Workshop: The Influence and Impact of Web 2.0 on Various Applications

Last week I attended the workshop on The Influence and Impact of Web 2.0 on Various Applications. It was organized by Mark Baker from Research3.org. It was an extremely interesting and productive workshop, below I will try to summarize some parallels between presentations and LiquidPub.

Duncan Hull presented Bibliography 2.0 – experience of Wellcome Trust Genome Campus in using CiteULike. The scenario was similar to the selection phase in Liquid Journals – people tag articles they somehow found on the Internet and then most tagged articles are sent out in a weekly digest. There was a set of tags people agreed on beforehand (which seem to correspond to several Liquid Journals or issues in our scenario). The results provided me with some insights on what we can see from selection behavior – typical long tail distribution.  Also, it was not clear how to evaluate quality of selection (e.g. whether people appreciate more papers from the guy who bookmarked 500 papers or from the guy who bookmarked “only” 150, but more interesting ones?). It might be possible to add hash in links so that we can see which links were actually followed by the users (RIN does something similar in their newsletters).

What Liquid Journals could add in Bibliography 2.0 scenario? Automatic gathering of content from Internet sources (publishers’ websites, Google scholar, etc.), privacy settings (share only with me, my friends, friends of friends, whole network), getting not only papers (but also blogs, datasets, scientific workflows). Also, from the scenario we can defines some “desirable” features for LJ – tell whenever [the right] Mark Baker published a paper, new paper matches your keyword alert (those two, apart from “the right” part, are already implemented by Scholar), and, going further, may be about new papers appearing in a certain community.

Roy Williams presented application of Web 2.0 for providing streams of news about astronomical events (appeareances of super nova, asteroids coming too close to the Earth, etc.). Even though the domain is completely different, their workflow of identifying “interesting” events looks extremely similar to the workflow of identifying “interesting” scientific contributions.

David de Roure presented MyExperiment, a virtual research environment for sharing scientific workflows (description of how you conduct an experiment). The presentation confirmed what I knew before the workshop – a scientific workflow could be one of contributions in LiquidJournals. MyExperiment lists more than 1000 workflows and also has a programmatic access via REST API, so they should be easy to crawl. Also, the notion of Research Object is extremely similar to the concept of SKO in LiquidPub. Possible future uses of such concept that were mentioned:
– can I have a copy of your research object?
– what can I do with your research object?
are similar to the questions we have when talking about licensing and copyright, in particular in Liquid Books.

If you want to know more about the workshop, presentations and videos will be available at the workshop website. You can also see the tweets of participants or Andy Turner’s blog, nicely summarizing all discussions.

P.S. And my presentation on Liquid Journals is available here.

Examples of Scientific Knowledge Objects (SKO) in the real world: EuroPLoP papers

While preparing EuroPLoP proceedings for submission to DBLP, it occurred to me that EuroPLoP conference follows versioning model followed by the SKO theory. Each year, authors submit papers, which are being improved for three months of shepherding and, sometimes, for two more months left before the conference. Then, versions of the papers being workshopped at the conference appear (and stay forever on the conference page, see, for instance, papers workshopped at EuroPLoP 2003). After the conference, authors have five more months to refine their papers, and post them to the final and official proceedings (e.g., EuroPLoP 2003 proceedings). What is missing now, is an explicit link between the two versions. However, it is easily fixable – author names and, often, titles remain the same. For instance, have a look at the workshopped and final versions of Pippin Barr’s paper on “Interface Ontology: Creating a Physical World for Computer Interfaces”: several figures were added and the paper was shortened, apart from more subtle changes in the content (to discover latter, you should read both papers). Of course, I guess, not all papers make it to the final proceedings, or some are exactly the same as at the time of the conference.

Similar format is adopted by interdisciplines. As far as I know, some conferences or workshops in computer science also have post-proceedings. This indicates that the real world provides a lot of examples for bootstrapping SKOs (in this case, the versioning aspect) and testing this LiquidPub line of research.