Common pitfalls in writing papers

Problems throughout:

• Overuse of the imperative. When describing an algorithm, be careful not to make it seem as if you are telling your reader what to do.
• Overuse of the passive. The phrase "this paper discusses the problem of sorting" is preferable to "the problem of sorting will be discussed."
• Overuse of lists. Although lists are easy to form in LaTeX (and HTML!), often a paragraph of text is a better way to express ideas.
• Lack of transitions. Dividing material into sections and subsections is not a substitute for transitional sentences and paragraphs.
• Arbitrary structuring of text. Think carefully about what constitutes a sentence; avoid run-on sentences created by overuse of the comma. A paragraph should be used to prove a particular point, not to divide text into arbitrary chunks.
• Lack of proofreading. Don't let the beauty of a typeset paper convince you that it is lacking any errors. Watch out for misspellings, inconsistency between subject and verb, inconsistency in use of tenses, and misuse of "it's" (which means "it is" not "something belonging to it").
• Undefined terms. Be sure that all technical terms are defined, and that the definition of a term precedes its use.
• Excessive notation. When summarizing related work, you do not have to make use of all the notation found in the original paper or papers. Often it is much clearer to develop your own notation or do without.
• Mixture of symbols and text. In the text, it is usually preferable to use English rather than symbols. For example, "we consider values > n" could be rewritten as "we consider values greater than n."
• Use of citations as nouns. The phrase "the best algorithm can be found in [N99]" could instead be written "the best algorithm is due to Nishimura [N99]."

Problems with the abstract:

• Verbatim repetition of text used elsewhere in the paper. Don't plagiarize from yourself, and don't waste your reader's time.
• Vagueness. Be specific in quoting the results you have obtained.

Problems with the introduction:

• Attempting to give a complete primer to background material. Make sure that all information present supports what follows, whether by presenting techniques that are exploited or modified in your work, or by indicating related topics not to be covered.
• Inconsistent assumptions about the level of background of the reader. It doesn't make sense to define a cycle in detail and then use the term "strongly chordal graph" without definition. Giving notation for simple concepts is fine.

Problems with presentation of the results:

• Lack of intuition and/or structure. You may know why you are presenting what you are presenting when you are presenting it, but if you do not supply the context for your reader, you will need to put in more effort in convincing them of the correctness of your work. Think of your presentation as telling a story. You need connections between subsections and a sense of direction through the material.
• Ineffective or inadequate use of figures. Although pictures can be very helpful, some of the potential use is lost if there is no discussion in the text of what the picture is intended to depict. A figure can be harmful if it replaces the precise description of an object, with the author instead relying on a reader to figure out what may be implied instead of explicitly shown.

Problems with the bibliography:

• Inconsistent use of abbreviations, such as of names of journals and names of authors.
• Non-standard ordering of references. The typical order is alphabetical by last name of the first author.
• Overreliance on LaTeX. Always check to make sure that names embedded in titles are capitalized appropriately.

Naomi Nishimura's home page