We submit this proposal under the assumptions that (1) the sprint will be online and (2) there will be few submissions so that organizers may consider a project outside the normal maturity limits. It is “bims: Biomed News”. It’s mature. But it’s misunderstood for many reasons: (1) it’s unprecedented; (2) it has no external support; (3) its site is drab; (4) its benefits can’t be demonstrated immediately; (5) it solves two problems at the same time. Bims connects readers with new papers selected by an expert on an area of interest. To qualify, papers should be new, contain new results pertinent to the subject area and be reasonably correct. It’s the job of the selector to ensure that. The benefits are: (1) the selector stays current in the subject area and (2) the selector can circulate the results to enhance recognition. There are three public benefits. First, democratizing access to knowledge. Most people can get access to papers by writing to the authors. We tell them what papers to get. Second, we could help emerging preprint efforts. Preprints are not classified by journal. We can help them find an audience. Third, we can help in the fight against fake science.

Bims has a basic technology in place. There is a public facing web site at http://biomed.news. This looks very simple, but there are serious technical chops. Thomas developed a PubMed indexer that finds the new papers every week. He wrote a specialized web interface used by selectors at http://ernad.biomed.news. It has five screens. The first or last screens are used to select issues to work on or tell users that there are none left. The second is to select papers that are on the topic. The third screen is essentially the same as the second. It’s used to narrow down papers that are on the topic by filtering for other aspects. Finally, there is a screen that allows to change the order of the papers. To aid selectors, Thomas makes sophisticated use of SVM-based machine learning. It’s great if the report is on real scientific papers. It’s good enough otherwise.

As we pointed out, we are sending this proposal under the assumption it will take place online. We also expect that the amount of time spent by participants is reduced per day as we can’t expect participants in remote time zones to be up at ghastly hours. We expect essentially three activities. First, Gavin to provide background into the organizational aspects. Thomas will shed light on technology aspect. Then participants will be tasked with going through the web site to try to find ways that we can better communicate what bims actually is. We expect that there will quite a bit of time, as clearly communicating the nature of the project is an unsolved challenge. Finally, participants will have a chance to test the machine learning and experience for themselves how much more powerful it is compared to repeated keyword-based queries.
We do not expect to make any substantial change to that infrastructure during the sprint. Instead, ahead of the sprint, Thomas will reverse the course of time, so to speak. He will prepare an enhancement that will allow the release of back issues. Thus, Sprint participants can try the system on issues going back to February 2017.

As we said, we aim this to fit a reduced event. We don’t intend to do technical work. Thus, we need subject experts and some good communicators. But since there is broad coverage of themes, anybody can start working on a subject just to get a feel of how it works.

Participants will have a broad interest in scholarly communication. Bims is right of the heart of it. In that sense most participants will be able to relate to it. In the system test, people will be working on their own issue. PubMed has both technical and non-technical biomedical research papers, so everybody will have some topic that they can work on, and potentially keep after the sprint is over.
As to what contributors get out of it, well let’s face it … a lot of the benefit of the sprint is actually the fun of travelling and seeing new places and meeting new people. This will be gone under the assumption of an online event. Bims is radically new. It gets very interesting once you start looking into the details of how it works. Nobody will leave feeling bored. In addition, we think we can make contributions by recruiting participants through talking to the people we already have lined up as selectors. True this will not be masses, otherwise bims would not need the sprint at all. But the project has a tiny dedicated fan base.