Writing Tips

First, read about writing

Start your career in writing by reading about writing:

Step 1: The Elements of Style by W.Strunk and E.White. You can find an older edition online for free here or here. It is worth buying this very inexpensive and useful book. Read it at least once a year.

Step 2: Read another style book, such as The Reader Over Your Shoulder (Graves and Hodge) or Bugs in Writing (L.Dupre). See Writing, Briefly: a neat essay on writing by essayist Paul Graham.

Step 3: Many professors have webpages on writing technical papers, or have written papers on the subject:

My advisor James Kurose's page on writing
Oded Goldreich's How to Write a Paper.
Henning Schulzrinne's page on on writing technical articles. (Henning and I were Jim's grad students at about the same time in UMass).
R.Levin and D.D.Redell. An evaluation of the ninth SOSP submissions -or- How (and how not) to write a good systems paper. ACM SIGOPS Operating Systems Review 17(3):35-40 (July, 1983).
R.E.Johnson et al. How to Get a Paper Accepted at OOPSLA
See the huge list of links to writing resources maintained by Prof.Honavar in Iowa State's CS Department.

Step 4: You're now ready for a more comprehensive resource, the book Writing for Computer Science by Justin Zobel.

The tools

Learn Latex. Latex is a typesetting language for science and engineering writing, based on Donald Knuth's Tex typesetting language. Because it's not WYSIWYG, it can initially appear daunting. However, there are innumerable on-line guides and quite a few books on the topic. Many people have created Latex templates that make it easy for you to insert content and modify only as needed.

Figures. You will need to learn how to create drawings and then include them in your latex paper. There are many, many tools for drawing. Examples include: xfig, idraw, StarOffice (Unix), Word/Powerpoint (Windows). Similarly, you can create graphs using a variety of tools. Examples include: gnuplot (Unix), Excel (Windows).

Beyond Latex. Aside from Latex formatting details, you'll inevitably need to learn how to include figures and generate PDF using pdflatex. You can include PDF figures as well as JPEG images.

Getting started

Step 1: Create a skeleton of sections in your Latex file, for example: Abstract, Introduction, Related Work, System Description, Analysis, Experimental Results, Discussion, Conclusions and Future Work.

Step 2: Think through what you want to say in each section. You can either jot down notes on paper or write them in as Latex comments in the file itself.

Step 3: Fill in some technical sections. If you have proofs, diagrams, detailed system descriptions, then these are easy to write independent of the flow of the paper.

Step 4: Write as much of the Related Work section as you can and add the references as carefully as you can. The best time to worry about the accuracy of a reference is the first time you type it in. You may have to come back to the Related Work section after you start the Introduction, because inevitably, you'll learn something about your own work and how to describe your contributions after you write some of the other sections.

Step 5:. Write the Introduction. Here are some useful tips in writing an Introduction. Although it's tempting to say "I'll write the technical stuff first and worry about fluff like the Intro and Abstract later", I prefer writing the Introduction first. My belief is, if you can't write a tight, solid Introduction, you don't have a paper.

Step 6: After the Introduction is done, the paper should pretty much write itself.

Step 7: This may be a good time to write the abstract, after every other section has been written. You should have a complete first draft now.

Step 8: Read through your draft carefully at the sentence level. That is, look for typos, awkward sentences, grammatical problems, stylistic problems and punctuation problems. At this stage, since you are "too close to the text" you might have difficulty seeing bigger problems with the writing.

Step 9: Put aside the draft for a few days, or enough time to forget some of the details. Then, pick a quiet moment when you are fresh and read through the draft with a skeptic's eye. This is when you might notice bigger problems such as: unsupported claims, weak arguments, awkward flow between paragraphs, too much text on some points or too little on others, and that most fundamental problem of all, making the central point of the paper apparent.

Step 10: Revise and edit the paper after your fresh read. Go over it again for typos and small fixes.

Step 11: Ask someone else to read and comment.

Common problems

Over the years I've seen quite a few samples of student writing. Here are some typical areas where students need help:

Sentence flow. Sentences are not stacked or merely juxtaposed. Instead, they need to flow from one to another. Did you notice the use of "Instead" in the previous sentence?

Passive voice. This is the passive voice: "The problem formulation is given in Section 3". Instead, write in the active voice: "Section 3 formally describes the problem" or "We formulate the problem in Section 3".

Understanding how to state the problem and contributions informally. This is the type of writing students find the hardest. I find it hard.