Last night (Wednesday Sept 25, 08) I attended a Talk Science event at the British Library and took part in a discussion on social networking and Web 2.0 tools in science. Timo Hannay (Web Publishing manager at Nature) gave a great keynote talk ("Scientific Researchers and Web 2.0: Social 'NotWorking'?), which was following by a discussion amongst the 100+ participants, both physical and virtual thanks to Second Life.

It was the tag line of Timo's talk that got me to attend. Despite the hype, social networking has struggled to make much of an impact amongst most mainstream scientists. Specifically it has had little visible impact on the traditional scholarly communication process, and it was this that I was hoping to hear about. Timo talked about blogging, Wikis and Nature's experiments with Web 2.0 tools. He also ran through some of the properties that make social networking "work" and also what hasn't worked. Nature has had some major successes. Notably their popular reference manager (Connotea) and the Nature Network platform, which is largely focused on blogging. But other experiments have had mixed success. For example, Nature dabbled with an open peer review system last year, but closed this when the results failed to add much beyond traditional peer review. Data repositories were also discussed, although Timo rightly pointed out that for many data types, no such repositories exist, and that Nature is (rightly) not in the business of creating discipline specific repositories, or defining data standards when a disciple lacked consensus. Unfortunately, as of right now this rules out repositories for almost everything but raw molecular data.

Much of the debate that followed focused on blogging and Wikis. While these are important I can't help but feel the discussion missed the bigger point, at least with respect to the formal scholarly communication. Participatory tools have the potential to truly change the way we share, manage and publish science. But as yet, I have seen nothing that lives up to the hype, and arguably blogging and Wikis are not going to drive this change, at least not in terms of formal communication. Blogging is a uniquely individual activity, allowing anybody to advocate any position without interference. At the other extreme, Wikis enforce sociality, such that ones efforts can be consigned to a glorified log file, at the whim of any editor. Neither of these extremes is good for science and in particular, they are not good for the people who do science. While we can ticker with ways of adapting blogs and Wikis to our needs (e.g. peer review of blog posts, controlling who has editing rights, and bean counting edits to incentivize contribution) these tools are not the game changers people suggest.

Perhaps the closest I have seen to what we need are Content Management Systems (CMS), specifically those tailored toward the needs of particular communities. These are flexible data repositories that make it easy for communities to self assemble around a common interest, and for individuals to make permanent contributions to group effort without significant formal management. Arguably they are best suited to the data driven sciences, where information management and curation makes up most of the work. For this reason I think it is the low impact, long tail sciences like biological taxonomy where they could have the biggest impact first. They can be designed with the needs of particular communities in mind, adding value to the content present which incentivize further contribution. CMS systems also centralize content such that it can be repurposed elsewhere and when controlled by self-motivated communities ensure that content always remains authoritative (potentially unlike Wiki's), without becoming authoritarian (like blogs). To my mind it is these CMS systems that have the greatest chance of supplanting replacing traditionally scholarly publications, especially for those disciplines where the publishers cannot readily (or are unwilling to) add any value to the products they offer. What is interesting is that I think it is publishers who are the least well placed to build such systems, since publishers (unlike scientists) invariably don't know the audiences they supposedly to serve.

In conversation with the Talk Science organizers, it appears they will be running a session on "taxonomies" (though not just the biological kind) in February 09. This is something to watch out for.

Thanks to Timo an the Talk Science organizers for putting on this event.