About me
For students
How to write a computer science paper
Writing tips
Short intro to JBoss and EJB
Short introduction to Subversion
Getting started with LaTeX
LaTeX notes
XNote
Software notes

Tips for scientific writing

This page shortly lists some tips to improve the writing of master's theses, PhD theses, and scientific articles. It further lists frequent mistakes, especially as made by German speakers, to allow you to avoid them. If you think that something is not correct or you would like to add something, especially if you are a native speaker, please send me an e-mail.

  • Avoid standard headings: It is better to write "Overview of messaging middleware" instead of "State of the art", for example.
  • Don't use "that" if it can be avoided. Often, people write something like "... the text that is displayed on the device ..." → This can be improved to "... the text displayed on the device ..."
  • Use standard vocabulary ("Fachausdrücke") – and use it consistently. This contributes to the understandability of your text.
  • It is OK to repeat words in English texts, especially standard vocabulary. This is contradictory to what you might have learned for German texts in school. Therefore, use the standard terminology of your community and don't try to find several different words with the same meaning.
  • Figures and equations have to be referenced and explained in the text.
  • Abbreviations have to be written out the first time they are used.
  • Carefully consider whether sentences should start with "But" or "And" – see grammar.ccc.commnet.edu/grammar/conjunctions.htm.
  • References: Better less but with high quality, e.g., scientific papers published by IEEE, ACM, Springer, IFIP or standards. Avoid Web links – they are volatile sources and the details may no longer be accessible after a few months. Wikipedia, although useful in practice, is no reference for scientific publications. Here you should refer to the core publications of experts in the field. Moreover, refer to the initial publication and not to a publication that refers to a publication that ... that refers to the initial publication. You can find further details in the IEEE Computer Society Style Guidelines for References.
  • Before you submit a paper to a scientific conference, journal or workshop, let your colleagues review the paper and discuss different points of view. This substantially improves your work.
  • Stay technical: Concentrate on facts and outcome.
  • No story telling: Don't write sentences like "We identified that...", "... issues arose ...", or "... after spending several hours and consuming much coffee ...". (Well, the last phrase should be obvious not to belong into a scientific text.)
  • Replace "which" with "that" if appropriate – see grammar.ccc.commnet.edu/grammar/notorious/that.htm
  • Don't use the "-ing" form too often – use "to" plus base form of substantives instead, e.g., "... explain the user how the registering works." → "... explain the user the registration process."
  • Use "such as" for lists of examples and "like" to point out similarities.
  • Do not use "Note that..." too often. Just say what you want to say.
  • Replace: "a lot" → "many"; "a bit" → "slightly", "at last" (different meaning) → "finally" (especially as introductory element).
  • Refrain from the use of "very", "great", etc. This is often used in combination with a subjective expression or an opinion. Moreover, you might be imprecise in stating your achievement, for example if your solution is able to "achieve a very great performance improvement", this does not tell anything specific. Is it 10% or 50% or 200% improvement compared to the best solution in the state of the art?
  • Use a comma after introductory elements such as "Therefore," "However," "Finally," – see grammar.ccc.commnet.edu/grammar/commas_intro.htm and owl.english.purdue.edu/handouts/grammar/g_commaint.html
  • Be definitive and precise: Write "The reason is this and that ..." and not "The reason could be ...", for example. This shows that you thoroughly investigated an issue and know the details for sure. Of course, this also has to be the case. Consequently, if you want to write "... reason could be ..", then your research about the topic is most probably not complete.
  • Name your topics: For example, don't write "... the thing is ..." → use "... one influential disturbance factor to consider is ...".
  • Cross-references such as section, figure, table, etc. are usually written with the first letter capitalized if combined with a number, e.g., "Section 2.2".
  • Footnotes: As recommended by the IEEE author guidelines, "use footnotes sparingly (or not at all!)" as they reduce readability due to "jumping" between text and footnotes. My general recommendation is: If the footnote contains valuable content, integrate it in the main text. Otherwise, check whether your text would "lose something", if you removed the footnote at all. In most of the cases, you can avoid footnotes.