Researchr is a new tool for indexing, managing, and sharing bibliographic information of scientific publications.
Rather than laying down a mission statement with a grand vision of all the things that researchr is and may become, I plan to talk about the features of the site and the motivation for them in a series of blogs the coming months.
WebDSL is my first production application of WebDSL, the domain-specific language for web applications we are developing in Delft. (Other applications in production are tweetview and the webdsl.org site.) So a motivation for building researchr might seem to be to exercise WebDSL, which it is. But another perspective is that WebDSL made it feasible for me to even contemplate building such a complex application. I’ll talk more about the various aspects of the design and implementation of researchr and the way these are supported by WebDSL in the future. Tools for bibliographic information have been a long time hobby; the Stratego Bibtex Tools still feature on the first Google page. I use these tools to generate the various publication lists on my homepage.
A concrete trigger for starting the researchr project was my SCAM 2005 keynote paper. A very cute paper about the definition of extensible languages with Stratego/XT. The ideas in the paper are the basis of current work on composite, extensible DSLs, which WebDSL is an example of. However, the paper is completely ignored, despite good reception of the talk at the conference. I’m probably to blame for not advertising it better, but partly I blame DBLP.
As you probably do, if you publish in the computer science literature, I regularly check my DBLP page to see what I have published. Or rather, whether what I know I have published is also indexed where it counts (or is counted, literally, when I would apply for a job). But my SCAM paper would not show up. I assumed it was an issue with IEEE proceedings not being indexed by DBLP, but at some point I found out that all papers in those proceedings, except mine were indexed by DBLP. Assuming this was an error, I wrote an email to Michael Ley. And after getting no response, another one, and another one, until I gave up. I understand that such things can happen. There is just too much data to process. I don’t know how many people are responsible for managing the DBLP data set, but whatever it is, it is probably too few. This results in arbitrary decisions being made about which publications are indexed and which don’t. And that is the problem of DBLP and and similar `cold’ digital libraries; they are managed by a few editors, and are not open for us, the researchers producing and consuming the data.
So the first problem researchr addresses is to allow us producers and consumers to manage our own bibliographic data. And not just ‘our own’ in the narrow sense of the data of our own publications, but the data of all publications we care about, i.e. read, review, and cite. In addition the data should have the same high quality as the data that is in DBLP, and the effort of preserving that high quality should be shared.
If you would be interested in beta testing the site send an email to firstname.lastname@example.org