Opening science: How unconferences changed my life

As I mentioned, I spent most of last week and weekend attending two unconferences, BioBarCamp and Scifoo.

By their very nature, unconferences tend not to converge on a single topic; over the past week, I paricipated in discussions whose topics ranged from the importance of database annotation to how mushrooms could save the world to the current technical considerations involved in settling Mars. Nonetheless, even in the anarchic environs of an unconference, self-reinforcing trends arise over the course of the discussions, and themes do emerge (though each participant might perceive different patterns and come away with a completely different report of an event’s most important themes).

For me, the most powerful and important theme emerging from the week was the idea of “open science.” This term refers not to any one initiative or project, but the cloud of concepts that includes open access publication, use of open source solutions (especially for protocols and software), commons-based licensing, and full publication of all raw data (including “failed” experiments). It also incorporates more radical ideas like opening one’s notebook in real time, prepublishing unreviewed results, replacing current models of peer review with annotation and user ratings, and redesigning (or ditching) impact factors. The world implied by these concepts is one of radical sharing, in which credit still goes where credit is due but by dramatically different mechanisms.

Open science isn’t so much “pay it forward” (though there is a bit of that) as an effort to create a (scientific) world in which no one is paying at all, a world in which there’s no incentive to withhold or protect ownership of data. The science fiction writer Iain M. Banks once wrote that “money implies poverty” — indeed, many of the current models of data ownership and publication, and their accompanying “currencies” of proprietorship, prestige and closed-access publication, imply a world in which data is scarce and must be hoarded. But data is not scarce anymore.

Given a suitable set of one-to-one and one-to-many agreements between the stakeholders, then, the benefits of sharing could come to outweigh any conceivable advantage derived from secrecy. Perhaps “open science” could be defined (for the moment) as the quest to design and optimize such agreements, along with the quest to design the best tools and licenses to empower scientists as they move from the status quo into the next system — because (and this is very important) if it is to ever succeed, open science has to work not because of governmental fiat or because a large number of people suddenly start marching in lockstep to an unnatural tune, but because it works better than competing models. Proof of that particular pudding will be entirely in the eating.

During the meetings, I met quite a few people involved in this mission, and I want to mention their organizations and projects here:

  • OpenWetWare, “an effort to promote the sharing of information, know-how, and wisdom among researchers and groups who are working in biology & biological engineering” – including tools for protocol sharing and open notebooks;
  • Epernicus, a social networking site for scientists that automatically connects peers based on institution, history, skills and research focus;
  • JournalFire, “a centralized location for you to share, discuss, and evaluate published journal articles” (still in beta);
  • Science Commons, the scientific wing of the Creative Commons, which “designs strategies and tools for faster, more efficient web-enabled scientific research. We identify unnecessary barriers to research, craft policy guidelines and legal agreements to lower those barriers, and develop technology to make research data and materials easier to find and use.”;
  • Nature Precedings, “a free online service launched in 2007 enabling researchers in the life sciences to rapidly share, discuss and cite preliminary (unpublished) findings”; and
  • UnPubDatabase, a discussion of ways for scientists to rapidly and efficiently publish “negative” results, both to allow re-analysis of data and to prevent the scientific community from following the same blind alley more than once.

Academic scientists aren’t the only ones to potentially benefit, by the way — pharmaceutical companies routinely run the same experiments as one another and often find that expensive trials could be avoided if they’d only had access to data mouldering in a competitor’s vault — so open science can benefit the profit sector as well, and there are already plans underway to make that possible.

I’m enthusiastic about bringing open science into my own project and my own laboratory — indeed, in a fit of post-conference ecstasy I basically put myself on record promising to do so. For reasons that have everything to do with available energy levels, I suspect that full-blown openness is probably easier to accomplish when it’s present from the beginning of a project, so I’m especially eager to put these ideas to the test in a large-scale collaboration that is just getting underway. I have no idea how it will go — I expect to meet resistance, especially to the more radical ideas like open notebooks — but it’s nonetheless an exciting time. Will I be able to convince my collaborators to try out open science approaches? Once implemented, will they work? I don’t know, but I am convinced that it’s a hypothesis worth testing.

Advertisements

12 comments

  1. Very good post. Glad to hear a report of the unconferences you attended.

    I commend you highly for your pledge to open science.

    brainhealthhacks.com

  2. It was an inspiring talk – I think you may have pushed a few other people over the edge from thinking to doing. We’ll see how it plays out.

    I agree that it is easier to do this with a new project and where all collaborators are made aware at the very start how it is going to be. IP can be a huge obstacle. In my case, the (like-minded) collaborators came first and the details of the project followed their competencies.

    I look forward to your reports on how things proceed!

  3. Nice post! It was good meeting you at BioBarCamp; sounds like SciFoo was even more exciting. Can’t wait to hear how the experiment with open science goes.

  4. Yesss!!
    Great idea!
    If we can accelerate this collaborative process we might save the world.

    First find collaborators and find ways of working with them. Leverage each others skills.
    Many people in many fields are discovering this, and would like to contribute,… if they could only find you, or a place to contribute.
    Spread the word. Share the best web sites (good start already) and best practices.
    Encourage development of the tools to assist focused collaboration, and those with common research &/or development purposes to find each other and collaborate.

    Apple’s iPhone is taking off, not just because of an excellent basic product, but because they shared their code and opened it up to creative people to develop applications.

  5. Sounds a good idea. I dont know to what extent a corporation would be willing to risk publishing it´s data, but at least at an academic level it should be popular

  6. I also doubt that corporations are going to start publishing their data, but the data suggest that some kind of sharing of “negative results” would make the drug pipeline a lot more efficient.

  7. […] Chris Patil @ Ouroboros: “Given a suitable set of one-to-one and one-to-many agreements between the stakeholders [in scientific research], then, the benefits of sharing could come to outweigh any conceivable advantage derived from secrecy. Perhaps ‘open science’ could be defined (for the moment) as the quest to design and optimize these agreements, along with the quest to design the best tools and licenses to empower scientists as they move from the status quo into the next system — because (and this is very important) if it is to ever succeed, open science has to work not because of governmental fiat or because a large number of people suddenly start marching in lockstep to an unnatural tune, but because it works better than competing models.” […]

  8. To me researchGATE is a very promising approach to combine a social network with advanced search tools. I think it is fantastic.
    Here some texts from their blog:

    “To make them even more useful, we upgraded our database to include the content of several major scientific databases, namely PubMed, Citeseer, RePEc, ArXiv, IEEE, NASA. Currently, more than 30.000.000 scientific articles from a broad range of scientific fields are included. Searching for keywords is one thing – relating research output based on content is another. We developed semantic tools that are able to do that.”

    Researchgate offers a powerful search, a new semantic search with the possibility to paste whole abstracts, which will be analysed semantically and closely related papers will be shown.

    The group functions are pretty nice. Filesharing tool, voting tool, etc.

    http://www.researchgate.net

    here is their video in youtube:

Comments are closed.