Write a paper - David's checklist

David's checklist for research papers (in the order he'll be checking your draft for them):

1. the paper has the following ingredients (typically as sections)
   1. abstract
   2. introduction
   3. related work
   4. methods
   5. results
   6. discussion
   7. conclusions
2. the title is specific. It does not describe a problem area or a class of solutions (unless the class is completely represented in the paper). It captures the contributions.
3. the abstract begins with the contributions. Typically, this means the paper starts, "we present...," or "this paper presents...".
4. the abstract summarizes the conclusions (see point below about conclusions)
5. the results demonstrate the contributions. If you results do not demonstrate your contributions, then remove them. If nothing demonstrates your contributions, then your paper is probably not ready to submit. If your results involve statistics, this link may be helpful.
6. the paper ends with conclusions. If there are no conclusions, why should anyone read the paper? Conclusions are not just a summary. From this work, what should readers conclude, or what do you conclude? Often the conclusions summarize and emphasize the most important elements in the results and discussion sections.
7. there is no "future work." Readers do not really care what you plan to do in the future. If there are interesting open problems to describe, label them as such in the discussion or somewhere else.
8. the related work section is not a literature review. A related work section should explain how this work is different from other work. It is not necessary to detail all of the other work, only to describe it enough to explain the relationship to your work. Of course, if you're writing a review article, then this point does not apply.
9. the methods section gives only enough detail to reproduce the results. It is not a narrative of how you spent your last N months. It is also not the place to discuss design decisions or funny results. It should very concisely outlined the steps someone else would need to take to reproduce these results.
10. the introduction should briefly describe the problem and its significance. It may also include related work as a subsection.

What is a contribution? It is something that expands human knowledge. Examples include:

1. a new and more useful algorithm, for which the results will demonstrate the novelty and the "more useful"ness. The "more useful" part must be with respect to the best available.
2. experimental results comparing computational approaches, visualization approaches, display environments, interaction algorithms, etc.. The results are the comparisons and your analysis of those comparisons.