Recent Changes - Search:

My Research


Advice / PaperRecipe

Let me tell you a little trick that I tell all the students:

sentence, paragraph, page

Mike's paper recipe:

  1. come up with 1 sentence that tells the "story" of the paper - the main message. you need to make this into one sentence. if the sentence is a long, run-on sentence (or is a paragraph, not a sentence), you don't have a focused enough message.
Keep the sentence in view at all times (print it out in big letters and put it next to your computer as you work!
  1. write the 1st paragraph of the paper - how you will hook the reader in, introduce the problem, ...
  2. write a 1 page outline of the paper
in each part, keep in mind that your goal is to tell the "story" of the paper - which is why the sentence is so important.

Update (3/6/2015)...
Here are a few other questions you should try to answer while developing those three things:

  1. Our key idea is...
  2. Our primary contribution is ...
  3. The contributions of this paper are: (list)

The first two of those should be sentences. Not run-ons. Not lists. Your "primary" or "key" thing must be a thing, not 3.

Here's a nice article on Tomorrow's Professor about writing. They echo my feelings about the importance of the "sentence." Some of their other advice, I don't implement myself - but wish I did!

Addition (9/25/2018) - sent from Mike to Manfred, Sept 2018 ...
Here's another set of questions that is often useful to think about: The 5 questions:

  1. What is the PROBLEM
  2. Why is it interesting
  3. Why is it hard
  4. Why do other approaches fail
  5. How is our approach different

The new paper recipe - New thing, March 2019

This is based on how the intro needs to "work".

The intro has to do 4 things. I think there should be a single sentence that answers each, which makes it easy for a reader to identify/get a summary of these things. Some intros are less explicit, but I think all good intros adress these points.

1. What is the problem? 2. What is the solution? (what is this paper) 3. What is the key idea? (why is this a paper) 4. So What?

1. What is the problem?

This usually gets set up by saying what happens now, and then gets to the "However". The first part is something the reader can connect to (the world they know). Then you connect it to the new stuff by defining the problem. This common pattern is discussed in the Williams and Bissup Style book.

Key sentence 1: However, ... (we have a problem)

2. What is the solution?

What does this paper give the reader? Hopefully its an answer to the question/problem set up in 1:

Key sentence 2: In this paper we ... (give a solution)

3. What is the key idea / takeaway?

At some point, you need to say why it was worth writing a paper about this. An explicit contribution statement can do this. But I think its even better if you can identify an intellectual core from which the paper will build. What made it possible for you to solve this problem (where no one else has before).

Key sentence 3: Our key idea is ... (give an intellectual core)

4. So What?

Why should the reader care? What actual progress have you made at addressing the solution? Why is the world a better place with the ideas in your paper in it?

I don't have a formulaic sentence for this. It can take many forms. One common pattern is: Our results/system/method/technique/ideas enables/provide/allows ...


I don't think that every paper has to have the formulaic sentences as explicit sentences in simple form. But I do think that writing these 4 sentences early in the writing process is valuable. If you can't express these 4 functions concisely, you will have problems building a coherent paper.

Other Resources

Seth Hutchinson, IEEE Trans Robotics, on how to think about revising papers in response to reviews:

History - Print - Recent Changes - Search
Page last modified on June 18, 2019, at 04:47 PM