What's keeping us from Open Science? Is it the powers the be, or is it … us? - David Dobbs, Melody Dye, Jan Reichelt, Kristi Holmes, John Timmer, Sara Wood
Timmer: more research – research that creates a lot of images from microscopes for example – takes the place of stopping to add metadata and to share – incentives. The exception was the PI who had had a postdoc falsify data, so in that lab there was good documentation, preservation, sharing. There are all kinds of policies about sharing, but it’s not enforced. Any of the players can add incentives and enforce policies are not doing so. This needs to be done
Dye: opening up peer review. there’s a really interesting discourse between you and the reviewers, trying to make sure they understand what you’ve done – that valuable discussion is lost. Brain and behavior has expert commentary on pieces – that’s useful (but not anyone). ArXiv.org is her other example. Makes and open, transparent community discussion possible.
Problems – greater reliability. Reviewer agreement is low. If you open up comments, then the specific disagreements can be discussed. We need a revenue model to move over to open – she suggests we don’t need a revenue model in place from the get go. The role of the traditional journal is filtering – it’s not clear that we need that in the age of the semantic web (huh?)
Wood: data is all different types. data from industry (finance, sports) is seen as a commodity and is easy to find. things like the poverty rate of Vietnam are hard to find by googling (however, not as hard to find if you ask your librarian and she shows you the UN stats).
problems- what exists, culture (reputation), standards, $$, analysis/real-time
Holmes: semantic web overview. vivo mention
Reichelt: Mendeley. some of the incentives are that it helps the scientists use the data they have on their computers better – find things that they’ve saved, reuse articles. Its not a yes or no about scientists being willing to share, it’s about the granularity of sharing. Share only some things, or only to some people.
Q: tied to a platform, long term sustainability when research will be going for 20 years
A: ability to export, apis, data are aggregated and reused. can synch to different platforms.
Q: scientists are willing to share at meetings but what about if the meetings are online – even worse from another person, can’t even share any unpublished data at meetings
A: do need fine grained control of what is released. If you are documenting your data as you go, then could set it up to share the data upon publication.
General commentary: it’s not all about easy to use tools. It’s incentives and making it part of the workflow.
Comment: it’s not only about the data – articles tell a story. If the project takes 3 years – who will follow that story if you don’t make your papers available.
The comments of the reviewers should be part of the stream
culture – not to receive comments directly on the paper – even when it’s just a typo, sticks out as a negative mark about the paper.
Problem when a direct comment on a paper is seen as insulting the author – compare to the discussions on the talk pages on wikipedia
Beyond open data, we need the paper, we also need open methodology to go with the data
guy works across two research areas – one area shares everything, the other nothing.
what can equipment manufacturers do to support this sort of thing – make it easier to share in a standardized format.
A: it’s awful for scientists to use the equipment so they need to fix that first – but the tools aren’t even interoperable with each other, nowhere close to the web.