How do you approach literature review efficiently?

Question

I'm curious what strategies people use when conducting a literature review, especially in a subfield they are not already familiar with. My personal preference is as follows: Do blind key-work searches (or ask someone) until I have some clue what extremely specific term I need to search to find anything, and then click around until I find a paper that seems seminal(ish) in terms of citation count. Then, follow the cites from that paper forward until I've hit a critical mass of maybe 5-10 papers that are at least somewhat related. At this point, however, I ditch google scholar and instead start going by author- for each of those papers, I look up that specific author to find other related work, and in particular, pay attention to co-authors. In each paper, I check the citations in the lit review and and search those papers/authors. This process eventually winds down at some point when you don't find any new authors.

However, I have no idea if this is the *right* way to do a lit review. What strategies do you use when doing a lit review? Is there any software or site you find particularly handy?

Answer 1

I do pretty much what you do. But don't get too hung up on the literature, on finding out what others have done. Doing something yourself, even if it turns out you have been anticipated, is more fun and gives you a better understanding of the problem, and can be more effective in leading you to other important discoveries. A wonderful example is Feynman's reaction to reading the manuscript of Watson's "The Double Helix" (quoted from James Gleick, "Genius", p.387):
  "At one point Goodstein remarked: 'You know, it's amazing that Watson made this great discovery even though he was so out of touch with what everyone in his field was doing.'
    Feynman held up the paper he had been writing on. Amid scribbling and embellishments, he had inscribed one word: DISREGARD."

Comments

I agree completely about not reading too much before writing something down. I'm mostly worried at the moment about ensuring credit is given where it is due, and making sure I can explain my paper's position in the literature.