New workflow for reference management

2 minute read

Managing references is an important yet borisome routine for academic researchers like me. Recently, with the help of Mendeley, I have much improved my workflow for reference management.

My old workflow

  1. Papers are scattered. Perhaps like many of you, I usually work on multiple projects at the same time. For each project, I will have a folder, say, projects/project-X/papers where I put all relevant papers for project X. In addition, as I use bibtex, I will have a bib file project-X.bib in the same folder as well. The benefit of doing so is obvious: you can easily find all papers with the same topic as project X. But as you can imagine, there are several drawbacks. First of all, it might well be the case that a paper is revelant to several projects and as a result I often end up with several copies of the same paper. What’s more, I have to maintain several bib files to keep them all update to date.

  2. Most of the workflow is manual. After I download a paper I will have to rename it manually (not only because it is really uncomfortable for someone with OCD like me to leave a file with its original name, but also for the benefit of looking it up in the future). Later on when I decide to cite it, I will have to google it again to get its bibtex added to project-x.bib, and possibly also project-y.bib, project-Z…

My new workflow

With the help of Mendeley, I have come up with a much better, new workflow.

  1. All papers are in one place. Now papers from different projects are all in a single folder /projects/papers and there is only one bib file papers.bib there. In such way I have no duplicates and only need to take care of one bib file.

  2. Most of the workflow is automated. Here are the tricks.
    • First, Mendeley can help watch the folder /projects/papers such that whenever I download a new paper there, it will be automatically added to Mendeley’s library and matched against its catalogue to fetch all essential information such as authors, year, and etc. One thing good to know is that this automatic fetching does not work every time, especially with old papers in scanned pdf. In such cases, I will get the paper’s doi (which normally can be found on the paper’s jounal page) and supplement it into Mendely which then helps to complete all the rest missing information.
    • As for the bib file, Mendely can be link its library to the papers.bib file and it will be automatically updated whenever a new paper is added or modified. As a result, the bib file can always stay up to date. But of course that demands a clean library, which admittedly may not be easy when starting with lots of messy legacy files!
    • Last but not the least, Mendeley can also help rename the original pdf files based on their information and the naming rules is quite flexible. Besides, the renaming can be done in bulk. What a bless for people with OCDs!
  3. Systematic paper look-up is much eaisier and faster. Having a virtual library in Mendely enables me to filter papers by authors, journals, user-defined tags and etc., which is much more versatile than creating separate physical folders to contain papers of different projects/topics.

If you are convinced, just Google a bit and you will find lots of guide on how to start with using Mendeley. In case you need some more advanced tricks to play around, here is a pretty good source.