM (#8886879)
I think one big plus for LaTeX is its slowly changing nature. Word was appalling 15-20 years ago for writing technical papers. If you gave up then and learned TeX/LaTeX, you're most likely over the learning hump by now and have just needed to learn the occasional new package.
Re:An Efficiency Comparison of Document Preparation Systems ... by Fomite on Wednesday January 14, 2015 @11:18AM (#8887175)
On the other hand, with Word's vastly improved equation-handling capabilities, if you haven't gotten over the learning hump yet, there's a diminishing return for doing so.

I have, for example, never really felt the need, and learning LaTeX has always been binned for "It'll be slightly more of a hassle in Word, but it'll still be done today."

Re:An Efficiency Comparison of Document Preparation Systems ... by Someone1234 on Wednesday January 14, 2015 @09:58AM (#8886577)
Strangely it doesn't list which version of either was used. Which is super odd to me as it may be relevant when others try to reproduce the study.
Re:An Efficiency Comparison of Document Preparation Systems ... by cefstat on Wednesday January 14, 2015 @10:12AM (#8886678)
The main problem in reproducing the study is that it is very vague what counted as a formatting error. It would have been very reasonable to just upload the output from each participant so that the analysis of the results could be reproduced.

But the main problem of the study is not its reproducibility. The main problem is that it is designed in a completely wrong way and it cannot seriously claim to answer the questions it pretends that it answers.

Re:An Efficiency Comparison of Document Preparation Systems ... by cjoelrun on Wednesday January 14, 2015 @09:58AM (#8886579)
Org mode to LaTeX conversion makes the production much less time consuming.
Re:An Efficiency Comparison of Document Preparation Systems ... by plg on Wednesday January 14, 2015 @09:38AM (#8886456)
and I may suffer a loss of efficiency (time & effort, online-driving-difficulty, problem solving, required training) driving my standard transmission sports sedan compared to if I were to drive an automatic transmission minivan

guess what though

I prefer the sports car