At least I had a good idea of what papers I had copies of, and in what format. I use EndNote to generate a list of papers cited within a document and keep my list of papers I’ve seen organized. I set up extra fields in EndNote just to keep track of my own filing system. One I call "label" into which I type a short description of the general topic (e.g. "geothermobarometry" or "crystallization"), which is also the name of the folder in which the paper is filed (if I have it in pdf format). Then, if I am looking for a paper on a specific topic, I can sort the full list in EndNote and look at the list of just the papers on that topic. This makes it much easier to find a specific paper than if I needed to look through the full list (344 references in the list so far!). I've also got a field named "format/where" in which I record if I have that particular paper in pdf, or I have a paper copy, or I checked it out from the library, or it is in the department thesis cabinet collection, or if I borrowed it from my advisor and have since returned it. This makes it simple to find papers or books again if I need them. However, I eventually found that knowing where to find a paper isn’t always good enough.
Two years into my project I was conversing with one of my friends about a 100-a-day skills challenge wherein the participants practice 100 repetitions of their chosen skill a day (usually physical skills, such as juggling, or dance steps, or whatever). The goal being to see how many days in a row one can manage, and see how much the skill improves whilst doing it. Thinking of that challenge I decided to apply a variant of it and challenged myself to read something from the geologic literature every day. 100 being too small a number where reading is concerned, I settled upon "1,000 words a day" (or roughly the equivalent of reading for abstracts) as my goal. And promptly set up a spreadsheet to track the days, what I read each day, and if I had any comments to make on the reading. Initially I only recorded the author name and year of publication, but after a bit of time I started also recording the title of the paper (or book) and the name of the journal in which it was published. However, other than those comments, I still wasn't really taking notes.
Sometime more recently one of my friends persuaded me to obtain a program designed to help one organize their "to-do" list. Realizing that I'd been downloading more papers "to read later" than I was actually reading, I begin a category in that program called "things to read" and started making a note there whenever I downloaded a new paper. Sure, that information was also in the spreadsheet recording my "1000 words a day", where the "comment" would say something like "downloaded a copy to read later", but I wasn’t actually looking back into the comments field of the spreadsheet. Creating a single list of everything I’ve obtained that I haven’t read yet made it easer to figure out what I should read next. Months elapsed before I realized that when I added a new paper to the list, I should also use the "notes" field to record *why* I'd downloaded it. Did I find it while doing a search for a specific subject? Was it referenced in another paper and I thought I should follow up on it? Did my advisor give it to me to read?
I'm now working on a part of my thesis wherein I need to talk about *why* I chose certain methods to do certain tasks. This means that I need to make reference to papers written by other geologists who chose similar methods, or to papers which used a different method and explain why their technique won't work for me. So now, I am actively taking notes as I look through the list of papers on the general topic, trying to find those which address the specific issue upon which I'm currently writing. As I find them, I type (or, when possible, copy-paste—how much easier a scholar’s life is in the modern age of papers published as searchable pdf files!) quotes into a file to sort out later, once I’ve finished the collection part of this minor literature search. However, in addition to putting these quotes into a single file, I am also copying them into that scheduling program in the “notes” section associated with the paper (most of which I can move to the “papers read” folder at this point), so that I still have them later, if I ever need to look such things up again.
All of these little techniques to keep track of what I’m reading has really paid off, and, so long as I manage to continue them throughout my career, will make the process of keeping current with the literature, and knowing what sources to cite when writing my own papers ever so much easier. I only wish I had thought of all of them sooner, so that I had better records from the first part of my project as well!
PS: for my 1000-words-a-day challenge, I keep track of the number of consecutive days I meet the goal and read 1000 words or more (often much more!) from the geologic literature. Occasionally I forget, and then I have to start my count over. In the past 455 days I have forgotten on eight separate occasions. My current count is 93 days in a row, which is my second-highest count yet (my record is 112 consecutive days, and my worst was only 7 days before I got distracted and forgot a day). How many of you would be able to beat my record? How many would want to thus challenge themselves?