Timo Hannay, Nature Publishing Group
Focused on five main areas:
- audio & video
- open data
- peer review & blogs
- online collaboration
With regard to the first, he pointed out that what was once a very expensive venture, namely, the creation of professional audio and video content, is now within the reach of most people in the industrialized world, and becoming more accessible for everyone else. Nature has been experimenting with podcasting for a couple of years. He pointed out that listening to a podcast of a scientist explaining their work in five or six minutes makes it much more accessible than reading their published articles, so it creates new roads to their research. It's had a couple of unintended benefits for them. For one, some people have asked for longer podcasts for their commutes or other downtime. Also, many non-native English-speaking scientists use them as a language learning tool, and actually asked for transcripts to be able to read along, which Nature now provides.
Their move into video is somewhat more hesitant, since both production and consumption require more than audio. He noted that much of it is still 'talking head' video and not terribly compelling. He did show Jove, which he referred to as "cookery television for scientists" (n.b.- I do not believe that this is a Nature production, but am not sure). It shows experiments being conducted and offers instruction and training to viewers. Neither audio nor video offers any inactivity, which he views as something of a limitation.
Moving into databases, he showed how a database structure for journals allows new utilization paths, such as interlinking. Some forward thinkers in science postulate that in the not-too-distant future the journal will be replaced by a database. In other words, the online journal will be less of a facsimile of the print edition than a richer presentation of research, inlcuding queryable figures and text, backed up by data sets, etc.
Turning to open data, he first showed RSS feeds as an example. Moving beyond the basic function of a feed, they added richer data to their feeds, such as Dublin Core data, with no clear idea, initially, of how it would be used. In other words, they added it and hoped for applications to arise, mainly because it's only nominally more difficult to issue more information since it's all available in the source data. As things have evolved, the apps have come along, so it was a successful experiment.
One step they've taken to assist indexing services and the like is to offer access to their article data in non-human readable format (OTMI). It's useful for text mining and indexing, but does not compromise their publishing model. They offer it in word vector format, or in snippet form, where the sentences are offered out of order in an XML wrapper. OTMI Twease at Cornell is a proof-of-concept search interface that shows that this model works. Showed another example of one of their journalists, Declan Butler, who offers--via his blog--an open data source in GoogleEarth format on avian flu.
Regarding peer review, he noted that Nature has a self-interest in being part of the changes to peer review, since that is part of the added value that scientific publishers add to scholarly communication. They did an experiment in open peer review, inviting open commentary on articles. It was not successful. They are now looking into post-publication commentary for their publications, to see if it generates more interest.
They've also dabbled in automated peer review, using database information to assess the web of authorship attached to an article and identify the best reviewers. He's not sure of its current status, but deems it an interesting idea.
Nature also offers blogs. He points out that it mimics, in some ways, the scientific communication process, since one cites sources, discusses information, offers commentary and countercommentary. He does not feel that scientific blogging is not yet mainstream, but wishes that more scientists would blog. One obstacle to wider adoption is that they get no credit for blogging in academic terms. There are developments in this area, including sites (postgenomic.com) that track scientific blogs and offer tools to finding the most-discussed papers and so forth. A new experiment Nature offers is Nature Precedings. Again, he notes that the lack of incentives hampers the adoption of such services.
On his last topic, online collaboration, he showed sites such as Nature Protocols, OpenWetWare, and UsefulChem, which are intended to share information about lab protocols and other lab-related information. It's an "open approach to science." He also mentioned Connotea, of course, which is a del.icio.us-like tool for science, in other words, social software for science. An implementation of Connotea data is in EPrints, where Connotea tags can enrich the page for a given publication. They've also set up Nature Network, which he referred to as "Facebook for scientists."