A couple of weeks ago, a friend and colleague submitted her Chemistry PhD thesis. Another is currently thinking about writing hers. It prompted me to think about writing. For most people, a thesis is the longest document they have ever tried to write. It is also the longest document many will ever write. Mine came in a little under 220 pages, including all the front matter (stuff like the table of contents, acknowledgments and such). I’ve seen them vary in length from a little over 100 pages to nearly 400, depending on the subject matter. One aspect that all theses share is their complexity of structure, something that is an inevitable result of condensing down around 3 years of research into a coherent story. Good research does tell a story; in some regards it is similar to a detective story. You begin with something that is not understood; a mystery that you want to solve. You make observations, theorise as to their cause, and set about proving or disproving your theories. Akin to the most complex fiction, there are often misdirections, which are also of significant importance: discounting suspects narrows down the options! Hopefully in the end, there is a successful conclusion, but in many regards it is the path taken to reach the end that is most important.

The analogy with fiction-writing also brings us to another aspect of thesis-writing. If you have not really considered what the story is before you write it, the task of bringing together all the threads is all-the-more difficult. This is where paper-writing comes in! If you have written several journal articles on your work before you have to write a thesis, part of the story is pre-formed. I liken it to editing a 2 hour film from 200 hours of shot footage: if you begin with 200 hours of footage, with no screenplay to follow and without having individual scenes in place that can be combined, it is an near-impossible task to create a coherent story. Alternatively, with these things in place, the task is easier and you are in a position to produce a much more polished result.

Unlike most works of fiction however, a major component of a thesis is referencing and cross-referencing. We typically refer to citing external work, such as related journal articles or ‘papers’ as referencing. This forms a large part of our evidence, building on the work of others that are the foundations of what we are ourselves investigating. Cross-referencing is the citing of different parts of the thesis; referring to a different chapter or section, or a table, figure, reaction scheme,...

Each of these places a demand on the author, above and beyond content creation. When you write a thesis, you are all at once an author, copy-editor, proof-reader, graphic designer and critic. Being able to concentrate on each of these tasks individually, rather than simultaneously, relieves a significant burden from the writer.

The last thing you want to be worrying about when you are trying to put down on paper your greatest ideas is whether you have numbered all your references correctly or you have remembered to change the font of the section heading to the right combination of bold 16pt san-serif text that you used last time. Imagine the scenario where you have cited 300 papers in your thesis (quite probable in chemistry), and you suddenly find a reference that you need to include on the first page of your introduction. Now remember, the scientific citation convention is to label references by number, relative to a list that appears in citation-order at the end of the thesis. Now also imagine that you have input all your citations by hand. You now have to spend a couple of hours adding one to most of the citation numbers in your thesis. Ok, that is doable. Painful, crazy, but doable. Now imagine that you decide that one of your sections is in the wrong place, and suddenly 20 of your references have moved. It starts to get confusing and time-consuming, and mistakes become inevitable.

Equally, you refer to a section in a later chapter from the introduction, and from a few places near the end. It contains some figures and some tables, which are directly referenced in the text at various points. You move the section, including the figures, to improve the flow of the writing. You have to renumber all the cross-references, update the table of contents, list of figures and the references; keeping track is a difficult task!

You might think this sounds crazy. But people do this. Fifty years ago, people had to do this, there was no choice (except perhaps to get someone else to do it for you). Today, well, people can choose not to. The most sensible idea is probably to not try and write a thesis ;), (those of us who do write theses are not renowned for following sensible ideas). But assuming you do want to write a thesis, there are more sensible ways to approach it, leaving all the decisions to a computer that can trivially work all these things out.

This is where Donald Knuth comes in. Knuth wanted to write a cross-referenced book that contained a large proportion of mathematics. He found there was no method of producing a professional-looking manuscript. Out of this need, TeX was born. He developed more than a typesetting program: an entire programming language dedicated to typesetting, immensely powerful, producing professional, print-ready documents. Later, Leslie Lamport would create LaTeX, a series of macros written in TeX, that make many repetitive tasks easier by automating them; sectioning, referencing, cross-referencing, page layout,...

LaTeX is widely used in the scientific community. It doesn’t have the ease-of-use of your standard word processor (it has a steeper learning curve than simply starting to type, and it isn’t WYSIWYG), but I for one do not find word processors easy to use. I am always fighting against what it assumes I am trying to do. No, I don’t want to change (c) to a copyright symbol, and I wouldn’t even be typing it if you could make a numbered list properly! I am a long-time convert; I first learned how to use LaTeX back in 2004. Now I use it whenever possible. I would rather have WYWIWYG; what you want is what you get!

Thankfully my friend who just submitted her thesis wrote it in LaTeX. I can’t take credit for her decision. I’ve tried (and failed) to convince many people to use it; to my knowledge, I’ve never been the determining voice. Thankfully though, many of these people have been encouraged from other sources, and they have at least been spared one of the pains of thesis writing, which is arguing with your word processor.

If you choose the path of insanity, and choose to write a thesis, at least do one sane thing and write it with LaTeX! The process is far less painful. The final selling point: LaTeX is entirely, completely, 100% free!!

Google has endless information on LaTeX—the top results are, surprisingly, about typesetting. Unless you use the images search! The next post on LaTeX will be a little more practical introduction, including details on how to obtain it for your OS and some useful basics.