Some Hints on Writing a Paper:
Dr. Bill Young
Below are hints for writing a research paper. I'll add others as I
think of them. Please take these into consideration as you edit your
paper for the class.
- Assume that you are writing for a technically literate audience,
but one that knows next to nothing about the specific area
you're covering. In particular, if you mention acronyms, specific
jargon or terms of art, or groups relevant to the area, define them
the first time you use them.
- Especially if there are several authors, papers often read as if
they are cobbled together with no thought to the overall flow.
Assume the reader will read the paper from start to finish. Don't
put information in section 1 that depends on the reader being aware
of something in section 5. But also don't repeat the same
information in sections 1 and 5. The paper should be a seemless and
- Start with a section-level outline of the paper. That will keep
you from wandering all over the place. Into each section, put only
what's relevant to that section. If you can't think of what sections
you need in the paper, you are not ready to write the paper.
- Divide your paper into sections with section headers. Use
subsections, if appropriate. Unless your paper is very long, you
probably don't need headers below the subsection level.
- Have an Introduction section at the start that describes the
problem you are addressing. The Introduction should conclude with a
paragraph describing what is in the other sections of the
- End with a Conclusions section. It should summarize your paper
and state what you discovered.
- Don't have Introduction and Conclusion subsections within
sections; that is overkill.
- If you are not a native speaker of English, have a friend who is
edit your paper for you.
- Try reading the paper aloud to see if your sentences are really
sentences, make sense and read fluidly. Given some of the bizarre
non-sentences I find, I can't believe some students actually ever
read the papers they write.
- Don't be too informal. "Companies need to get off their butts
and improve security" may be OK when talking to a friend, but it
doesn't belong in a research paper.
- On the other hand, avoid excessively flowery or pretentious
writing. Simple, direct prose is best: more Ernest Hemingway and
less Edward Bulwer-Lytton.
- If you're tempted to use "said" as an adjective ("said persons"
or "said document"), rethink it. That is legalese that has almost
no place in standard speech.
- Avoid being overly subjective and personal. Don't say something
like: "I believe that what I've learned about X goes hand in hand
with what we've been learning in class ..." Just give the evidence
for the point you're making.
- It's probably best to avoid first or second person. Use third
person unless there's a good reason to do otherwise.
- Avoid long and complex sentences. Don't use words that you don't
fully understand to make yourself sound smart.
- It's usually considered bad form to have a paragraph that is a
- If you have two distinct thoughts in a sentence, consider making
it two sentences. In no case should you just separate them with a
comma; that's called "comma splicing." Use a semicolon instead, but
only if they are closely related. Example: "This is one thought; this
- Use a spell-checker.
- "It's" is a contraction for "it is." If you can't replace a use
of "it's" by "it is," you meant "its."
- Any time you use the word "which," ask yourself if "that" would
work as well. If so, use "that" instead. Consider using your editor
to do a "which hunt" and substitution, but be careful. At times,
"which" is the appropriate word.
- In written American English, quotation marks always follow commas
and periods. They may fall inside other punctuation. Example,
"foo," "bar," and "baz." Did she say "braf"?
- Whenever you have a list of things, make sure that the items in the
list are parallel. That is, all items should be of the same type
(nouns, phrases, sentences, etc.).
- If you are tempted to say "as I told you above" or something
similar, rethink it. If you said it before, you probably don't need
to say it again, particularly in a short paper. Wording like that is
appropriate if the reference is distant in the paper. In that
case, say something like, "As described in section 2.3, ..."
- Make sure you are not plagiarizing. Any significant fact
that you got from another source should have a reference. Every
direct quotation should have a reference, unless you have a bunch of
them in a row from the same source. If you copied a sentence verbatim
from Wikipedia or another source, you are plagiarizing.
You must credit the source and rewrite it into your own words
(unless it's a direct quotation).
- Make sure your references and bibliography follow a style
accepted for the field. Don't trust your text formatter to get them
right. Look them over.
- Use figures, pictures, tables and graphs where appropriate. A
long prose description of a complicated algorithm is boring and much
harder to follow than a graph.
- Don't pad your paper with excessive spacing, big margins,
extraneous figures, etc. It's better to have a short paper that is
substantive than a long paper full of fluff.
- Long quotes should be set off from the body of the text. The
rule of thumb is that any quote shorter than four lines can be
in-lined. Longer quotes should be block indented.
- There are a variety of acceptable styles for citations used
within technical papers in CS: ACM, IEEE, APA. There's a useful
guide. The most important thing is to decide on one and follow
it consistently throughout your paper.