Editing for technical papers

A few writing rules

I read a lot of papers, especially ones written by my students. Here are some mistakes to avoid. They are strictly mechanical, so there's no excuse for having them in a paper, unless you believe that your time is more valuable than that of the person reading.

• Small numbers, between zero and ten, should be written out as words, not in digits.
• Use "e.g." and "i.e." (note the periods) only within parentheses; otherwise use "for example" and "that is."
• Possessive pronouns like "his," "her," "theirs," and "its" do not have apostrophes.
• Words should be spelled correctly and consistently (e.g., don't use "Chebyshev" and "Tschebyshev" in the same paper).
• Mathematical symbols in text should be italic in general; look at the LaTeX book by Lamport for examples that clearly illustrate this. Avoid using symbols at the start or end of a sentence whenever possible.
• References are meta-text, not nouns. If you want to say that Smith and Jones did some work that's described in reference number 4, write "By contrast, Smith and Jones [4] showed that metamodels are transiently subcubital" or "By contrast, Smith and Jones showed that metamodels are transiently subcubital [4]" but do not write "By contrast, [4] shows that metamodels are transiently subcubital."
• Use "that" in place of "which" in any sentence in which it makes sense when you try the substitution. This sentence, which is contrived, is an example of one where the switch is inappapriate. But a sentence which has no subordinate clauses probably needs a "that" instead of a "which." (The last sentence *should* have its "which" replaced.)

Here are some things to think about before you even start writing at all:

• Decide for whom you're writing this. I recommend that most papers written by my students be aimed at a beginning Master's student in CS/graphics.
• Decide on your goals for the paper. What are the main things that you want to communicate? All papers involve PERSUASION; decide on your persuasive strategy as early as possible. If you find something that doesn't fit in, perhaps it belongs in another paper.
• Write your introduction carefully: it should give the reader a promise of what s/he will get by reading your paper, and a reason to do so.
• Write your "related work" section carefully: it's your best chance to get your paper rejected. If you unfairly criticize others' work, or mis-represent it, the laws of fate will dictate that those you've most maligned will review your paper and hate it. Instead, point out the important accomplishments of former papers, and how your work builds on those foundations. You can also point out what they left unaddressed (easy to find in their "future work" sections!) and describe it as having provided motivation for your work. And it's fine to say that you do something differently from the way someone else did it and then compare; when you do so, try to find a case where their method works better than yours, or at least a reason why someone might prefer their method. These go alongside your reasons why people might prefer yours, of course.
• You learned in high-school to "tell 'em what you're gonna say, say it, then tell 'em what you said." Authors usually take this to mean that you should write an Introduction, a Body, and a Conclusion. I strongly believe that for most papers in computer graphics, the conclusion is an admission of defeat as an author. It says "you started reading these eight pages just 15 minutes ago, but I wrote them so badly that even now you've forgotten what I told you, so I'll have to say it again." I therefore advise against writing a conclusion unless your other writing is horrid.
• There are a couple of rules of thumb about writing: one is that any magazine article can be shortened 20% without anyone, including the author, noticing it; another is that the first page of most magazine articles can be deleted as a way of improving them. Keep these thoughts in mind as you write.
• Use your word processing tool to help you. If you can't figure out how to say something, put it into a different font, into a box, into an indented section, whatever...and write it informally. Just get the ideas down on the paper in some order that lets you play with them. Wordsmithing should happen relatively late in the writing process.