Short notes (or summary) taken while watching lecture How to Write a Great Research Paper given by Professor Simon Peyton Jones (Microsoft Research).

Primary, author giving 7 suggestions about writing a great research paper.

1. Don’t wait: write

Different models of writing research paper exists. One of the most command one is Idea -> Research -> Write Paper. Instead of model mentioned, the following one is more appropriate Idea -> Write Paper -> Research. Because writing papers is a primary mechanisms for doing research (not just for reporting it). Try to put any idea on the paper (take a note) and if its' possible - discuss or present it to other people.

2. Identify your key idea

The paper is a mechanism of conveying your idea from your head to other heads, not the tool to get an promotion. Your paper should have just one “ping”: one clear, sharp idea. After reading your paper, the reader should easily figure out what was the main idea behind.

3. Tell a story

Imagine yourself explaining your work in front of the white board. Narrative flow: (a) Here is the problem; (b) It’s an interesting problem; (c) It’s an unsolved problem; (d) Here is my idea; (e) My idea works (details, data); (f) Compare to other people's approaches. State your contribution and refer to the work done.

Avoid using following phrase that usually exists in most papers (however, no one is reading it) “paper is structured as follows …“. Instead, use forward references from the narrative in the introduction section.

5. Related work: later

Don’t put related works at the beginning of the paper, instead place in just before the conclusion. Presenting related works before the main idea of the paper is just tracking the use attention from your own idea. Despite the primary goal of the related work section is to drop some sunlight on to the domain that you are working in, it’s mostly confusing for non-experts readers, because of limited space for the describing the vast amount of information. Given credits to other authors, respect them.

5. Nail your contribution

Make a clear list with provided contributions in the paper. While describing the contribution use referencing, so the reader can simple jump to the section with detailed explanation of contribution (experiment, data, etc. ).

6. Put your readers first

Explain the intuition first, instead of starting from the generalized approach.

Introduce the problem, and your ideas, using examples and only then present the general case. Chose the most direct route to the idea, instead of showing all the steps and wrong deciding that didn’t solve the problem.

7. Listen to your readers

Get your paper read by as many friendly guinea pigs as possible. Each reader can only read your paper for the first time once! So use them carefully. The key points, which you want to get from your readers, are addressed by the following phrase “I want to know, where you got lost during the reading”. The process of giving someone reading your paper is not dedicated to make them feel stupid or learning something form you, rather it’s author’s learning process from reader. The process that will help to convey author’s research idea to other in most understandable way.