3

I have problems in reading papers, specifically papers where the subject is not so new to me and I am familiar with the problem which the writer wants to solve.

How to find the those parts which talk about the main idea?

When I go to their suggested work to understand the main idea, I have to have first read their previous sections, because they define some notations and definition which they use them on their suggested work. When I go to read those parts to understand the notation. I am not sure if my reading method is good. because after reading the notations and definitions , I have to start their suggested method section. This one has its own challenges too.

My advisor said you should understand the main idea of a paper in one hour, or at maximum in two hours. But I need at least 5 hours.

(Maybe help: I am reading a paper in the field of computer networks.)

M R R
  • 3,389
  • 7
  • 25
  • 40
  • 3
    I would say there is no SOP (standard operating procedure) for reading a paper... – Yes Jan 18 '15 at 06:13
  • What are you trying to get out of this paper that you are reading? The answer depends on this... – jakebeal Jan 18 '15 at 06:18
  • I want to know just the main idea. Not understand all parts like reviewer or a person which wants to work on that paper. – M R R Jan 18 '15 at 06:26
  • This answer to a related question gives more detailed advice on reading papers. Beyond that, I agree that there is no standard way to read a paper. – ff524 Jan 18 '15 at 06:39
  • 2
    Have you considered the possibility that maybe the language barrier is what's slowing you down? For example, I tried to correct the English in your post, but I can't understand the part from "I tell my self is work ok?" to the end of that paragraph. – ff524 Jan 18 '15 at 07:11
  • @ff524, maybe. but not generally. my problem is "How to detect those parts which give me a point of view about their suggested algorithm as fast as possible – M R R Jan 18 '15 at 07:18
  • 2
    You skim the paper and find the parts you're interested in. It does sound like a language issue... skimming in a language you're not very comfortable in is bound to be difficult. (And the rest of that paragraph - "I should start to learn their suggested method. and also that one still have its challenges." - is still not understandable.) – ff524 Jan 18 '15 at 07:20
  • But I need at least 5 hours. So? Even 5 days is worth it if the paper is good. – Nobody Jan 18 '15 at 07:52

1 Answers1

8

How to efficiently read a paper is not something that can be taught over the internet in broad generalities. It seems pretty futile to try to teach it "in general" in general (!!), by which I mean that if you're trying to teach someone how to read a paper, you had better have at least one actual paper in hand. You can try to draw broader morals, but whatever lessons are there to be learned should be grounded in the reading of that particular paper. Thus for instance

How to find the those parts which talk about the main idea?

would be silly to address without seeing the paper.

I find more promise here:

My advisor said you should understand the main idea of a paper in one hour, or at maximum in two hours. But I need at least 5 hours.

Your advisor knows me?!? Sorry, just kidding. Anyway, that's exactly whom you should be talking to. But even with your advisor you are stuck on a way too general discussion. Go to her with with paper you are trying to read now, and learn how the advice applies to that paper.

Let me also say that I am a bit skeptical of the hard limits your advisor is imposing, or more precisely with the way you are construing whatever your advisor told you. Obviously 1-2 hours cannot be the amount of time needed to understand the main idea of every paper: some papers are just significantly longer, more difficult and/or more obscure than others. Speaking for myself: after some years of experience for reading for the main idea, depending on the paper, sometimes I can now grasp the main idea just as soon as I run my eyes over the first page; sometimes it takes me an hour or two; sometimes it takes me a full day's work; sometimes at the end of a full day's work I find that I don't understand the main idea.

Where is this 1-2 hour figure coming from, and what does "should" mean here? Maybe she means "Come back if you've spent more than two hours on it and I'll give you more help." Or maybe she means that's an upper bound for how long it would take her to read the paper. But she's not a student, she's supervising one. If there's one matter in which students should be skeptical of their advisors, it's in their advisors' estimations of how long it will (or "should") take a student to do something. It's not a matter of arguing with your advisor but rather letting her in to understand your process. If it's taking you so much more time than she expects, show her what you're doing and see what she has to say.

Let me also say that a longer time spent on something isn't inherently problematic. If you are reading in a language different from your own native one, it will take you longer -- and skimming will take you much longer. Most academics eventually realize that the people who can complete tasks more quickly are often but far from always the ones who can do the best work on those tasks. Real breakthroughs are things that people chew on and work through for weeks, months or years. Some speed freak types simply don't have the patience or focus to linger on something for much longer than it takes for others to see how quick they are. So if you are eventually understanding what you're reading quite well and you're not lagging far behind in other tasks: ask yourself whether you really have a problem or just a certain skill that you'd like to optimize when you get the chance.

Pete L. Clark
  • 130,631
  • 25
  • 361
  • 522