Fred Brooks recommends this system for organizing research papers into a queue for processing.
Rank each paper according to two categories:
Usually an intuitive ranking of the paper can be reached by simply reading the abstract or, lacking that, skimming the first few paragraphs.
Fred recommended only 4 levels within each category:
This yields a convenient two-axis sort. Ideally, you would read the 1A's first, and then you can decide how to proceed, prioritizing either relevance or importance.
One advantage of this approach is that you don't have to spend time adding tags or categorizing papers. This yields a single, flat list of all papers that can be sorted using importance and relevance. This saves time spent organizing for reading, processing, and writing.
One caveat is that as your research questions and interests change, Relevance may also change. You may need to re-rank the relevance of the papers occasionally, but in practice this seems to happen fairly rarely.