Ask HN: Do you write papers directly in LaTeX?
Hello,
what is your workflow for creating scientifical papers?
Do you write your research/results directly in LaTeX or do you have a different approach, e.g. creating paper notes before putting it into your word processor etc.
Thank you for replies.
1 comment
[ 0.23 ms ] story [ 13.3 ms ] thread3 months before...
“huh, I have that coming up, I should probably get something down”
writes an abstract on latex then gets distracted by an article on containers
2 months before
“I really should start thinking about that...”
reads around relevant papers and review results I should have done last month, makes some paper & electronic notes
1.5 month before
“I’ll look at it once I’ve got this experiment done”
gets distracted by my .bashrc file
1 month before
“S%%t! S%%t! S%%t!”
write 10 pages of _only words_ into a latex document in a git repo
3 weeks before
edit all the words, making sure to use branches / tidy commit names... render & send to supervisor
1.5 weeks before
tables, references, diagrams... make edits according to supervisor comments
1 weeks before
formatting, final editing, send to supervisor for final review
Last day
“thank fking christ”
final 5 page version
...
So... I’m a spree type of person. You may be more methodical.
I’ve found latex & git to be sufficient however.
Your first draft is supposed to be awful. Things should be wrong. Sections should be a mess. Forget tables and diagrams. References aren’t that important right now. Getting information down is.
Splurging your knowledge onto paper is the name of the game.
Any word processor application can do that. So latex, word, notepad, whatever. Just get it down.
After that is editing. This is where version control (git) comes in SUPER HANDY and I’d recommend avoiding MS Word if at all possible.
It’s easier to look at the changes in documents on GitHub instead of checking out specific commits, closing word, reopening word, sigh wrong commit, repeat.
Latex templates (e.g. ieee articles) also force you to use specific formatting that’s difficult to change unless you go digging.
Saves you time moving that picture around in word until it’s just right when you can just align top centre according to the journal’s spec by default.
I also find referencing much easier in latex when combined with bibtex. Word is horrendous.
So yeah, whatever I can splurge into for step 1, everything else is latex & git for versioning.
Edit:
Forgot to mention. I have reams of notes I’ve made on papers and photos of whiteboards I’ve written “explainers” on. My desk is covered in paper in various different piles.
I usually refer to the actual paper when doing splurge/editing, but if I’ve forgotten a few things then paper notes/whiteboard photos come in very handy for re-understanding what the hell im actually talking about.