Have you used LaTeX to write your thesis, in a department/field where everyone uses Word?

• I'm due to start writing my PhD thesis very soon, and am agonising over whether to use Word or LaTeX. What did you use to write your thesis, and did you come to regret your choice part way through? I would particularly like to hear answers from people who are not in one of the normal LaTeX-using disciplines (maths, computer science...), but all opinions are welcome! More info after the break. So I am a physical geographer, about one or two months away from starting to write up my PhD thesis, and in my department, no-one has ever heard of LaTeX. Microsoft Word is the go-to program for writing PhD theses (and anything else). I am having great difficultly in deciding whether to use LaTeX or Word. The issues/pros and cons I have are as follows: - I have already tackled much of the fairly steep learning curve involved with using LaTeX. I decided to write my 20,000 word, end-of-second-year "Thesis Outline" document using LaTeX, as I figured that learning to use a new thing is rarely a bad idea, and I was keen to see whether I could master it. This involved multiple chapters, referencing, figures, tables, lists of contents, figures, tables and acronyms, equations, etc. I found it logical and pretty satisfying when I got my document to look how I wanted it. I used JabRef to manage my references and I found that referencing, in particular, was much better than anything I've ever used in conjunction with Word. - I really like how my document looks and how TeX handles the things I mentioned in the previous point. In particular, I like not having to worry about where my figures are going to go and whether my tables are going to run over onto another page and look terrible. Having produced a fairly sizeable document already, I guess a lot of the stuff like structuring the document, setting up a database of references, doing all the acronyms, and so on, is already taken care of. - I still haven't got used to the fact that you can't see the changes happen as you edit your document, as you would with Word. I find that this slows down the writing process quite considerably, and I don't know whether this will be improved with use, or whether I'll always be thinking "ugh, I can't tell what's going on with all these codes everywhere". - I work on a Mac at home and a PC at university. While this is no problem for LaTeX, or Word more generally, the only Word-related referencing software we have at university (RefWorks) is very messy when constantly swapping between Mac and PC. I guess I could get around this by taking my Mac to uni or installing Windows on my Mac. - Once I start writing my thesis I don't want to have to stop and change from one approach to the other. I know of LaTeX to rtf converters but they aren't perfect and the further through I get, the more work it will be to change approach, in either direction. - I won't have very many equations or formulae (probably fewer than 10) and while I like how LaTeX typesets them, Word is likely to be fine for my requirements. - I won't have any help from inside the department if I go down the LaTeX route, although there are potentially people who use TeX in other departments in the faculty. However, my SO used TeX to write his thesis so will be on hand and is happy to help (although I don't want to be constantly asking for help). I've generally found that people far cleverer than me have found or devised solutions to most problems and that a small bit of googling will usually reveal the answers to most things. - In relation to the last point, I like how genuinely helpful LaTeX users seem to be. People are happy to post their solutions to problems on forums and blogs, and those who really know what they're doing can understand how it all works and get TeX to do pretty much anything they want. This is not the case for Word and every now and again it will do something weird and unpredictable (not to mention the dreaded repaginating, which I think will become an issue in a 50,000+ word document). TL;DR: I can use both LaTeX and Word. I find Word to be less pleasing, but everyone in my department is scared of TeX, and I do find that not being able to see the words as I type them slows down the writing process a little. What should I use to write my PhD thesis?

• Answer:

  I'm a physicist and wrote my thesis in LaTex on a Mac. Word drives me crazy even for a three-page report, so my personal recommendation is against that. The one guy I know (also a physicist) who used Word for his thesis regretted that he did. The reference handlers that are supposed to work well with Word also often don't in my experience. The one reason I can see for doing it is if your supervisor insists on giving his comments using Word functions rather than on paper. I think you'll have a nightmare with that though, with a thousand different versions getting emailed back and forth and mixed up. I think it's crucial to have the "infrastructure" set up in a good way: If you enjoy using your tools and have your files in good order, it's easier to get started at times when the writing itself feels slow. This is what I used and was happy with: TextMate for LaTex editing. I tried several, and this is the only one I really liked. It's clutter-free and has good LaTeX and SVN integration. SVN for version control. You need this! It will do so much for your peace of mind, knowing that whatever you delete or change, you can always come back to a previous version. If you can have a repo on a server, it will also make it a lot easier to swap between different computers. Illustrator for figures. Some figures were originally Matlab plots, some were drawings. I put all of them in a single Illustrator file, each on a separate artboard. That made it easy to ensure sizes and typefaces were consistent, and to copy elements between figures. When you save, you can have each artboard in a separate .eps file. I found this very convenient, although the file became quite big in the end. Mendeley for references. It's nice with an online handler if you work on different machines, and it's BibTeX support is ok, if not perfect. For my bibliography to look nice, I had to use biblatex to exclude some BibTeX fields that Mendeley insisted on including. Make a few macros to include figures and generate labels, so your figure's label and file name always agree. Feel free to MeFi-mail me if you want to know more details. On preview: I don't think you should worry about valkyryn's comment. In every case that I know of, PhD theses have been distributed in PDF format.