Reading is an important part of doing research.

Part of the problem is keeping track of what you’ve read. I strongly recommend that you keep a bibliography (database) of everything you read that is research/work related.

Update: back when I wrote this, there weren’t any good technological solutions to this. That has changed a bit. As of 2020, I am using Zotero. Medeley is in some ways a better alternative, but there are issues that the company that runs it treats your data in an unacceptable manner.

There are now a bunch of tools, online and desktop, for research reading management. I recommend you look at them and choose one that is appropriate for the way you work. I am using Mendeley (although, it was recently conquered by an evil publishing empire). It’s far from perfect, but I have made it work for me.

I recommend that you keep this as a bibtex file - that way you don’t have to retype things when you need to put it into a paper.

At a bare minimum you should keep the full citation information (the more complete, the better - get page numbers, volume numbers, editor, and all those other BibTeX field that you probably don’t really care about). I also keep the following information along with each record:

  • from - what pointed me at this reference? (example: “searched google for terms ‘psychophysics’” or “referenced in paper X”)
  • return - should I return to this paper? (example: “yes” it deserves re-examination, “cite” it should be cited)
  • read - did I actually read it? (example: “yes”, “skim”)
  • annote - some commentary on it. sometimes a short summary, but usually just enough to jog my memory as to what the paper was about, and any key bits I might need from it
  • pdf - where to find it on the web? (this is a link - so maybe “html” would be a better field name)
  • proj - which project does it pertain to?
  • date - what date did I read it/enter it into the database?