Tips for writing your thesis

Summary

• Get a good Latex editor
• Make a Latex outline of your thesis
• Write notes into the outline as you go as if writing for your supervisor
• Get a good reference manager
• Download more references than seems strictly necessary
• Expect the gremlins to move in with you and not pay rent
• Buy a gigantic backup drive to do automatic 30 min backups of your whole system

Get a Good Tex/Text Editor

My first piece of advice would be to find yourself a great tex/text editor and spend a weekend getting really good at Latex and learning all the keyboard shortcuts for your editor. My personal favorite is TextMate which I think is the best one for the Mac there is and imho, is better than Kile for KDE in some ways (but it doesn’t have all the fancy buttons….).

TextMate is not free but is cheap considering its capabilities and it has made my life a million times easier since I got it. It has bundles available for most scripting/programming languages (even IDL) and is shockingly customizable. For Latex editing, I use it with PDFsync and Skim

Get a Reference Manager

Get a very good reference manager and immediately start collecting references from day one to begin your Bibtex library. The best and most incredibly fast way of doing this is to start with a thorough review article from your topic and then use the link on the ADS page to download all the references they cite. Don’t worry about downloading bibtex references for papers you’ll never use because they don’t end up in your bibliography if you don’t cite them. It’s better to be safe than sorry because there is nothing more annoying than getting into a great rhythm with your writing and then have to stop to download a reference. This will really kill your momentum. If it were possible and not completely OTT, I’d download the entire ADS in bibtex format.

My personal preference for reference managers for Mac is Papers which is also not free but just as with TextMate is well well well worth the small amount of Euros. They also kindly offer a good student discount and you can load the program on more than one machine as long as it is yours. Papers is very capable. It can import directly from ADS or arxiv, download pdfs, auto-match citations from pdf imports, download abstracts, auto-organize your entire library, auto-generate citation keys, and can import and export to a large number of formats (like bibtex). Seriously, it is great. A good free one for all platforms is Jabref and a free one on the mac included in the TexLive package is Bibdesk.

Make a Thesis Outline

Sit down with your supervisor at any point of your course, and make an outline. It will definitely change as you go along, but just having it on paper is a huge help for getting a grip on just how much work there is to do. Ask them for a list of the definitive papers in your field, review articles, and for a list of recent papers to consider. Then in Latex format, make a rough outline with all the important sections you will need to have. Introduction chapter, science chapters, and conclusions chapter. This will help you clarify what exactly the bigger picture is and how your research fits into it. What are the big questions in your field? What are the nuanced debates? How does your research help to answer some of these questions? You obviously don’t need to know or fully understand everything when you start, but having time to frame the debate and have it sit in your head for a while will make the writing a LOT easier later on.

Fill Sections In As You Go

The worst mistake that I made was to take notes on paper, in various plain text editors, on post-its, in multiple computers, and on top of print-outs. While this was definitely better than taking no notes at all, it was enormously painful to convert little notes to myself into reasonable sounding thesis-speak when the time came. The things that take the most amount of time in your thesis are the small, completely tedious tables, little numbers, calculations, and paragraphs about reduction techniques(or modeling methods). Instead of writing it on paper, when you find the instrumental dispersion of your instrument and your estimate for the wavelength calibration accuracy put it directly into your outline! Make a quick latex table with the information, write a few sentences about how you got it, and put in whatever plots you made to show your supervisor with a caption explaining what it is. No matter how small the detail, it will be recorded and the tables will be made.

Basically what I’m saying is, whenever possible, write your notes in Latex directly into your outline.  This also makes writing papers easier too.

It is infinitely easier to take stuff OUT and revise something already there than to make it from scratch 1 month before your deadline from a dusty old notebook which didn’t even have any context because you wrote it for yourself and not for someone else. Two years later (or even two months later) you WILL be someone else. Someone else who has no idea what this weirdo first year was talking about when she wrote “calib acc 5% tested” in the margins next to some dark print out with some lines on it and no labels. Write your notes directly into your thesis as if you were writing them for your supervisor to read….say…..3 years later.

And more than anything, when the time comes to write, you will just want to WRITE and not be making dumb tables. Man, I hate those tables….

Double All Time Estimates for the Gremlin Wars

I didn’t quite have this luxury since I was on a huge deadline to finish on time for a job, but if there is a date you are aiming for, start the writing-up process a good 4 months in advance and anytime you think “this will take me a week to finish” secretly, and very quietly plan for 2.

Here’s why: I’m not sure what it is, but there is something about writing your thesis that sends a signal to the evil grelims from hell that NOW NOW NOW is the time to cause the most statistically improbable amount of havoc in your life. It’s almost as if you getting your PhD is a threat to their livelihood. They will stop at nothing to stop you!

Plan for any or all of the following kinds of things to happen at the worst possible time: all your family members start having problems, your boyfriend/girlfriend/special friend decides to dump you and/or get into lots of fights with you, ALL YOUR COMPUTERS WILL BREAK, you lose your entire documents directory, you get diagnosed with several weird disorders/diseases, all your best friends get married and then refuse to talk to you for missing their weddings, long lost loves emerge from nowhere with pre-paid tickets to Paris the weekend before your thesis is due, the most important presidential election in centuries fills up the news, Rafael Nadal beats Roger Federer, the most amazing Olympics in memory takes place, a hurricane nearly knocks down your house, oh and the global financial system completely collapses……etc, etc, etc, not a joke. Invest in a sword, and some good armor, or at least a good water gun.

Ok, it’s different for everyone but the basic thing to know is that you could write 3 different 50 page scientific papers consecutively and stuff wouldn’t go wrong more than usual. But for some reason, writing 150 pages of THESIS means that all hell breaks loose. I’ve come to realize that you don’t get a PhD for all that research that you did but rather as a badge of honor for slaying the gremlins and not calling it quits. This battle really caught me off guard and I spent way too long worrying that the apocalypse was nigh or “WHY MEEEEE????”-ing myself into a philosophical hole. But you’ve been warned so just prepare for and expect the worst and you should be ok!

Buy a 1TB Backup Drive

Speaking of which….

I probably should have put this first. Let’s pretend that I DID put this first but I’m putting it here because you’d skip over it if it was first and go yeah yeah yeah, but maybe now that you’re privy to the existence of the anti-PhD gremlins and are sufficiently fearful, you will listen to me.

Buy a 1TB backup drive (ok 500GB would do) that you can permanently hook up to your computer then have it automatically backup your work every 30 mins and save the all backups until it is full. In Mac OS X Leopard, the Time Machine application does a great job of this. If you are running another operating system, Carbon Copy Cloner, or various other free drive cloners with automatic incremental backups will do the job just fine. Do not use the drive for anything else but as a bootable drive for your machine with as many 30 min iterations of back ups as it can fit. Just hook it up and let the program save everything in the background so you never have to worry about it….you’ll have enough to worry about.

Furthermore, identify 2-4GB worth of stuff that you would absolutely jump off a cliff if you lost and either make weekly back ups of this to DVD, or use an automatic and free web service like Mozy to backup onto secure servers on the web. Dropbox is also another good option. If that’s not possible, then at the very least, zip up your thesis directory and send it to a friend to store for you. If you only have one backup drive that is in the same room as your computer, you are pretty much ensuring that the gremlins will either set fire to the house, or break in and steal all your stuff. But putting your most important stuff at another location will make them try another tactic…..like making it rain every day!