I spent the last two days in the Alpine town of Grenoble to discuss how we support science2.0 in Stellar. It seems clear from these discussions that there are quite diverse views of what constitutes science2.0.
In any case, I believe that sharing is at the heart of science2.0. This can involve
- data: For instance, we can share experimental data, so that colleagues can verify our analysis or add their own. Or we can share attention metadata for recommendation algorithms: when we develop new such algorithms, we can compare them more easily if we have a reference set of observations – much like the Netflix challenge. (Disclaimer: this is the topic of a recent proposal we submitted.)
- services: In an open research infrastructure, we can mash up common services for a specific purpose. We can configure feeds to remain informed of new publications or events that are relevant to us. Or professionals and amateurs alike can take pictures of the sky remotely through services for remote telescope operations. In Stellar, we will be setting up a directory of such services (à la programmableweb), so that you can share and find more easily what you have or want…
- applications: Shared applications like for instance the visualization tools at gapminder or manyeyes make it easier to apply sometimes quite sophisticated analysis to our data – as we recently did to analyze the growth of repositories of learning material (warning: takes extremely long to load!). Widgets are a specific kind of such applications that run in web browsers – a very early example of such widgets for science2.0 in Stellar is under development: we’ll be setting up a widget store, much like the apple iphone store, where you will be able to find and use widgets for notification of new publications, tracking citations, etc.
Another perspective on science2.0 is to focus on how generic web2.0 tools can be used by researchers, so as to scale up and make more efficient how we work together and get notified of new developments in our fields.
As an example of that perspective, I find it somewhat mind-boggling that trackbacks and other techniques notify my whenever someone blogs about me, comments on a blog post I’ve done, tweets about my work or retweets what I’ve tweeted, adds a bookmark (if I subscribe to his feed), etc. whereas it may take months or longer before I am aware that someone cites my work in a publication. That is very awkward, as the act of making such a citation is more complex and as we spend much more time and effort at making these citations ‘trustworthy’ and ‘complete’. Surely, we must be able to make citing a more lightweight and transparent process? Of course, that is what mendeley, citeulike, bibsonomy, etc try to address…
For me, all of this is yet another example of the “Snowflake Effect” at work: we need to make the research tools and environments hyper-personalized, so that what we do in research is more scalable and transparent!
BTW, if you have any good examples of science2.0 tools, practices or … whatever, then please do let me know: I will be presenting on this topic at the JTEL Winter School next week and would appreciate any help I can get.