Writing a thesis

Writing a thesis is an overwhelming task at the beginning. This tutorial should support you by suppling you with some ideas, hints and rules for a good thesis style.

Your main points

Before starting to write, think about your main points (3-5):

  • What are they?
  • Why are they important?
  • Why is your solution cool?
  • Why is it non-obvious? (What has changed to make it obvious to you? What different angle are you looking at?)

Scientific style

First of all you should be clear in your mind that a thesis is a scientific document. That means that a scientific style is desired. To clarify the term of "scientific style" take a look at the following rules.

  1. No personal pronouns. Do not use personal pronouns. If you absolutely need them, use first person plural ("we", "us"), even for a single author.
  2. Active. Avoid passive formulations, try to use active formulations (e.g. "This approach is able to accomplish this tasks." instead of "This can be done by this approach.").
  3. Short sentences. Use simple, short sentences since this makes it easier to read complicated associations.
  4. Be concise. Use only words which cannot be deleted from the text without replacement. Every word which is not necessary for the sentence should be deleted.
  5. Avoid idioms. Do not use familiar quotations/idioms. Just use words and terms which can be understood in every language.
  6. Avoid "one". Avoid "one" (germ.: "man") - terms (e.g.: "It is obvious that..." instead of "One can imagine that...")
  7. Avoid modal verbs. Try to use as few modal-verbs as necessary.
  8. Leitmotif. Ensure that the "leitmotif" (German: "roter Faden") in your text is obvious. Each sentence should be connected by content to the sentence after and the sentence before.
  9. One concept, one word. For the same concept, always use the same word throughout the document (no need to search for synonyms).
  10. Repeat important points. You are allowed (even encouraged) to repeat yourself, at least partially. Describe the major concepts in the abstract, the introduction, the technical part, and the conclusion. Describe the main problem from different aspects, but repeat your main points. Do make sure the readers get your main points. Therefore, you need to be clear about what your main points are, otherwise, how should the reader find out about it? [This item is also an example of repeating the same stuff over again with different words :-)]

You will see that writing a text regarding these rules is not as easy as it seems but it will result in a text which ensures formulations often seen in other scientific texts.

Layout

A ready-to-go template for latex can be found under https://svn.uni-konstanz.de/disy/General/Templates/ThesesTemplates/. Ensure that your files are UTF-8 encoded and that no errors are thrown while compilation. Refer to the common latex - documentation for information about latex itself and to BibTeX documentation for how to organise your literature in your thesis.

Please use vector-based graphics in your thesis since these graphics are scalable and therefore are looking much better in documents like (non-converted) pdfs. If you have to use pixel-based pictures, ensure to convert them into a pdf and to use them in the original resolution without any scaling.

Evaluation

The following criteria will be used when evaluating, so please make sure they are well visible from the thesis. Feel free to repeat some facts (abstract, intro, when you describe the feature, and conclusions), if appropriate (of course, you do not want to copy text verbatim, but mention the same feature):

  1. Contributions: What is novel, new, exciting? What can we learn from this (which was not known before)?
  2. If there is anything implemented: Was the implementation tricky? Were there any specific, non-obvious design decisions? Why was it done the way it is?
  3. Do you introduce any new concepts? Distinguish your concepts from those that were introduced by others.
  4. How about your analysis? Is there any? (If not, explain why; you need a good reason not to analyse your results!) What does it show? Is it what you would expect? Where some results seem strange (e.g., "Ausreisser", unexpected scaling behavior, …), explain them!
  5. What are your results?
  6. Make sure your information is well presented. Structure it readable: Have explanatory titles, helpful graphics. When writing the thesis (better yet, before you start), make a list of the top-5 items that are so great about your thesis. Make sure the reader gets to know these items at least once.