Literature management

"A single book at the right time can change our views dramatically, give a quantum boost to our knowledge, help us construct a whole new outlook on the world and our life. Isn't it odd that we don't seek those experiences more systematically?" [Steve Leveen]

Rich, Kutti, and Kayesdee discussed recently their literature management systems. I personally use also JabRef in combination with DOI/PMID identifiers and PDF copies. For reading I prefer always paper copies, because I use text markers. After reading and extracting relevant information I keep only articles, which are not available online. Non-available articles are usually too old for having a DOI identifier, so I keep them in alphabetical order using my own article keys.
Single author: FirstThreeLettersOfLastName_Year[a,b,c for multiple occurences]
Multiple authors: FirstLetterOfEachAuthor_Year[a,b,c for multiple occurences]

This system is ways better then sorting by categories, because I was often confused when articles had multiple categories. But how do I store then the categories? Simply add your categories to the BibTeX keyword section in JabRef or create your own document with multiple sections/categories.

Anyway, my personal experience is that my keywords and categories are mainly meaningful for me, but maybe completely useless to others, no? So, I would recommend to use whatever you like. Just ensure that your system has a good long-term stability and that you are able to exchange literature references with others (and yourself, when working on multiple machines). This exactly is the main reason why I heavily recommend to use BibTeX and unique article identifiers. Those will be still valid in ten years and I can export and import those things in almost any text processing program. Why? JabRef supports multiple export filters, which are simple to modify, because they are based on text-files. This means that special export function for text-based outputs, like CSV, HTML, and XML are simple to create.

For a fulltext-search I can recommend Docco with all (mandatory) plug-ins. Although the full-text search for JabRef is already registered in the feature-tracking system (id=1674076), it might take a while till this lucene feature is implemented there, too.
Or you could also check a desktop search engine indexing your PDF files.

See also social citation systems

• CiteULike
• Connotea

References

• Image source: Old book bindings at the Merton College library. Date: 25 August 2005