What good can massive-scale collaboration bring to science?
There are many cases when observable macro-phenomena appear from the combination of many micro-phenomena. This is true of the interaction of elementary particles which compose gas, but this has also been proven to be true of human-related phenomena. The so-called micro and macro-economic theories describe just that, in one particular domain.
Our societies are global constructions but, at a much smaller scale, one can think of a company as being an example of what voluntary cooperation of many can build. There are many tools and techniques which are used inside a company to make sure that everyone’s work is used to advance toward the common goal. One of these is the hierarchical structure. Another and necessary tool is communication. Communication can be from one human to another through oral language, it can be from one to many through seminars, written reports, it can be from many to many through meetings and workshops, it can combine many voices into one through vote.
In recent times, technical advances in the information technology (in particular the internet) have led to a breakthrough for communication tools. Even before the web, the e-mail had drastically improved intra and inter-company communication. Outside the entrepreneurial world, forums have enabled members of self-formed, amateur communities to share their experience and help one another.
Since then, new tools have emerged, some of them allowing so-called real-time collaboration. One tool, namely wiki, has made possible mass collaboration toward the creation of a global knowledge bank. Knowledge being necessary to intelligence, I would like to call this a first step toward the creation of a global intelligence.
A lot of tools have been developed to enable mass collaboration. Sometimes, this is uninformed collaboration: people play an active role in helping out automatic character recognition via the system of reCaptcha1. This is also an example of a case when people can help out without any particular technical knowledge.
At an ACL2 seminar last year, I listened to a talk of Prof. Mike Ernst on how people can help science while playing games (he was trying to apply it to specific software verification problems but anyways). There is one well-known example of this: it is FoldIt. But this approach cannot be applied to any type of problems and it requires a lot of thinking to transform a complex scientific problem into a game.
I’m more interested in finding ways to enable people with some minimal technical knowledge (even when this is self-taught knowledge) to take their part in the scientific journey. Galaxy Zoo, a website where people can contribute to astronomy by looking at images from telescopes like Hubble and annotating them, is a good example that you do not necessarily need to gamify science to make it attractive.
Through education, people generally learn a lot on many subjects. Most of this accumulated knowledge, while helping them to forge their own view of the world, will have no future usefulness for them at all. Sometimes, people can be genuinely interested in one subject and remember a lot of what they learned, but they won’t necessarily use any of it in their future work. I am sure that some of them would be happy, if given the opportunity, to contribute to projects where this knowledge can be useful. That could mean contributing during their leisure time and for free. In fact, there are already lots of examples for what I am talking about. People classifying stellar objects are one, developers contributing to open source applications are another. There is interesting scientific literature on why people contribute to these projects2.
Given that people are willing to spend some of their leisure time to achieve collaborative side-projects, we (the computer scientists, the engineers, the developers) have a responsibility to create new tools to make it easier. And everybody has a responsibility to spread this new culture.
A lot of tools already exists. For open-source software, a lot has been thought and done already. This is not surprising: developers are helping themselves. It’s time to help other communities too.
From collaboration to mass collaboration
There is no clear limit when a collaborative project becomes massively collaborative. Given the project, the critical threshold can vary. But for all projects, the more people are involved, the more necessary specific tools are for the community. There are also lots of examples of projects that are both massive and not: a lot of open-source software rely on a small community of developers but have a much larger community of testers and even more people contribute once by sharing an idea for an improvement.
This is most likely that not all projects are suitable for massive collaboration. However, no massively collaborative project can succeed if it has not been thought carefully first and if the necessary tools are not available. Wikipedia wouldn’t be this wonderful giant encyclopedia if not for a software allowing everyone to make changes and regular users to track these changes. Can you imagine what it would be if not for this control tool?
There is a collaborative project I’m very much interesting in. It is called Polymath. People have discussed math and created proofs together for a very long time but Polymath is using the internet to invent proofs with tens to hundreds of participants. Yet, there are many more than hundreds of people who have the knowledge and ability to help with a proof. What about a proof that would involve thousands to millions of participants? At that scale, a new tool is required3.
Comments: please do not hesitate to leave a comment below. I would especially appreciate if you would take some of your time to share any information you have on any new and interesting collaborative projects you heard about or any pointers to scientific literature on the subject.
In this great TED video, the inventor of Captcha and reCaptcha also presents a new massively collaborative project called Duolingo. It seems however that Duolingo is now moving away from crowdsourced translation, not because the idea was a bad one but because the company has decided that it wants to be an education company and not to devote its resources to becoming a better translation provider. ↩
See for instance: Von Krogh, G., Haefliger, S., Spaeth, S., & Wallin, M. W. (2012). Carrots and rainbows: Motivation and social practice in open source software development. Mis Quarterly, 36(2), 649-676. ↩