# How to do a literature search

How do I do a literature search of the research literature on cryptography? Assume there's some topic in cryptography I'd like to learn more about; how do I search the cryptographic research literature to find research papers that might be relevant? What's the best process to use for finding relevant research?

-
I'm hoping this may generate a good reference question that will be generally applicable, so when people have questions about how to find research papers on topic X, we can refer them to this question for a good starting point. –  D.W. May 10 '13 at 19:37
This may be made a generic question and moved to academic.se ? for greater good ? –  sashank May 11 '13 at 3:10

It depends on whether you're in university or not. If you are and your school has the for-pay journals, start by searching them for the topic. Also try 'eprint.iacr.org' and google scholar.

To find the most important or seminal papers in a topic you should look at the number of people that have cited it. This is an imperfect but useful heuristic for culling a large number of papers down to a few gems.

And don't overlook Wikipedia. If the topic has a well-fleshed-out article, the citations will be a useful starting point.

EDIT: And don't be afraid to send out a few emails. Professors work really hard on their research and they love talking about it! For example, I recently needed to learn more about order-preserving encryption. I emailed Adam O'Neill, a recently minted PhD who published several important papers on the topic. He was extremely helpful and pointed me in the direction of his and his colleagues' work that I might not have found otherwise.

-

Here is how you do a literature search, to find relevant research papers in the literature:

1. You identify some search terms related to your topic, and search for them on Google Scholar and other places (e.g., Crypto.SE, via web search, on Citeseer). (For cryptographic work, also try searching Google with site:eprint.iacr.org and your search terms, to turn up papers on the IACR's ePrint server.)

Go through each paper that you find in this way and very quickly skim the title and abstract of each paper. Keep all the papers that look relevant or related to your topic of interest, based upon your skim of the title and abstract.

2. Take the papers you identified in step one. If there are any research papers that you know are related, add them to the list. This is your starting set of relevant papers.

3. For each relevant paper, you use that paper to help you find other papers that may be relevant:

• You read the paper's related work section, to see how it describes related work. Anything that sounds relevant to your interests, you take that citation and add it to your list of relevant papers.

• You search for that paper on Google Scholar and on CiteSeer. You use them to generate the list of all other (newer) papers that cite that paper, and you look through those results. Keep all the ones that look relevant, and add them to your list of relevant papers.

As you discover new relevant papers, apply the same process to each of them recursively.

4. Keep doing this until you don't find anything new (i.e., until the set of papers is closed).

The result of this should be a (possibly large) set of relevant papers. Now go read all of those papers, to understand what they are doing.

Once you have some research papers that look like they might be relevant, here is how you read go about reading them:

1. Start by reading the abstract and introduction; that should give you a good overview of what the paper's contributions are. Don't try to read the rest of the paper. This should take maybe 15-30 minutes.

2. If the paper still sounds relevant, you read the paper at least once straight through (skip or skim all proofs or other technical parts that take hard thinking to follow); your goal is to understand the main ideas in the paper, but you don't need to grok every little detail. That should give you an idea of what the paper is doing. Normally, this should take an hour or two.

3. If it still sounds useful, you can then do a careful read-through where you try to understand every detail in the paper, and think critically about why they're doing what they're doing and how it's helpful to you.

-