Writing Scientific Documents Using LaTeX (2009) [pdf] | Hacker News Reader

Writing Scientific Documents Using LaTeX (2009) [pdf]

mirror.unicorncloud.org

53 pointslainon23 comments

If you're just getting started with LaTeX, you can do so without needing to install anything by using a cloud-based service such as Overleaf [1] or ShareLaTeX [2] (who've recently joined our team [3]).

ShareLaTeX also provide a very useful LaTeX learn wiki [4], and my co-founder at Overleaf created this three part introductory course [5].

Hope this helps anyone getting started, and feedback on Overleaf/ShareLaTeX is always appreciated if you've used it previously.

[1] https://www.overleaf.com

[2] https://www.sharelatex.com

[3] https://www.overleaf.com/blog/518-exciting-news-sharelatex-i...

[4] https://www.sharelatex.com/learn

[5] https://www.overleaf.com/latex/learn/free-online-introductio...

I used to be a big proponent of using LaTeX for scientific documents. I even wrote a botanical flora using LuaLaTeX. Being able to drop down into Lua was wonderful for formatting scientific names correctly (something that may seem simple, but is not). I don't use it as much anymore for two reasons.

1. It's fragile. Documents I wrote in 2010 no long "build" with the latest TeXLive release, and fixing and debugging my documents has been a huge time sink. With something like a flora or plant list, that needs to be updated periodically with new information, getting sidetracked by breakage can be very frustrating.

2. I've found it's a lot quicker and simpler for me to write in MS Word and then (if needed) import into Adobe InDesign to style my writing. This combination also seems to work better with the citation software I prefer (Papers).

> 1. It's fragile. Documents I wrote in 2010 no long "build" with the latest TeXLive release

Do you have any general guidelines as to why?

Did you use obscure packages or features that no longer exist?

I had been thinking one of the greatest things about the book I am almost finished is that it will serve as a template for me for decades.... maybe I was wrong.

I suspect OP's issues are with his custom Lua stuff.

In order to reply to this message with some facts, I just rebuilt my phd thesis for the first time in 17 years using a modern TeXLive with nary a LaTeX issue. The only effort needed was to re-learn my own work flow and to install some old tool I used to post-process the PostScript.

I'd hazard that LaTeX has better shelf life, in terms of rebuilding, than say C.

> I suspect OP's issues are with his custom Lua stuff

It was due to a conflict between packages (memoir and morefloats) the only seems to happen in recent version of LuaLaTeX. My Lua code was not directly involved, as far as I can tell.

It was due to a conflict between the memoir package and the morefloats package that best as I could tell was only present in LuaLaTeX. My Lua code wasn't a direct cause beyond the need to use LuaLaTeX.

LaTeX is a wonderful gateway for many students to expand their exposure to code.

A lot of students will shy away from coding prefering more hands on work in the lab. Some students will use Matlab or LabVIEW to perform some essential task and hate every minute of it. It might be some stastical analysis they don't understand built with spaghetti code they've inherited, or it might be for controlling a contrary piece of lab equipment. Either way they feel threatened by it or don't want to spend the time learning to understand it.

Show them a basic LaTeX document and how it can make writing papers and, importantly their thesis, much easier than the horror show that is Word and suddenly there is a moment of wonder.

I'd hope just the dramatic improvement in appearance would be enough. Seeing any scientific content in Word makes me ill.

Yes, generally the response to first seeing the output is wonder at how elegant and professional it looks.

> LabVIEW

Oh god, when I was an undergrad we had to spend a semester of junior/senior year doing an applied stint in a physics lab on campus, and I still have nightmares about the decade old Frankenstein's monster of a LabVIEW document that was thrown in my lap... trying to debug thousands of nodes in this flat data flow nightmare while interfacing with shaky equipment held together with scotch tape was horrific.

I immediately noped out to applied math.

I love LaTeX, but I've been really enjoying Markdown in Pandocs as an alternative (with equation support, of course). All of the problems with tables (which are really some of the worst things in LaTeX) are super simple and everything can be written in Markdown format for text/layout along with the usual LaTeX math commands and typesetting.

I've begun using it to write a lot of my notes, since TeXMacs has sadly been failing lately, and to be honest it's wonderful having human-readable text along with a document that compiles nicely to a PDF with all the bells and whistles.

This. And a small Makefile to run pandoc.

Are there any LuaTeX users here? At the moment, I use XeTeX to generate software manuals. Technically, it all works and I'm able to import code samples directly from source into the manual using the listings package. I'm also able to abuse CMake to run certain test problems and import their output into the manual in order to keep documentation more up to date with the code. That said, it's a pain to maintain and it takes much, much longer to build than the actual software project itself (likely due to overuse of listings.) I'm curious if using LuaTeX would make producing complicated documents easier than pdflatex or XeTeX.

For anyone interested in this, may I suggest using emacs org mode and just calling latex (or any other language) when needed for writing, and then you can export to latex+PDF (or HTML, etc).

All the nice looking output if latex but all the nice features of emacs and orgmode.

I love LaTex. Writing without thinking about formatting and get marvelous looking documents is kind of magic. You use any excuse just to put some math in your documents. I've stopped using it due to 2 reasons:

1) Tables are hard. There must be a more visual way do layout a table.

2) All your documents look like a scientific paper. It doesn't have beautiful styles which you can include and have a completely different layout. Like you can do with Word templates.

Does LaTex improved in the last ten years? Or we all must be content with markdown?

For point 1, I always use this. [0]

For point 2, this is still the case.

[0]: http://www.tablesgenerator.com/

I've typeset books, articles, reports, fliers, resumes, chess diagrams, recipes, presentations, and probably a bunch of other stuff in LaTeX over the years... and, no, they don't all look like scientific papers.

LaTeX is really what you make of it. It can look like anything you want. It's just a matter of which styles/packages you use, which options you select, how much tweaking you do, and how much knowledge/help you have when trying to make it look the way you want it to.

For 1), if you're an Emacs user, you can use org-mode tables, and convert them into LaTeX. You can even do that from the same LaTeX document using "radio tables" (https://www.gnu.org/software/emacs/manual/html_node/org/Radi...), which you can edit with all the facilities of org-mode, and then simply convert to LaTeX at the appropriate place.

Regarding 2), you can nowadays find a wealth of document styles to use, if you don't want your writing to look like a CS paper... While creating proper document classes is a non-trivial amount of work (and so it is in MS Word and other WYSIWYG tools), it can then greatly simplify the task of creating the documents themselves. For example, I have recently done a CV class ( https://github.com/mvarela/orbit-cv ), which provides a bunch of macros to automate the creation of your CV/résumé, and looks nothing like a scientific paper...

Regarding 1), I use org-mode with inline LaTeX to write first couple of drafts, which has a nice visual table layout. It also has a Markdown-like syntax for markup. I export to LaTeX after I'm fairly satisfied with the result. For math-heavy documents, I start from LaTeX. In my opinion, that's where it really shines.

Regarding 2), there are many document classes that don't look like scientific papers although LaTeX ships a limited number of document classes (book, article, report, and letter). Here are two third-party classes that come to my mind:

- https://ctan.org/pkg/hitec is a nice class that is

- https://www.ctan.org/pkg/tufte-latex is inspired by Edward Tufte's books and provides a similar-looking handout/book style.

And here is a list of some other classes: http://texcatalogue.ctan.org/bytopic.html#classes

If you haven't, check out LyX, as it dramatically improves the LaTeX experience.

I never liked LyX, at the time I found it difficult to extract a standalone tex file from it.

What are the vital PROs of LaTex over easy-2-use word processing programs like Ms Word ?

I am putting the finishing touches on my first book that will go to print, using latex for the first time since University.

Things like gutters and margins are precisely controlled, can be different for left and right pages, etc.

The ability to define your own macros, that can be used through the doc.

The ability to use conditionals within your document, so it's easy to produce say a print-ready pdf and an epub from the same source.

Really, really easy to put your doc unto source control (cvs or git) and get legible diffs.

IMHO, if you are printing a major document or book for professional purposes, you want to use latex.

Reply on news.ycombinator.com