What is it that means one open source project takes off, while another doesn't? There are a lot of ways to analyse this question depending on the example at hand, but a more general study of the "remixability" of online content has found a surprising correlation -- there's a trade-off between originality and the chance it will inspire new versions.

Researchers Benjamin Mako Hill from MIT and Andrés Monroy-Hernández from the Berkman Center for Internet and Society at Harvard University wanted to look at a particular dilemma -- despite "proponents of remix culture often speaking of remixing in terms of rich ecosystems where creative works are novel and highly generative", actual examples of it happening "can be difficult to find", Monroy-Hernández writes on Hill's blog.

"Although there is a steady stream of media being shared freely on the web, only a tiny fraction of these projects are remixed even once. On top of this, many remixes are not very different from the works they are built upon. Why is some content more attractive to remixers? Why are some projects remixed in deeper and more transformative ways?" Examples of "peer production" like Wikipedia and Linux are the exceptions, not the rule.

Luckily, a large dataset from an active community of remixers was available for them to test out some hypotheses on. MIT Media Lab's Scratch is an online community for children that teaches them how to make simple animations, simulations and programs, and encourages them to share them online for others to remix.

It's extremely popular -- not only have more than three million projects been created by more than a million users and shared by its community, educators around the world are considering  adding it to their IT curriculums -- and its open standard gave Hill and Monroy-Hernández a sample dataset of 536,245 projects by 105,317 unique users taken from across the year-long period of 2010.

As they write in their paper -- "The Remixing Dilemma: The Trade-Off Between Generativity and Originality", published in American Behavioural Scientist -- the two researchers used the sample to explore several hypotheses about remixing, originality and "generativity" (the word they use to describe how likely a work is to inspire remixing).

These can be grouped into two apparently paradoxical main theories. The first is that the most generative projects are going to be ones that are moderately complicated first-generation projects by prominent authors, while the second is that the most original projects are going to be ones that are moderately complicated remixes of remixes by less famous creators. In other words, as Monroy-Hernández writes, "we suggest that each of the factors associated with generativity will also be associated with less original forms of remixing".

The theoretical reasoning behind these hypotheses is  discussed in detail in the paper, referencing P Diddy's 1997 "Every Breath You Take"-sampling hit "I'll Be Missing You" along the way, but the key thing is that the more likely your work is going to get remixed, the less likely it's going to be in a genuinely new and interesting way. They call this trade-off between originality and generativity "the remixing dilemma". The Scratch data confirms each side of the idea.

To test whether the two are linked, though, required re-testing with a new dataset of every project on Scratch that has ever been remixed and tracing it back to its original project. They then created a value for what they call the "edit distance", a combination of variables that together represent just how original a remix is -- including the amount of code the remix shares with the original, the number of times a project was viewed versus the number of remixes made and the demographic breakdowns of users and the time between remixes appearing.

The data showed the remixing dilemma held true, with originality increasing as generativity decreased. This is a troubling conclusion for peer production, open source projects and so on, because it implies that truly innovative technology is the opposite of what collaborative remixing produces.

In the blog, Monroy-Hernández writes: "We feel that our results raise difficult but important challenges, especially for the designers of social media systems. For example, many social media sites track and display user prominence with leaderboards or lists of aggregate views. This technique may lead to increased generativity by emphasizing and highlighting creator prominence. That said, it may also lead to a decrease in originality of the remixes elicited. Our results regarding the relationship of complexity to generativity and originality of remixes suggest that supporting increased complexity, at least for most projects, may have fewer drawbacks."

"As supporters and advocates of remixing, we feel that although highly generative works that lead to highly original derivatives may be rare and difficult for system designers to support, understanding remixing dynamics and encouraging these rare projects remain a worthwhile and important goal."

Image: Shutterstock

Latest on