Keeping track of bibliography references for an entire research group

Question

This is a related question to my previous one about keeping advisees aware of literature.

Given the relatively large number of papers that are out there, it's inadvisable to force every student to start from ground zero in building up a reference library. To me, this suggests that there should be some centralized ways of keeping track of bibliography references.

The low-cost but high-maintenance route to me would be to have an SVN repository to which people can update their own personal bibliography files. Are there other more time-efficient routes to manage this process when:

• people have different computing platforms and workflows (Windows with Office, OS X with iWork, Linux with TeX, etc.)?
• working with collaborators at other institutes?
• it's important (according to university/workplace regulations) not to have data stored "in the cloud"?

Answer 1

I accomplish this using the groups feature of Mendeley. It works on all three major OS's, allows you to share bibliographies easily with both your group and external collaborators. It also allows something that I think is very important -- lots of bibliographies on particular topics within the realm of what my group does. See, for example

• http://www.mendeley.com/groups/1045561/runge-kutta-stability-regions/papers/
• http://www.mendeley.com/groups/1070421/nonlinear-hyperbolic-pde-solvers-on-gpus/papers/

I should mention that Mendeley's web interface to bibliographies is awful. But the desktop interface is quite nice and (most importantly) can export Bibtex.

Mendeley does store your data in the cloud (if you consider a bibliography to be "your data").

Update: I stopped using Mendeley when it was bought by Elsevier. I haven't found a satisfactory replacement.

Answer 2

If you're using LaTeX, you can set up a .bib file on a shared drive and everyone can reference and add to it as necessary using whatever program they like. I'm a fan of BibDesk, personally.

Answer 3

If you have a PHP server running somewhere, I would definitely recommend Aigaion. From the website, they say:

Aigaion provides a bibliography management environment that supports a user in just this: Organizing and managing a complete bibliography, from small personal bibliographies to bibliographies for e.g. a complete research department.

I've been using it in the past on a project, and it's really helpful to manage a bibliography for an entire group: you can define collections, you can annotate each entry, and you can export easily the bibliography at least to Bibtex and RIS (probably more, but I don't remember exactly, as I was using only Bibtex). You can at the same time manage references to be read by students, and publications generated by the group. Plus, it's open source :)

Answer 4

We used to maintain our own list on our (sadly now defunct) group website when I was a PhD student. The sources for the group website are in a shared (internal) version control repository to which we all have access. We maintained three shared bibtex-files:

1. one for internal references — everything any of us ever published

2. one for external references — anything not internal any of us has ever cited or found otherwise useful

3. one for new references; we go through them at our weekly meetings, then merge them into (2).

Simple, but works very well.

Answer 5

If you want to sync refs across multiple users but don't want to host in the cloud, you should check out sparkleshare. You'd set up a git repo on a local server to host a bibtex file, then have your users install sparkleshare on their computers and connect to the git repo. You would then use Mendeley, which has a bibtex syncing option. This will achieve a system that will distribute new refs added by any user.

Answer 6

For small scale projects, such as you and a few other authors working on paper(s), I would just consider plain text file (with standardized fields) in a location you can all access (plus version control). Bibtext would be an obvious solution (and most reference software that I know of can import bibtext files). Easy and minamalist to implement and update. Mike's answer seems like a better solution than this, but this is dependent upon all the members of the group utilizing such software, which sometimes isn't worth the effort to get people to convert.

For large scale operations (like you need to enter in over 1,000 papers for a lab) I would consider rolling your own database + user forms to enter in data. Fields you want from papers are fairly easy to delineate, so setting up the initial database is not too dificult. Here are a few other reasons this is nice;

• You can have flexibility to put what you want in the database. For instance if you were conducting a meta analysis you may want to extend forms to include relevant statistics.
• Querying on a variety of characteristics becomes trivially easy.
• If you are saavy enough with the database, you can write some scripts to export the data in whatever format you want. For instance, I use the one statistical software I am most proficient in (so not a true database, but several seperate tables) to write my bibtext library, plain text citations in approximate APA format, and VBA code to find-and-replace latex like citations in word documents.

Alot of flexibility with your own database, and if you are doing a very large project it might be worth the upfront effort to develop and customize to suit your own needs.

Answer 7

The solution is the open source I, librarian server. It manages bibliography for groups. You can try it out online here: http://www.bioinformatics.org/librarian/ Look for "demo"

Answer 8

Why not just use vanilla git ? Everyone sets up their private repo and you can use github for shared syncing as needed.