"Open" vs remixable

"Open" just isn't enough. I've carefully avoided defining it, in part because there are people much more legally inclined than I am who know far more about it, and also because it's a topic that inspires online religious wars, and also because even advocates mean different things by it. Plus, corporate interests are attempting to redefine "Open Access" to mean whatever is best for their bottom line.

So what should 'open' mean? We need to maximize reusability or remixability above and beyond access. (I think the right distinction is gratis vs libre, or "free as in beer vs free as in freedom". We want research to be libre, not just gratis.)

The scientific process we've developed through several centuries is predicated on remix with provenance: you publish a paper to communicate your results and work, I build on the paper and write my own paper, which cites your paper. (Michael Nielsen has a great and much more detailed description of this in Reinventing Discovery.) This is generally compatible with the idea that very few, if any, ideas are entirely new -- we stand on the shoulders of giants -- and our job, as scientists, is not only to advance our field but to do so in such a way as to elevate the perspective of those who follow.

This is in what is danger: our ability to remix and reuse not just research ideas, but research data and process.

The pernicious effects of Bayh-Dole on academic research is one example of this threat -- I hear that negotiations over material transfer agreements one must sign in biomedical research is stifling both basic and translational research, for example, because of the amount of money that could be made. Closer to home, the GATK folk are close-sourcing their variant caller, presumably to make a buck or three.

I think a bigger threat to remix culture in science is likely to be the rise of "mixed model" online startups that are attempting to make a buck off of social data and online material. For example, courtesy of Greg Wilson comes this coulda-seen-it-coming story about Coursera effectively blocking free institutional use of their materials. Less obviously silly examples for research (but just as limiting) are the changes to Twitter's TOS that block If-This-Then-That; the rise of "alternate" social networking platforms that aim for researcher lock-in but don't provide data liberation APIs; and pretty much any 'free' but not 'open' set of services. There are obvious market reasons for such companies to restrict export of data and broad free use of their software, but I think that they are bad for the future of science.

If you want a positive example, I think Figshare gets it right, by explicitly buying into a cultural model of remixing by using Creative Commons licenses.

Irony in this space abounds, BTW. I feel compelled to mention the Executable Papers Grand Challenge, brought to you ... by Elsevier. Or one of the first two IPython Notebook executable papers being published in ISME, a non-open-access journal.

The bottom line is this: we need to ensure that open access, open science, open data, and the rest all permit and encourage remixing, in much the way that open source has institutionalized it with licensing. This is just as important for the social networking metadata of science (or, as we call it, "collaboration" :) as it is for explicit research products.

Generational change isn't fast enough

I'm sure you've all seen this quote --

"a new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it."

This quote, from Max Planck via Thomas Kuhn, is used by Kuhn in illustration of the dynamics of paradigm shifts. It's perhaps more succinctly but even more cynically stated as,

Science advances one funeral at a time

So, sure, if open* is inevitable, we could simply wait for all the opponents to retire, and switch the fight over to making sure that it's remixable (see point #1, above). But I think we need to figure out how to accelerate the process. Why?

We face lots of big, complex societal problems, and my bet is that the single-scientist approach isn't going to work to solve them. In my own domain of expertise, biology, we increasingly rely on big and heterogeneous data sets, produced by other people, to interpret our own data and generate hypotheses; these resources, while necessary, rely on open principles that are not well supported within the current incentive structure.

Now, what's interesting is that it's clear that the funding agencies get this, at least from the informal discussions that I keep on having with program officers about Big Data, data use and reuse, data integration, and publication. Everyone is aware that it's a problem and that the production, curation, and interpretation of big data sets is critical, as well as the development of effective and usable software. But we don't really know how to incentive this.

As an increasing amount of effort is put towards generating data sets and correlating across data sets, funding agencies are certainly trying to figure out how to reward such effort. The NSF is now explicitly allowing software and databases in the personnel BioSketches, for example, which is a great advance. Surely this is driving change?

The obstacle, unfortunately, may be the peer reviewer system. Most grants and papers are peer reviewed, and "peers" in this case include lots of professors that venerate PDFs and two-significant-digit Impact Factors. Moreover, many reviewers value theory over practice -- Fernando Perez has repeatedly ranted to me about his experience on serving on review panels for capacity-building cyberinfrastructure grants where most of the reviewers pay no attention whatsoever to the plans for software or data release, and even poo-poo those that have explicit plans. And if a grant gets trashed by the reviewers, it's very hard for the program manager to override that. The same thing occurs with software, where openness and replicability don't figure into the review much. So there's a big problem in getting grants and papers if you're distracting yourself by trying to be useful in addition to addressing novelty, impact, etc.

The career implications are that if you're stupid enough to make useful software and spend your time releasing useful data rather than writing papers, you can expect to be sidelined academically -- either because you won't get job offers, or because you won't get grants when you do have a job. A few program managers are very concerned about this, because it means that the more competent and hands-on the person is, the more likely it is that they will not be able to stay in academia. I'm watching this happen with some of my own students, who are very good at data analysis and software development, but don't want to try to make it in academia; and because they have plenty of other good options in industry, they leave. It's a real problem.

So unfortunately, I don't think it's going to be as simple as getting the funding agencies to push. Where are other lever points?

One lever point that I think is ripe for attack is tools. We lack good tools to robustly support good publication of process and data, and it's unreasonable to expect scientists to learn data-base backed Web programming in order to publish a paper. (As Greg Wilson likes to say, we don't think we can teach people enough about Web programming to let them do anything but create security holes.) I'd guess that we need both incentives at the funding level -- because honestly, it's one of the only ways to get scientists to do anything -- and enabling technology that lets scientists publish process and data easily. And yes, this guess is the underlying motivation for many of my wanted tech and ideas for the future

I think it would also be interesting to figure out how to hack academic culture, but I'm not sure how to begin that mammoth undertaking :). Good tools would certainly help.

Enabling citizen science

Another big obstacle, especially here in the US, is the lack of engagement with scientists. Corporate and liberal anti-science agendas are crippling our ability to intelligently discuss evolution, climate science, vaccines, and nuclear energy -- just to name a few hot-button topics :). Most people just aren't that involved in the scientific process, which makes it easy to snow them. What's the solution?

Personally, I think citizen science is a pretty neat idea. I was up at Mozilla in Toronto a few months ago for a Software Carpentry summit, and a group of us sketched out a bunch of ideas on how to enable non-academics to interact with data and process. (In fact, the first four ideas in my wanted tech list come partially or completely from that discussion.)

One specific idea I had at the time was to integrate a sort of storify-style ipython notebook interface with the Microbiology of the Built Environment data, to enable individuals to examine and correlate their own household microbial fauna across households and place them in a global context. The Earth Microbiome Project also plans to have a citizen science component to enable contributions; I think integrating citizen-contributed data into a broader context is a pretty neat idea.

I also believe that Rich Lenski and BEACON would probably be interested in putting up an integrated interface to all of the Lenski Lab Long Term Evolution Experiment data. This is less "citizen science" and more "outreach" but it would still be neat.

So, what's my utopian dream? It'd be awesome to enable more, better, and deeper citizen science by enabling easy publication of and rich interaction with big, open data sets. Part of this is swiping good community hacking ideas from massively online collaborations like the Polymath Project and the Galaxy Zoo; another big part is developing tools; and a third is figuring out what fields are ripe for this kind of thing. Coming from biology, I feel like efforts such as microbe.net, the Earth Microbiome Project, ENCODE and caBIG are data rich and ripe for more interaction with a larger community; surely this is true of many other fields, too.

Concluding thoughts

We've made a lot of progress even in the year or two since Michael Nielsen's book was written -- the advances in open access alone have been stunning. But I don't think Open Access is enough; we need to enable more scientists to be makers, not just consumers, of data, code, and online communities.

--titus

p.s. I want to thank Fernando Perez, Carole Goble, and Cameron Neylon for pointing me towards a bunch of this stuff, and Greg Wilson for critical commentary. Interpretations are my own but I bet they agree with me!