This is a good question. Obviously it becomes easier the more you understand the subject area. As student you might have little understanding. That's why you

This is a good question. Obviously it becomes easier the more you understand the subject area. As student you might have little understanding. That's why you are studying! In my case, initial downloads were readings for my courses or papers referenced in handouts. I just let my curiousity guide me from there. But I'll admit I also had some background knowledge to begin with, gleaned from work experience. In some (most) scientific discplines, every paper begins by citing background material. You should gather and save those papers. No matter how old. When you read the more general papers, and the other background material, in chronological order, you are in a better position to understand more specific, recent work. Basically you are building a foundation for understanding the current state of research in a subject area. At least you will always be able to get up to speed more easily by having the proper foundation. Of course, when you're at university you may be busy with other pursuits and you may not have the time nor inclination to read all those older papers. This is what you download and save. Sometimes there are scientific "encyclopedias" or other compilations. These can be good in their own right and also a further source for references pointing to more worthy downloads. When in doubt a general rule is to download papers that are highly cited. There is a vast amount of general knowledge that will never be offered on Wikipedia.[1] It is the best kind: peer-reviewed. Wikipedia and clones (Knol, etc.) wanted or tried to become "peer-reviewed" sources but this idea has never really succceeded on the open web. That is why this issue with academic publishing will continue to simmer. 1. At least, not until academic publishing undergos some changes.