1. Writing mathematics | 3. Whitespace |
2. Typesetting text | 4. Typesetting mathematics |
One of the questions that students ask most often is “how much background material should I provide in my thesis?”, or variants like “should I include this theorem?” or “should I include this proof?”. What the student is really asking is: Who am I writing for?
When writing a bachelor or master thesis, you can decide who your target audience is. (Writing a scientific paper follows an entirely different logic, and is not covered here.) A common choice is to pretend you are writing for one of your fellow students – perhaps someone who has taken many of the same courses as you have, but who does not necessarily know about your thesis subject. Thinking about your target audience every now and then will help you decide what background material to include, and what you can safely assume is common knowledge.
That being said, the only person you can really be sure will read your thesis is your thesis examiner. While your examiner will know much more about the subject than your fellow student does, it is probably not so easy to predict exactly what they know, and what they don’t. Therefore, it's a good idea to choose a fellow student as your target audience.
Don’t let your results speak for themselves! While it might be obvious to you why your theorem is important, it might not be so obvious to everyone else – even to a professor working in your field. You need to explain what you have done and why you have done it. Use phrases like “We will show …”, “We have proved …”, “We implemented …”, both in the introduction and in the main text. (It is common to use the first-person plural we, even when you have done everything yourself.)
You should consider ending your thesis with a Summary or Conclusions section where you list your main findings. This is also the place to mention what you haven’t done, but would have liked to – your thesis examiner will appreciate that you are able to think a little bit further.
When explaining your findings, you shouldn’t be afraid that you are over-explaining or bragging about your results. Your thesis examiner is experienced in reading text, and will simply skip a sentence or paragraph they already understand. On the other hand, if your examiner doesn’t understand the significance of your result because you haven’t explained it well enough, then you risk getting a worse grade than you deserve.
Like any other scientific text, your thesis will be structured into \section
s and \subsection
s (as well as \part
and \chapter
, if you are using the book
or report
document classes). Mathematical texts require more structure, however, which is where definitions, theorems, proofs and so on come in.
The article
document class provides three levels of headings: \section
, \subsection
and \subsubsection
. With the book
and report
document classes you get two additional levels: \part
and \chapter
.
Before you start writing, work out a tentative framework of your thesis with chapter and section titles, as well as a short description of each part. Start with only one or two levels of heading (e.g., chapters+sections, or sections+subsections). Only use higher levels of heading (i.e., \subsection
and \subsubsection
) if you really need them; as Robert Bringhurst writes in his book The Elements of Typographic Style,
use as many levels of headings as you need: no more, and no fewer.
The structure of any level of heading, be it a chapter, a section or a subsection, follows a simple logic, summarized here:
Definitions, theorems, examples and its siblings are all defined in theorem environments such as
Structuring your text around definitions, theorems, lemmas, remarks, examples, etc. will make it more easy to read, and will help you structure your thoughts. It makes it much more clear to the reader what are hard facts, and what is more informal discussion. Your text should include a healthy dose of both.
In order to increase readability, don’t let definition or theorem environments (or their siblings) “touch”; try to write at least one sentence before such environments as an informal introduction.
We all learnt it in school: Write complete sentences! Many of us still break this rule, particularly when we are writing mathematical texts. To understand the issue better, it helps to know some basic ideas from syntax.
A sentence is usually composed of an independent clause (e.g. Bob sings) and zero or more dependent clauses (e.g., because he is happy). An independent clause consists of a subject and a predicate:
Dependent clauses usually begin with a dependent word (e.g., because) and continues with a subject and a predicate (e.g., he is happy).
An incomplete sentence is any sentence which does not adhere to this simple pattern. Some examples:
Mathematical expressions can themselves act as both subject and predicate (as in the example x = 5), but you should be very careful about sentences in which mathematical formulae constitute the entire subject or predicate:
Both examples are harder to decipher than they could have been. As far as possible, avoid subjects or predicates consisting entirely of mathematical symbols. Instead, use both words and symbols in both subject and predicate, or push the formula into a dependent clause:
It can be disheartening to receive negative feedback on a manuscript. But try to make the best of it, and take it for what it is: Someone is helping you to do better.
If your supervisor returns your manuscript full of strange symbols, they might be copy editing marks. These are a standardised set of symbols employed by copy editors to correct manuscripts (or, at least they used to, before the advent of Microsoft Word). The NY Book Editors has a nice overview of copy editing marks.
There are plenty of good TeX editors available. In my own opinion, Texstudio is by far the best; others include Texniccenter, Texmaker and Texworks. Although the choice of editor is mostly a question of preference, a good editor should at least have the following features:
When referring to a particular theorem, definition, chapter, section or similar, you should capitalize the first letter: Theorem 5, Chapter 8, Section 3, etc. Remember to insert a non-breaking space between the label and the preceding word: Theorem~5
, Lemma~\ref{mylemma}
(see §3.5 Learn when to use a tie).
If you use the cleverref
package then the command \Cref{somelabel}
will print the type of environment (say, “Theorem” or “Chapter”) as well as the label.
To cite, or not to cite? Before citing a book, paper, website or whatever, make sure that you have read it, that you have understood it (at least partly), and that it's relevant.
If you need a theorem which is not common knowledge then it's certainly appropriate to include it in your manuscript, as long as you cite the original author:
If you need to point to a specific part of a reference, such as a theorem or chapter, then write either, say, Theorem 3 of [1], or use the alternative argument to the \cite
command: \cite[Theorem~3]{SomePaper}
.
See e.g. the LaTeX Wikibook for more on bibliographies.