Mendeley, Zotero, Papers, Citeulike, and others are all playing the social reference management game. You store your PDFs in their excellent programmes and then you can start to be social; form groups, browse subject categories, subscribe to other people’s reference lists. My question is this: is the current implementation the best way forward for us, the users, or is it driven by the interests and/or old-fashioned thinking of the providers. For those of you just looking for the punchline, I think the answer is the latter, best for them, not us. We should be learning lessons from social networks, and the open-science movement, and adopting open standards before some FaceBook-like behemoth emerges and dominates social scientific reference management for the foreseeable future.

Social Reference Management

Reading scientific literature is at the heart of all science, and whoever you are there is always a little paranoia that you haven’t read the latest important thing in your area. The social web has a lot to contribute here. When I find an interesting reference I click to add it to my (say) Mendeley library, and subsequently post it to a shared group folder such as “Gene Duplication” or “Stuff Dave is Reading”. Other people can then follow that collection from within Mendeley, but outside of the Mendeley ecosystem it becomes much more difficult. I’m not picking on Mendeley here, they have done a lot of hard work in creating some excellent software, and they are a commercial company after all. But wouldn’t it be better for me, and for science, if I could subscribe to lots of people on lots of networks without barriers?

Social Networks

It’s almost impossible to consider this topic without introducing analogies to social networks like Google+, Facebook, and Twitter. In these areas however there has been a lot of talk about privacy, data ownership, and freedom of connectivity. If I am on FaceBook I can interact with other FaceBook users just fine, but how can I bring in someone on a different social network? I can’t, I’m locked in and able to operate only within the environment designed by FaceBook. Similarly there are issues of data ownership- FaceBook has famously caused some concern regarding who owns the photos that are uploaded. But there is another side to data ownership, that is not the ownership of the data itself but of the data’s organisation. I might be reluctant to leave a social network, not because I can’t keep a list of friends, but because I have dozens of people partitioned into different groups, organised so that I can follow work-related contacts in a different way from drinking buddies.

Open, Distributed, Semantic Social Networks

The web discussion of social networks have lead in one very promising direction- away from locked in “data silos” towards a consideration of open standards where different social networks speak the same language and you can communicate across networks. In addition there are much fewer concerns about what rights to the content you post you are giving away if that content stays on your server, or otherwise in your control. Similarly, if we rely on the export tools provided by a service provider, they are always likely to be poor. Why invest time developing tools to make it easier to leave your network? Sure you have to have an export option, otherwise you will get bad press, but the rabble won’t rouse up across the web just because your export is mediocre and loses much of the organisational content people have invested in. And its not like people can fix this themselves, because you control the software environment. There are many open distributed social networks including Diaspora, OneSocialWeb, gnusocial and Friendica although they have struggled to make significant inroads in terms of user base and some are no longer in active development. What if we used their code for open scientific bibliographic social networking?

Open Social Reference Management is important

I would like an open standard social system of reference sharing. This could be done within their current system, it might not even look that different to the present, it could certainly be built to match the look and feel of whatever reference software system you are in. But the important difference is that you could follow a group, or recommend reading to anyone, no matter what system they were using.

Yes this would be difficult, but not very difficult. The open licences of some of the existing social software could be challenging for commercial ventures to include (though perhaps not Zotero, which is open source). Linking the reference to the PDF could again be challenging, especially if your business model is based on selling user storage. There are many other things that could also be challenging, but hey, really big advances are always difficult. I currently find a lot of references on twitter and G+ but they are disconnected from my library of literature, sure they are social, but nothing more.

Hyperbole alert: Science is built on knowledge. Reading that paper, which sparks the idea or makes the link that eventually produces (name of the coolest advance you can think of). The group that builds the infrastructure that truly links scientists’ reading and social knowledge-sharing across the world will save lives, and protect biodiversity, and build rockets to Mars. What could be cooler, or more important than that? And all we need is Diaspora* for journal articles, to link Mendeley to Zotero, to CiteUlike to Papers, how hard is that really?

This is worth reading: “A flock of twitters: decentralized semantic microblogging“