The first is the beginners system, and has almost nothing to recommend it. It results in a drawer full of pieces of paper ranging from slightly used napkins, to old computer printouts. Alas, one must resort to it from time to time. But care should be taken to regularly transform such scraps to the better format that one is using.
A notebook seems like a good idea, but when your interests broaden it becomes difficult to categorize things. The real problem with this system is retrieving the information. This worked fine for me for several years, but when I had several notebooks full of references it became unworkable. The only way to retrieve information was to page through the notebooks.
The use of notecards is a time honored system and works well. Ken May has a detailed discussion of how this should be done in his Research Manual. I suggest you read it carefully. One suggestion he makes is that you do not use index cards as they are too thick and it is difficult to riffle through them. Slips of paper are far superior. I keep several of these pads at hand to record references. The pads are easy to handle in the library. When you want to organize the information you can tear the slips off the pads and manipulate them any way you wish.
Most recently I have been using a data base called EndNote. It is a very handy program. The main advantage of using such a program is that one can very easily search for a given word. The disadvantage of this system is that you are tied to a computer. Unless you have a portable to take with you to the library this system can require excessive keyboarding time. I find that when I have my laptop with me that I take much more careful notes in the library (and also Xerox less stuff). A problem with this record keeping method is that if you have several computers then you need to take care that they contain the same information in the electronic file. This system also works wonderfully when you are collecting references from some sort of database off the web.
One place where EndNote is especially nice is that when a student asks about references for their paper, I can quickly sort out those of interest. It is best if the student is with me, for then I can discuss the advantages of the various papers. But I can also print out all of the references on a given topic.
A final piece of advice: Whenever you photocopy anything be sure to immediately write the precise reference on it including the title of the work (I have started just copying the title page of the journal or book) and the call number of the item and the name of the library in which you are working. Failure to do this can result in real headaches.