20
$\begingroup$

Apologies if this is too broad a question for this forum, but I'm interested in specific tactics and tips that researchers (in TCS) use to write the introduction of a research paper.

$\endgroup$
  • 2
    $\begingroup$ Conference or journal ? Boaz Barak's advice for FOCS authors (and Impagliazzo's rule) apply well even to the intro. More generally, you might find my answer on Academia useful: academia.stackexchange.com/questions/15276/… $\endgroup$ – Suresh Venkat Apr 13 '14 at 23:38
  • $\begingroup$ I was looking for answers relating to conference submission. I've read Boaz Barak's advice and it's really helpful. Same goes for the link you posted. I was wondering if there's more tactical advice available on this. Things like write bullet points first and expand later. I know this is can be very personal but I'm just curious :) $\endgroup$ – Madhav Jha Apr 13 '14 at 23:54
  • 2
    $\begingroup$ An alternate possibility might be to ask for examples of introductions that people really like, together with an explanation of why. This avoids the problem of too much subjectivity. $\endgroup$ – Suresh Venkat Apr 14 '14 at 1:10
  • 1
    $\begingroup$ Welcome, Madhav! $\endgroup$ – arnab Apr 14 '14 at 8:48
17
$\begingroup$

Simon Peyton Jones has an excellent web page devoted not only to advice on writing introductions, but whole papers, and there is a cool video as well. On page 18 of his slides he says that the purpose of an introduction is to:

  1. Describe the problem.
  2. State your contributions.

He then goes on for a while to explain what precisely that means in pracrtice. But there is much more to his talk, I highly recommend it.

My personal piece of advice: never ever start your paper by saying "In the recent years the foo technology has become important [1, 2, 3, 7, 10, 12, 13, 21] for saving the world peace, taking photos of kittens, and obtaining academic promotions. But all the people who did foo before us missed a minute point, which we are going to talk about at length in this paper. But first, let us bore you with a review of all the foos out there." Just get to the point, please.

| cite | improve this answer | |
$\endgroup$
  • 1
    $\begingroup$ I like everything he has to say. But the typical theory (A) paper puts the related work early rather than late. While I see no problem in putting it late, it does go against the prevailing norms. $\endgroup$ – Suresh Venkat Apr 14 '14 at 6:38
  • 8
    $\begingroup$ But remember that, if you include photos of kittens that were taken using foo technology, nobody will care about how boring your review is and the paper is guaranteed to be accepted. At least, no conference has ever rejected a paper of mine that contained photos of kittens. $\endgroup$ – David Richerby Apr 14 '14 at 7:33
  • 7
    $\begingroup$ Now that you mention it, all papers of mine that included photos of kittens were accepted, too. $\endgroup$ – Andrej Bauer Apr 14 '14 at 9:55
  • 4
    $\begingroup$ When I see a new paper, I usually scan it for three things: the problem, the contribution, and the context (which usually includes related work). Without context, the contribution feels like a piece of trivia, making me less likely to proceed to reading the paper. $\endgroup$ – Jouni Sirén Apr 14 '14 at 23:49
  • 1
    $\begingroup$ @Thomas: that assumes the formal statement can be made intelligible without first introducing any concepts or explaining things. There are many areas in which there is just no way you can do that. A better, more general advice is: in the intro state your contribution clearly, with forward reference to the exact place where that contribution has materialized. So, rather than demanding a formal statement in the introduction, you can have a reference to it instead. $\endgroup$ – Andrej Bauer Apr 26 '14 at 22:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.