Catastrophe-Proof Your Dissertation

Megan Poorman recently completed her Ph.D. in biomedical engineering at Vanderbilt University and is currently a postdoctoral researcher at CU Boulder and associate of NIST. You can find her on twitter @meganpoorman or on her website.

This is part two of my dissertation writing series where I will discuss how to write your dissertation in an organized and catastrophe-proof fashion. If you missed part one, check it out here: Pre-Gaming Your Dissertation.

We’ve all heard it: the horror story of the student who forgot to backup their dissertation and lost it all in a freak accident. While this scenario is terrifying, it is thankfully one of the pre-defense nightmares that you can completely avoid with a proper writing system. Before you even begin,it’s important to set yourself up for success by getting the right tools and using them well.

In this post I will share strategies to structure and backup your dissertation files. Obviously there are many ways to set up your dissertation, so I will cover some general advice and discuss why I chose to use LaTeX with Git version control to do mine.

Consider your program options
There are a million and one word processor options out there, but two of the most common are Microsoft Word and LaTeX. The debate between which ones is “best” is heated,but I know people who have successfully written their dissertations with both programs. I chose to use LaTeX because it lent itself very well to equations, formatting, and figure placement but it isn’t the easiest to read for sentence structure. What is best for you is the program within which you’re most comfortable with the features. I recommend automating as many processes (such as formatting and linking to figures) as possible within the document to make your life easier when you make edits. If you aren’t sure how to automate things such as bibliography generation or caption formatting, both programs have a plethora of how-to guides available online.

Find a template
Once you’ve selected your weapon of choice, it’s time to set up your document. Many universities have a style guide that you must adhere to in order to have your dissertation accepted. Thankfully, many also provide templates. Take a look at your graduate school’s website for a LaTeX or MS Word template to avoid having to start your formatting from scratch. You should still read through the style guide, as sometimes the templates aren’t 100 percent correct. I also suggest downloading a few dissertations of people from your department whose work you respect. This will give you an idea of how they structured theirs, how many pages are expected, and help you create a template if you cannot find one.

Create a file structure and naming convention
Now that you’ve got your template it’s time to create a file structure. Create a folder on your computer named “dissertation” and create subfolders for your “figures” and main “manuscript.” You could also create a folder for relevant literature, but I suggest using a citation manager to help you with the actual referencing. In my “figures” subfolder I created additional subfolders for each chapter I was planning to write. That way I was able to quickly find the relevant figure when I needed to add it to my document or make changes.

Try to save your figures with informative names so that you know what they are without opening them or even with the figure number so you know the order they will appear in. Any files pertaining to the actual writing (such as an outline or LaTeX subfiles) I saved in the manuscript subfolder.

When you save your main manuscript document, decide on a strategy to keep track of versions (like including the date and who edited) and BE CONSISTENT. Now, saving a new version each time can get annoying and add up in disk space very quickly. One of the main advantages of using LaTeX/Git is to avoid this pile-up of versions. I will discuss this in more detail below, but if you’re using MS Word consider saving your document with a service such as Dropbox or OneDrive that automatically has some version control features.

Back it up to multiple locations
ALWAYS, ALWAYS, ALWAYS, KEEP YOUR FILES SAVED IN MULTIPLE LOCATIONS. This is a situation where it’s okay to be paranoid. I was terrified of a freak accident and thus not only kept my files in the cloud with BitBucket (Git) but had a local copy on my desktop, a backed-up version with my backup service, a copy on my external hard drive that stayed at home, and a copy on my USB memory drive that lived in my backpack. While five copies may have been excessive, I knew that I would always have access to a copy of my dissertation regardless of any accident. If you’re not quite as paranoid as me, at the very least make sure you save your files in the cloud with something like Dropbox, Box, OneDrive, Apple, etc.

The dream team: LaTeX and Git
I’m about to sing the praises of LaTeX so MS Word users should either bear with me or skip to the last section (I won’t be offended). For me LaTeX was a natural addition to my writing arsenal. I will fully admit, however, that it’s not the best tool for everyone. In writing my dissertation, the advantages were:

1. It is awesome with equations –  you quickly type in-line code instead of clicking a ton of buttons just to get one summation to look right. Being in engineering, my dissertation comprised a lot of equations so this was a no brainer.
2. You only mess with formatting once – Using the LaTeX template provided by my university meant whether it was adding a new paragraph, sizing a figure, or changing heading style, LaTeX automatically set those for me. When I did make formatting changes, I only had to change the style in one place and knew the change would be automatically propagated to my entire document without me searching 100+ pages and inevitably missing something.
3. LaTeX is saved as plain text – This means that the files are very lightweight, don’t take a lot of computer memory to open/edit, and are easily parsed by Git.
4. Git version control means never losing anything – You simply link your files through a cloud-based Git repository (like BitBucket or Github) and you can access and edit them locally anywhere, regardless of operating system. It keeps track of version changes without having to rename and save multiple files. The system shows you what edits were made, when they were made, and highlights exactly what lines were changed. At any point you can revert to an old version and create a new branch to test out an idea without messing up your original thoughts. This saved me much confusion on file renaming and kept my manuscript folder clean. It also meant when my readers had suggestions I could easily find where they had made changes.

5. You can put chapters in separate files – Instead of scrolling for pages and pages to find the exact paragraph I wanted to work on, I was able to open only the chapter I wanted. Each file was easier to navigate and I didn’t have to open the entire document every time I wanted to edit, which was life-saving on a slow laptop. Instead, I just included the separate LaTeX files in the main document where I wanted it and compiled that document if I wanted to see the whole thing.

6. The automation is to die for – LaTeX automatically updated my bibliography, figures, and table of contents each time I saved my file. There was no forcing a refresh, worrying if my captions stayed with the right figures, or making sure my figures fell at a reasonable spot on the page. LaTeX kept an accurate, real-time record of where everything was located and placed figures where it made sense in the white space with no input from me.

You want to use LaTeX but don’t know how
So, I’ve sold you on LaTeX and Git and you want to give it a shot. While I will admit there is a (semi-large) learning curve, thankfully there are also a lot of tutorials online about using these tools and even some specifically geared towards writing dissertations. If you’re crunched for time, I suggest trying one of the online LaTeX editors like Overleaf. These sites make writing in LaTeX similar to writing in a conventional word processor and handle all the file structure, compilation, and backups behind the scenes. Some journals have even partnered with these sites to allow submissions through them, which could be particularly handy for STEM fields.

A final note on processing power
It quickly became apparent to me as I started thinking about writing my dissertation that I was in desperate need of some at-home computer power. My poor laptop took 30 minutes of chugging along at glacial pace before it was ready for use, and even then it could barely handle a PowerPoint presentation. When the new year rolled around I decided to pre-emptively reward myself for writing my dissertation and build myself a desktop computer. This decision made a sizable dent in my wallet but it allowed me to write my dissertation without the frustrating distraction of a slow computer.

Your dissertation document is going to be large, with hundreds of pages and many figures. You’re going to open, close, edit, and make copies of that document thousands of times. That is why you are going to need a computer that doesn’t make you want to bang your head against the wall while you wait for it to perform all these actions. If you have the means to do so, invest in a computer that will handle large files smoothly and save yourself the excuse to procrastinate on writing.

What program did you use to write your dissertation? How many locations did you save your files in to put your mind at peace?

[Image by Flickr user Alex Proimos and used under a Creative Commons license.]